Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Anemia and white matter: a diffusion MRI analysis of the human brain
(USC Thesis Other)
Anemia and white matter: a diffusion MRI analysis of the human brain
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
ANEMIA AND WHITE MATTER:
A DIFFUSION MRI ANALYSIS OF THE HUMAN BRAIN
by
Clio Gonz´alez Zacar´ıas
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)
December 2024
Copyright © 2024. All rights reserved. Clio Gonz´alez Zacar´ıas
Doctoral Committee:
Professor Paul Thompson, Chair
Professor Richard M. Leahy
Professor John C. Wood
Associate Professor Jason Kutch
Professor William J. Mack
“”If I have seen further, it is by standing on the
shoulders of giants.”
— Sir Issac Newton
This dissertation is dedicated to the giants who made my Ph.D. possible:
To my beloved father, Abdon Gonz´alez-Cisneros,
and to my esteemed Ph.D. advisors,
Professor Richard M. Leahy and
Professor John C. Wood.
iii
Acknowledgments
“Many times a day, I realize how much my own
outer and inner life is built upon the labors of my
fellow men, both living and dead, and how
earnestly I must exert myself in order to give in
return as much as I have received.”
— Albert Einstein
“Science is a collaborative effort, and our
understanding is built on the foundation of what
came before and what others contribute alongside
us.”
— Richard Philips Feynman
I am profoundly grateful to my advisors, Professor Richard M. Leahy and Professor John C.
Wood. I feel incredibly privileged to have learned from both of you, and the lessons you have
imparted will guide me throughout my professional journey.
To Professor Richard M. Leahy, your unwavering encouragement and insightful feedback have
been nothing short of transformative; when problems seem insurmountable, you were always there
to illuminate the way forward. Your vast expertise shaped both my research endeavors and academic
development in ways I never expected. Your remarkable ability to challenge me to think critically
and strive for excellence has been a continual source of motivation. Each discussion with you
opened new avenues of thought, inviting me to push beyond my limits and transform my ideas into
meaningful contributions to the field.
iv
I will be forever grateful to you and Stephanie for welcoming me into your family —the memorable
Thanksgivings, the joy, the music, the games, and the conversations that have enriched my life in
the USA.
To Professor John C. Wood, your unique perspective, constructive feedback, and enthusiasm
for science have been invaluable in sparking my curiosity and passion for discovery. You have been
an incredible mentor who has encouraged me to explore novel ideas. Your thoughtful advice has
profoundly enriched this dissertation, instilling in me the values of dedication and perseverance.
Your confidence in my abilities has been a source of motivation that I will carry with me throughout
my career. I want to extend my heartfelt thanks to you and Ruth for welcoming me into your home.
The lab celebrations were truly memorable, and the delicious meals brought everyone together. I
greatly appreciate your personal advice; it has been truly invaluable to me.
I want to take this opportunity to express my sincere appreciation to the members of my
dissertation committee: Professors Paul Thompson, Jason Kutch, and William J. Mack. Your
contributions have shaped both my research and my academic path. Your insightful questions and
suggestions have inspired me to approach challenges from new angles, significantly enhancing the
depth and rigor of my work. Each of you has played a vital role in my development, and I value
your time, patience, and belief in my potential. Thank you for your unwavering commitment to
my success.
I would also like to express my gratitude to everyone I worked with at USC and CHLA. I
especially want to thank Professor Anand A. Joshi, with whom I collaborated in the BIG Lab, for
all the times you stepped in to help me troubleshoot the problems du jour. I also want to thank
Dr. Peter Chiarelli for your valuable input on the interpretation of results, as well as your feedback
on my manuscripts and abstracts.
v
Additionally, I wish to extend my sincere appreciation to Professor Mary Ann Murphy. Her
dedication to facilitating effective communication, providing thorough feedback, and offering insightful recommendations has immensely enhanced the quality of my dissertation and publications.
Moreover, her guidance has inspired my development as both a writer and a scholar. I am genuinely
grateful for her willingness to share her time, expertise, and encouragement throughout my academic
pursuit.
I want to extend my sincere gratitude to the members of Professor Leahy’s Brain Imaging
Group for their invaluable assistance, collaboration, and camaraderie throughout my Ph.D. Their
input on my work and willingness to offer constructive feedback greatly enhanced the quality and
scope of my work —special thanks to Dr. Soyoung Choi, Dr. Jian Li, Yijun Liu, Dr. Takfarinas
Medani, Omar Zamzam, Pavan Magesh, Woojae Jeong, Wenhui Cui and Dr. Haleh Akram. I
deeply appreciate their support, mentorship, and the collaborative spirit they fostered, which made
the lab a place of inspiration, friendship, learning, and personal connection. I also want to thank
the Ming Hsieh Department of Electrical and Computer Engineering staff members for providing
logistical, technical, and administrative support that ensured that my work was always on track
—special thanks to Talyia White, Mayumi Thrasher, Gloria Halfacre, and Theodore Low. I am
extremely grateful for the opportunity to collaborate with such talented and dedicated professionals.
I am equally grateful to Professor Wood’s laboratory members for their insightful contributions
and technical expertise. I sincerely appreciate the team’s enthusiasm and collaborative spirit, which
enriched every interaction. Their collective expertise and support helped me navigate challenging
experiments. My sincere gratitude to Emma Carpenter, Samatha Mejia, Hanna Salcudea, Dr.
Sneha Verma, Botian Xu, Dr. Chau Vu, Dr. Jian Shen, Dr. Matthew T. Borzage, and Dr. Eamon
Doyle for creating a warm atmosphere that sparked curiosity and deepened the joy of learning.
I have appreciated the chance to work alongside such skilled and committed individuals whose
dedication and talent have been truly inspiring. I would also like to express my gratitude to the
dedicated staff at Children’s Hospital Los Angeles for their logistical, technical, and administrative
vi
support, which played a crucial role in ensuring that my work remained organized and on schedule.
Special thanks to Noel Arugay, Theodore Calleja, Silvie Surany, and Obdulio Carreras. From
seamless coordination to prompt assistance, their dedication made a significant difference, allowing
me to focus on what truly matters. Finally, I sincerely thank our esteemed collaborators at
Amsterdam University Medical Center, Dr. Aart Nederveen, Dr. Bart J. Biemond, Dr. Liza
Afzali-Hashemi, and Dr. Koen P. A. Baas.
Additionally, I extend my heartfelt gratitude to the physicians and their staff members who
supported my health and well-being throughout this demanding journey. First, I sincerely appreciate
the expertise and thoughtful advice provided by Dr. Daniel G. Arkfeld during some of the
most challenging moments. His support helped me stay focused and resilient. Thank you to
Dr. Sanddhya Ravikumar for her exceptional care and attention, which ensured that I remained
physically healthy, migraine-free, and able to face the challenges of my work. I am also profoundly
grateful to Dr. Susan M. Axtell for her insightful advice and guidance in managing stress, chronic
pain, and other difficulties. Her support was invaluable in helping me maintain balance during
intense periods. I also want to thank Dr. Vladimir Ayvazyan for his expertise and patience when
addressing my health concerns. His reassurance and practical solutions were greatly appreciated.
Also, I thank Dr. Emma Schiewe for her proactive approach and dedication to helping me maintain
a healthy lifestyle, enabling me to perform at my best. Lastly, I wish to express my sincere gratitude
to Dr Sharae Tejada for her exceptional care and dedication throughout my recovery journey. Her
expertise, encouragement, and personalized guidance have made an incredible difference in my
progress.
To all my friends, both near and far, who have supported and encouraged me throughout this
intellectual voyage. My dissertation represents not only my efforts but also your unwavering love
and belief in my capabilities. I am exceptionally fortunate to have you in my life. I want to extend
special gratitude to Artemis Zavaliangos, Yuni Kay, Brendan Miller, Amanda Rios, Adam Mezher,
Celia Williams, Kimry Jones, Carol Church, Jose Perez, Guadalupe Perez, Maria G. Maldonado,
vii
Jose Luis Maldonado, Omar Flores, Karina Mendoza, Dr. Monica Ledezma, Dr. Ulises Salazar, and
Estela G¨uitron. Your steadfast support and motivational words have provided me with strength
during the most challenging periods of this endeavor.
I would like to take this opportunity to express my sincere gratitude to my dear friends, Katherin
Martin and Carlos Hernandez. Their unwavering support, kindness, and encouragement have been
invaluable throughout my Ph.D, and their faith in me, particularly during periods of self-doubt, has
given me the strength to persevere. I am thankful for their presence in celebrating each achievement,
regardless of its size, and for their guidance during challenging times. Their friendship has been a
source of joy and a reminder that I was never alone in this journey.
I also wish to convey my sincere appreciation to Misael A. Garcia Vasquez. His kindness,
patience, and encouragement have been a constant source of strength throughout this academic
pursuit –—your constant care for my well-being has made a profound difference. Thank you for
being my rock and my cheerleader. You are among the few people who understand how much
listening to live music is healing for me. Thank you for all the amazing concerts, the dancing, and
the singing. Your understanding and sacrifices have meant more than I can put into words, and I
could not have done this without your support.
I would like to express my profound gratitude to my family, whose love, encouragement, and
unwavering support have been the cornerstone of my journey. To my entire family, thank you for
believing in me and being a constant source of strength throughout this process. Special thanks to
my uncle Antonio Gonz´alez-Cisneros, my aunt Ines Zepeda and my cousins Gloria Gonz´alez and
Antonio Gonz´alez for looking after my father while I being away from home.
Also, special acknowledgment to Sofia and Leia Gonzalez for your unconditional love and for
those countless moments of solace, whether through quiet companionship during late-night writing
sessions or playful breaks that reminded me to pause and enjoy the present. You have been a
viii
reminder of resilience and the importance of finding joy in the small moments, and for that, I am
forever grateful.
Also, a big acknowledgment to Diana L. Duarte-Arroy; your kindness and encouragement have
meant so much to me. Your ability to uplift my spirits during stressful times, your care for my
health, and your thoughtful words of advice have been invaluable.
A big shout out to my “little” brother, Abdon Gonz´alez-Suarez; thank you for constantly
reminding me that you are always with me, no matter the distance or circumstances. Thank you
for bringing light and laughter even in the most challenging moments. Your encouragement and
confidence in me have motivated me throughout this academic pursuit. I also want to tell you that,
I love you and am really proud of you, Dr. HH.
Last but not least, the most significant acknowledgment goes to my father, Abdon Gonz´alezCisneros, and your unwavering strength, wisdom, and belief in me have been a foundation throughout
my life, especially during my Ph.D. You promised Mom that you would take good care of me, and
you have not only fulfilled that promise but did a fantastic job.
Through your encouragement –—whether through thoughtful advice, steady reassurance, or your
quiet presence–— you have guided me, not just as a father but also as a mentor. Your support has
shaped my ability to persevere through challenges with resilience and determination. The values
of hard work, integrity, and dedication you instilled in me have been crucial to completing this
dissertation. I am deeply grateful for your sacrifices to ensure that I could pursue my dreams and
for always being a source of inspiration and motivation. Thanks to you, I have become not only a
woman of integrity, but I get to call myself a PhD. This achievement would not have been possible
without your unwavering support and love.
ix
Table of Contents
Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Chapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Chapter 2: Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 Anemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2 Pathophysiology of chronic anemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Thalassemia syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.2 Sickle cell disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 Cerebral hemodynamic response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.1 Cerebral blood flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.2 Cerebrovascular reserve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Watershed white matter areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5 MRI contributions to the characterization of chronic anemia . . . . . . . . . . . . . 20
2.5.1 Contributions from T1 weighted . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5.2 Diffusion MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.5.2.1 Diffusion Tensor Imaging . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.2.2 Interpretation of DTI Measures . . . . . . . . . . . . . . . . . . . . 25
x
2.5.2.3 DTI findings in Chronic Anemia . . . . . . . . . . . . . . . . . . . . 27
2.5.2.4 Diffusion Kurtosis Imaging . . . . . . . . . . . . . . . . . . . . . . . 28
2.5.2.5 Neurite Orientation Dispersion and Density Imaging . . . . . . . . . 31
Chapter 3: Chronic anemia: The effects on the connectivity of white matter . . . 35
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2.1 Participation criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2.2 Laboratory markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.3 Image acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.4 Post-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.2.5 Diffusion modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2.6 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3.1 Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3.2 Laboratory comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.3.3 White matter connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Chapter 4: Unraveling the link between chronic anemia and white matter damage:
A comprehensive diffusion MRI analysis . . . . . . . . . . . . . . . . . . . . . . . . 53
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.2 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.2.1 Participation criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.2.2 Laboratory markers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.2.3 Image acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.2.4 Post-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
xi
4.2.5 Diffusion modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.2.6 Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
4.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.3.1 Demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4.3.2 DTI Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.3.3 DKI results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.3.3.1 Additional Blood Markers . . . . . . . . . . . . . . . . . . . . . . . . 67
4.3.3.2 Processing Speed Index . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.3.4 NODDI results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4.1 WM alterations measured by dMRI models . . . . . . . . . . . . . . . . . . . 72
4.4.1.1 Diffusion tensor in CA . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.4.1.2 Diffusion kurtosis in CA . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.4.1.3 Neurite orientation dispersion and density imaging in CA . . . . . . 75
4.4.2 Mechanisms of WM Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.5 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
Chapter 5: Conclusion of PhD Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Appendix A: Supplementary information for Chapter 4 . . . . . . . . . . . . . . . . . 108
xii
List of Figures
2.1 The shapes of red blood cells for (A) healthy individuals, (B) those with thalassemia,
and (C) those with sickle cell disease are visually represented in this illustration. It
is generally observed that individuals with thalassemia have smaller and paler red
blood cells due to low hemoglobin levels, while individuals with sickle cell disease
may have red blood cells that deform into a crescent moon shape. Such deformation
can lead to blockages in blood vessels. . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 (A) Plot of CBF versus blood oxygen content for control subjects (solid dots), anemia
control subjects (e.g. thalassemia subjects, x symbols), and sickle cell disease (SCD,
open circles). The solid line is the best linear fit to the log transformed data. The
light grey lines are the 95% confidence intervals calculated from a historical cohort of
older nonhemoglobinopathy patients. (B) boxplot and linear correlations of oxygen
delivery. Figures adapted from [1] and [2]. . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3 Axial view of whole brain CBF and oxygen delivery maps. The panels show the
average values for healthy controls (N=25) and SCD patients (N=32). The top colorbar represents absolute perfusion in ml/100g/min, and the bottom oxygen delivery
in milliliter of oxygen/100g brain tissue per minute. Figure modified from [3]. . . . . 15
2.4 The plots display the median and interquartile ranges of CVRCBF in (A) grey and
(B) white matter, for patients with severe and mild SCD, as well as healthy controls.
Lines with asterisks indicate statistically significant differences that remain after
multiple comparison correction (p < 0.017), while diamonds represent statistically
significant differences that did not remain after correction. Figure modified from [4]. 16
xiii
2.5 (A) Damage to tissues caused by infarction exhibited densities ranging from 1% to
18%, with the most prominent concentrations observed in the deep white matter
of the frontal and parietal lobes. (B) Examination conducted on the lowest CBF
(average of < 45 mL/100 g/min) from an independent pediatric SCD cohort and
superimposed onto the infarct density map of the SIT Trial. The outcomes demonstrated
that areas with nadir CBF, situated in the watershed regions, strongly correlate with
the highest infarct density. Figure adapted from [5]. . . . . . . . . . . . . . . . . . . 21
2.6 (A) Illustration of gaussian (upper panel) and non-gaussian (lower panel) diffusion of
water in different environments. (B) This diagram shows the kurtosis shape (yellow
ring) distribution in relation to the diffusion ellipsoid (tensor) model. D1, D2 and
D3 are the eigenvectors (with their respective lambda values). W1, W2 and W3 are
the kurtosis values along principle directions of diffusion ellipsoid. Pictures adapted
from [6]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.7 (A) Cross-sectional electron microscopy image of a white matter bundle. The blue
regions depicts axons while the myelin (for a specif axon) is highlighted with green. In
addition, yellow is used to illustrate CSF and the free water in the brain. (B) Shows a
schematic illustration of the three compartments modeled by NODDI: neurite desity
index (NDI), free water fraction (diso), and the orientation dispersion index (ODI).
The arrows show some of the constrains and assumption to fit NODDI. (C) Diagram
of fibers orientation dispersion where dec,k
is the water diffusivity inside each axon,
dec,k and dec,⊥ denote the local extra-axonal diffusivities, and d’ec,k and d’ec,⊥ are
the apparent extra-axonal diffusivities. Figures adapted from [7, 8]. . . . . . . . . . . 31
3.1 FA analysis in WMS based on the coregistration of the parceled T1-w with dMRI
maps. This coregistration allows mapping the connectivity tracks between two ROIs
to the FA map. The exact process is repeated to create a connectivity matrix for
each subject. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
xiv
3.2 3D-rendering of left and right hemispheres of a representative subject, the white
matter surfaces (WMSs) where FA was controlled for age (log transformed), sex and
group and it was statistically significant in both statistical models (linear regression
and permutation analysis) after FDR correction. The specific regions of interest are
listed on Table 3.2. Top Row: green WMSs, comparison of healthy controls (CTL)
with non-sickle cell anemia (non-SCD). Bottom row: red WMSs, comparison of CTL
with sickle cell anemia (SCD). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.1 FA and KFA: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The
left three columns reveal a mixture of decreased and increased fractional diffusivity
(magma and blue color scales). Very similar patterns are observed in the right three
columns depicting the fractional anisotropy kurtosis when controlling for logarithm
of age and sex. No results were found when hemoglobin was added to the statistical
model. These findings strongly suggest that anemia drives the observed FA’s and
KFA’s decrements|increments in both the SCD and non-SCD cohorts. . . . . . . . . 63
4.2 MD and MK: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The
left three columns reveal an overall increase in mean diffusivity (blue color scale),
while the middle three columns depict an overall decrease in mean kurtosis (magma
color scale) when controlling for logarithm of age and sex. Additionally, the right
three columns highlight statistically significant voxels when the hemoglobin effects
are also regressed out from the MK maps. Notably, the only remaining signal was
found in the MK model of SCD vs CTL, whereas non-SCD patients showed no
statistical difference. Moreover, in both cases, the MD maps showed no statistical
differences when also removing the effects of hemoglobin. These findings strongly
suggest that anemia is the driving factor behind the observed WM changes in the
non-SCD cohort. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
xv
4.3 PSI: Statistically significant correlation (p<0.05) between MK and PSI while controlling
for logarithm of age and sex. Positive correlation coefficients are represented on the
blue scale, and negative correlation is depicted on the magma scale. The left panel
presents the results for SCD patients and CTL individuals, whereas the right panel
displays the results for non-SCD patients along with CTL individuals. . . . . . . . . 67
4.4 NDI: T-maps displaying voxels that were statistically significant (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The three
columns on the left reveal a mixture of decreased (mainly) and increased neural
density index (magma and blue color scales). In addition, when hemoglobin was
added to the statistical model, those voxels with the highest pink intensity (decreased
values) were preserved only in SCD patients. These findings strongly suggest that
anemia drives the observed NDI decrements in both cohorts, but in SCD, more
factors influence the changes in NDI. . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
A.1 RD and RK: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The
left three columns reveal a mixture of decreased and increased radial diffusivity
(blue and magma color scales). However, only smaller cluster of decreased RD are
present in the non-SCD vs CTL comparison. The middle three columns depict an
overall decrease in radial kurtosis when controlling for logarithm of age and sex.
Additionally, the right three columns highlight statistically significant voxels when
the hemoglobin effects are also regressed out from the RK maps. The only remaining
signal was found in the RK model of SCD vs CTL, whereas non-SCD patients showed
no statistical difference. Moreover, in both cases, the RD maps showed no statistical
differences when also removing the effects of hemoglobin. These findings strongly
suggest that anemia drives the observed RD’s increments and RK’s decrements in
the non-SCD cohort. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
xvi
A.2 AD and AK: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows).
The left three columns reveal a mixture of decreased and increased axial diffusivity
(magma and blue color scales), while the middle three columns depict an overall
decrease in axial kurtosis when controlling for logarithm of age and sex. Additionally,
the right three columns highlight statistically significant voxels when the hemoglobin
effects are also regressed out from the AK maps. The only remaining signal was found
in the AK model of SCD vs CTL, whereas non-SCD patients showed no statistical
difference. Moreover, in both cases, the AD maps showed no statistical differences
when also removing the effects of hemoglobin. These findings strongly suggest that
anemia drives the observed AD’s decrements|increments and AK’s decrements in the
non-SCD cohort. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
A.3 T-maps depict voxels exhibiting statistically significant differences (p<0.05) when
comparing MK maps between individuals with Sickle Cell Disease (SCD) and healthy
control (CTL) individuals, utilizing magma color scales. All analyses were adjusted
for the logarithm of age, sex, and hemoglobin (Hb). Additional blood markers were
systematically integrated to account for any residual signal from the Hb model.
Results, incorporating Leukocytes (Leuk) and fetal hemoglobin (Hb F), reveal clusters
with decreased MK closely resembling the Hb model. Moreover, additional blood
markers, including lactose dehydrogenase (LDH) and absolute reticular count (ARC),
elucidate the remaining Hb signal, as demonstrated in the far-left and right two
panels, respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
xvii
A.4 ODI and FISO: T-maps displaying statistically significant voxels (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows) while adjusting
for the logarithm of age and sex. The three left columns of the T-maps show
a mix of decreased and increased orientation dispersion index (ODI) values, with
predominantly increased values in non-SCD patients (represented by the blue color
scale) and decreased values in SCD patients (represented by the magma color scale).
The three right columns show mostly decrements in free water fraction (FISO) values
in both cohorts. However, the extent of the clusters is greater in SCD patients. No
significant results were found when hemoglobin was added to the statistical model. . 112
xviii
List of Tables
2.1 Hemoglobin levels to diagnose anemia. . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Clinical description of chronic anemia. . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 CBF, CMRO2 and CMROglc values across human life span. . . . . . . . . . . . . . . 11
3.1 Demographics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.2 Results for ∗FA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3 Results of FA when controlling for transfusion status and LDH. . . . . . . . . . . . . 44
4.1 Demographics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
xix
Chapter 1
Introduction
“Begin at the beginning,” the King said gravely,
“and go on till you come to the end: then stop.”
— Lewis Carroll, Alice in Wonderland
Worldwide, the prevalence of anemia is very high, affecting around 1.93 billion people and
causing more significant disabilities than asthma, diabetes, and cardiovascular disease combined
[9]. Of this, approximately 802 million suffer from chronic anemia (CA) [10], which is a condition
where the number of erythrocytes or hemoglobin (Hb) concentration is lower than expected and
incapable of providing sufficient oxygenation [11]. The resulting hypoxia is particularly damaging to
the brain because it is one of the organs with higher metabolic demands. Under acute circumstances,
the brain would normally receive preferential blood flow [12]. However, this is not the case in CA,
leading to hypoxia, neuroinflammation, and remodeling of white matter [13].
Chronic anemia is a standard clinical feature seen in patients with hemoglobinopathies [14],
mainly represented by quantitative disorders of Hb synthesis (e.g., thalassemia syndromes) and
qualitative disorders in Hb structure (e.g., sickle cell disease) [15]. Interestingly, these disorders are
becoming more common in countries like the United States and Northern and Western Europe due
to global migration movements [16], [17].
1
Chap.1 Introduction 2
Hemoglobinopathies have been associated with alterations in both gray and white matter
alterations [18, 19, 20, 21]; cerebral vasculopathies [22, 23]; and changes in cerebral blood flow
[24, 4]. These changes make hemoglobinopathies a valuable model for understanding the cerebral
changes caused by CA:
Chapter 2. In this chapter, we review the general pathophysiology of chronic anemia and
its associated features. We also outline the specific characteristics of Thalassemia syndromes
and Sickle Cell disease. Additionally, we will briefly touch upon the cerebral hemodynamic
response to oxygen fluctuations and expand on the concepts of cerebral blood flow, cerebrovascular
reserve, and the vulnerable areas in the WM to their fluctuations. Finally, we outline the
more common mathematical approximations to model the diffusion signal and discuss some
of the findings of the literature on structural and diffusion MRI in chronic anemia.
Chapter 3: The results of this chapter were already published in a special issue on sickle
cell disease in Frontiers in Neurology. In this study, we compared 19 SCD and 15 non-SCD
anemia patients with a wide range of Hb values, allowing the characterization of the effects
of chronic anemia in isolation of sickle Hb. We performed a tensor analysis to quantify FA
changes in WM connectivity in chronic anemic patients. We calculated the volumetric mean
of FA along the pathway of tracks connecting two regions of interest defined by BrainSuite’s
BCI-DNI atlas. In general, we found lower FA values in anemic patients, indicating the
loss of coherence in the main diffusion direction, potentially indicating WM injury. We saw
a positive correlation between FA and hemoglobin in these same regions, suggesting that
decreased WM microstructural integrity FA is highly driven by chronic hypoxia. The only
connection that did not follow this pattern was the connectivity within the left middle-inferior
temporal gyrus. Interestingly, more reductions in FA were observed in non-SCD patients
(mainly along with intrahemispheric WM bundles and watershed areas) than in SCD patients
(mainly interhemispheric).
Chap.1 Introduction 3
Chapter 4: This chapter presents our latest research, where we expanded previous investigations
by analyzing multi-shell acquisitions such as diffusion kurtosis imaging (DKI) and the multicompartment model neurite orientation dispersion and density imaging (NODDI). By doing
so, we sought to determine which changes could be attributed to anemia alone, as opposed
to disease-specific pathophysiology. In particular, we studied 76 patients with SCD, 20
patients with non-SCD anemia, and 32 control subjects. WM integrity was diffusely deranged
across the brains of all anemic subjects, as shown by DTI, DKI, and NODDI measurements.
However, DKI and NODDI models yielded systematically larger effect sizes; changes in SCD
and non-SCD were of comparable magnitude. After controlling for hemoglobin level, patients
with SCD exhibited significant residual derangements not seen in non-SCD subjects. These
remaining abnormalities disappeared after including a marker of hemolysis such as lactate
dehydrogenase (LDH) or reticulocyte count. In summary, these data demonstrate that
chronic anemia due to hemoglobinopathies is associated with diffuse remodeling of white
matter that does not co-localize with regions commonly affected by hypoxia. Based on
our literature review, we propose that demyelination linked to neuroinflammation is the
underlying mechanism responsible for these effects.
This page intentionally left blank
Chapter 2
Background
“I learned very early the difference between knowing
the name of something and knowing something”
— Richard Philips Feynman
2.1 Anemia
Oxygen is essential for normal tissue functioning. Lower levels of oxygen are the direct consequence
of a decreasing number of erythrocytes or Hb which causes widespread health problems and multiple
organ pathologies. A single erythrocyte normally contains around 300 million Hb molecules and
can transport up to 1.2 billion oxygen molecules. On average, if these numbers are reduced below
their 25%, body tissues will become hypoxic, and this condition of decreased oxygenation is also
known as anemia.
Common symptoms of anemia include weakness, lethargy, dizziness, headaches, shortness of
breath, and arrhythmias. [25] Unfortunately, the symptomatic description is non-specific and
insufficient to assess anemia’s true severity [26]. Instead, the degree of anemia is determined
based on the blood hemoglobin concentration. However, it is essential to understand that Hb
values naturally diversify by age, sex, pregnancy status, genetic and environmental factors, and,
5
Chap. 2 Background 6
Table 2.1: Hemoglobin levels to diagnose anemia.1
Anemia
Population Non-Anemia Mild Moderate Severe
Children (0.5-4.9 years) ≥ 11 10-10.9 7-9.9 < 7
Children (5.0-11.9 years) ≥ 11.5 11-11.4 8-10.9 < 8
Children (12-14.9 years) ≥ 12 11-11.9 8-10.9 < 8
Nonpregnant women (≥15 years) ≥ 12 11-11.9 8-10.9 < 8
Pregnant women ≥ 11 11-10.9 7-9.9 < 7
Men (≥15 years) ≥ 13 11-12.9 8-10.9 < 8
1 Hemoglobin levels for diagnosing anemia at sea level in grams per deciliter (g/dL).
Adjustments recommended for altitude and smoking behaviors prior to applying
cutoffs [11, 27].
potentially, race [11]. Therefore, the World Health Organization (WHO) has specified Hb parameters
for various groups and levels of severity (see Table 2.1) [27]. Additionally, some studies have
shown the need for additional parameters for race [28, 29] due to intrinsic genetic variations and
predispositions to some forms of chronic anemia [30, 31].
Anemia has been shown to be a public health problem that affects not only low but also high
income countries [10]. According to the National Heart, Lung, and Blood Institute, it affects more
than 3 million Americans, creating around 2.8 million visits to physician offices, 623,000 visits to
emergency departments and 5,987 deaths per year [32].
The pathophysiology of anemia is diverse and often multifactorial, e.g., genetic mutations in
hemoglobin genes, altered erythrocyte morphology, acute and chronic blood loss, vitamin deficiencies,
and iron deficiency [9, 33]. It is often classified based on the biological mechanism of causation:
deficient erythropoiesis (faulty or decreased production of erythrocytes) or uncontrolled loss of
erythrocytes (excessive destruction of erythrocytes or blood loss) [11, 34]. Anemia can also be
classified based on the erythrocyte morphology, e.g., microcytic (smaller erythrocyte size than
normal), normocytic (normal looking erythrocytes), and macrocytic (bigger erythrocyte size) anemia
[25, 26, 35].
Chap. 2 Background 7
Figure 2.1: The shapes of red blood cells for (A) healthy individuals, (B) those with thalassemia, and
(C) those with sickle cell disease are visually represented in this illustration. It is generally observed
that individuals with thalassemia have smaller and paler red blood cells due to low hemoglobin
levels, while individuals with sickle cell disease may have red blood cells that deform into a crescent
moon shape. Such deformation can lead to blockages in blood vessels.
2.2 Pathophysiology of chronic anemia
Hemoglobinopathies are hereditary conditions of abnormal hemoglobin [36], and they are characterized by the presence of different degrees (mild, moderate, or severe) of chronic anemia [37]. To briefly
describe their nature, it is crucial to outline that adult Hb is a protein that can transport four oxygen
molecules and consists mainly of a heterotetramer made up of two α-globins, two β-globins and a
heme portion [38]. Healthy adults also have a small percentage of fetal hemoglobin F (HbF ∼ 1%),
much of which during the first year of life is replaced by hemoglobin A (HbA ∼ 97%) and minor
components of hemoglobin A2 (HbA2 ∼ 2%) [39].
Hemoglobinopathies can result from [36] :
• Qualitative abnormalities of Hb: Production of abnormal hemoglobin S, C, O or E. Sickle
cell disease being the most common example.
• Quantitative abnormalities: Production of normal Hb but decreased amount of globin chains
such as thalassemia syndromes.
• Failure to switch globin chain synthesis after birth, i.e., hereditary persistency of hemoglobin
F (Hb F).
Chap. 2 Background 8
Table 2.2: Clinical description of chronic anemia.1
Disease Hb (g/dL) Type of Anemia
α− Thalassemia
Silent carrier normal no apparent illness
Trait normal mild HMA
Major 8-10 moderate HMA
β− Thalassemia
Trait low-normal mild HMA
Intermedia 6-10 moderate HMA
Major < 7 severe HMA
Sickle cell hemoglobinopathies
Trait (Hb AS) normal no apparent illness
Hb S/HPFH 11-14 mild HA with mild sickling
Hb SC 10-12 HA with moderate sickling
β
+-Thal (Hb Sβ
+) 9-12 HA with mild-moderate sickling
β
0
-Thal (Hb Sβ
0
) 7-9 HA with severe sickling
Disease (Hb SS) 6-8 HA with severe sickling
1 Clinical description of anemia and Hb ranges for SCD and
thalassemia syndromes. Adapted from [15, 48].
Consequently, despite the anemia burden, the underlying pathophysiology of hemoglobinopathies
can be quite different [15]. For example, the normal lifespan of an erythrocyte is roughly 115 days
[40], whereas, in thalassemia patients, it is close to 30 days [41], and approximately ten days
in SCD patients [42, 43, 44]. Hemolysis (i.e., erythrocyte destruction) itself carries additional
burdens because free Hb and free heme are liberated into the plasma, manifesting direct cytotoxic,
inflammatory, and pro-oxidant effects that, in turn, negatively impact endothelial function. This
creates further vascular damage that underlies the chronic organ damage observed in hemoglobinopathies pathology [45, 46, 47].
2.2.1 Thalassemia syndromes
With respect to thalassemia, patients do not have enough hemoglobin A (HbA) because there is a
disproportion in the globin chains: their cells cannot generate either α-globin (alpha-thalassemia) or
β-globin (beta-thalassemia). Thalassemia syndromes are primarily caused by faulty erythropoiesis.
Consequently, the erythrocytes are smaller and have less Hb than normal, giving them a pale
appearance (see Fig. 2.1).
Chap. 2 Background 9
This is the reason thalassemias are classified as hypochromic microcytic anemias (HMA) [49]. In
moderate to severe cases, thalassemia can also lead to more rapid erythrocyte destruction [15].
The clinical description of these two forms of thalassemia is based on the genotype, and the
degree of anemia varies accordingly (see Table 2.2). Roughly, α-thalassemia can be categorized in
three groups: silent carriers (no clinical symptoms), α-thalassemia trait (generating mild anemia)
and α-thalassemia major (patients have persistent anemia and considerable health problems, in
some cases patients might benefit from blood transfusion). In all of these cases some degree of
HMA is present, even in the silent carriers, where individuals have no significant health problems.
Similarly, in the case of β-thalassemia, there are also three main groups that present HMA: βthalassemia trait (generating mild anemia), β-thalassemia intermedia (patients experiene significant
health problems and might require intermittent blood transfusions), and β-thalassemia major
(severe form of disease where patients require lifelong blood transfusions) [15].
Some of the common secondary complications of thalassemias that can affect the brain include
iron overload from blood transfusions, enlargement of the spleen (that can actually worsen anemia),
and the creation of venous and arterial blood clots (specially β-thalassemia major or intermedia
patients can develop a chronic hypercoagulable state) [50]. Therefore, novel therapeutic approaches
are currently in development to target some of the main and consequential effects of thalassemias:
i. Correction of the globin chain disproportion: Bone marrow transplantation offers the
possibility of restoring tissue’s ability of producing functional hemoglobin and, thus, regulate
the globin chain imbalance. Better results have been obtained in patients that had had a
sibling donor before the development of iron overload and resulting organ injuries [51]. To
restore Hb synthesis in thalassemia other techniques have been tested like gene addition and
gene editing, resulting in Lentiglobin being approved in Europe in 2019 [52, 53]. Researchers
are currently working on gene-editing mechanisms to also reactivate Hb F because it has
shown some level of protection against hypoxia [54, 55].
Chap. 2 Background 10
ii. Targeting ineffective erythropoiesis: Transforming growth factor β superfamily ligands
that play diverse roles in embryonic development and adult tissue homeostasis such as growth
differentiation factors and activins, negatively impact late-stage erythropoiesis in thalasemia
patients [54]. However, luspatercept medication (recently FDA approved) is an activin
receptor ligand trap that improves chronic anemia by creating a rapid and sustain increase
of erythrocytes in disorders caused by ineffective erythropoiesis [56, 57].
iii. Modulating iron metabolism: Hepcidin is the main hormone that regulates iron metabolism
in the body [58]. However, thalassemia patients have low quantities of hepcidin despite
the presence of iron overload [54]. Consequently, in preclinical studies a number of agents
that restore hepcidin levels, e.g., hepcidin analogs, minihepcidins, and inhibitors of hepcidin
repressors, have shown beneficial effects [59, 60]. Similarly, by targeting ferroportin (a
cellular transmembrane protein that transports iron from inside-out), and its interactions
with hepcidin, the iron’s negative effects can be reduced [60].
2.2.2 Sickle cell disease
The term “sickle cell disease” indicates all forms and manifestations of abnormal hemoglobin S (Hb
S), from the least severe case called sickle cell trait (Hb AS) to the most critical form, sickle cell
disease with homozygous occurrence (Hb SS). This also includes a range of mixed heterozygous
cases such as, hereditary persistence of fetal hemoglobin (Hb S/HPFH), appearance of hemoglobin
C (Hb SC), and Hb S combined with some forms of β-thalassemia (see Table 2.2) [15, 48, 47, 61, 62].
In SCD, intracellular polymerization of the abnormal Hb S molecule stretches the round erythrocytes into a sickle form (see Fig. 2.1), leading to the agglutination and occlusion of vessels and
arteries [62]. This causes abnormal perfusion, which leads to tissue ischemia, infarction in every
organ or even stroke [63]. The erythrocytes then partially unsickle when re-oxygenation occurs [64].
This process continues until sickling is irreversible, and the erythrocytes undergo hemolysis [65].
Consequently, the two hallmarks of the disease are the vaso-occlusion and the chronic hemolytic
Chap. 2 Background 11
Table 2.3: CBF, CMRO2 and CMROglc values across human life span.
Population CBF CMRO2 CMROglc
mL/(100g·min) mL/(100g·min) µmol/(100g·min)
Newborn infants 6-35 1.3 4-19
Kids (3-10 years) 60-140 4.3-6.2 49-65
Young adults 46-57 3-3.5 25-31
Average PET values of cerebral blood flow and energy metabolism in
healthy individuals. Adapted from [66, 67].
anemia (HA) that emerge in patients even during their first year of life [15, 47].
Depending on the form, SCD can be a dangerous hemoglobinopathy, resulting not only in the
long term developmental disorders, but in episodic acute clinical events, such as extreme pain crises
(also referred as vaso-occlusive sickle crises) [47], high susceptibility to various infections, spleen
crises, acute thoracic syndrome, and increased risk of stroke.
In the last 50 years, immense efforts have been made to comprehend the complex pathophysiology
of SCD; however, effective treatments have been developed at a slower rate [47, 68]. Prior to 2010
the only true successful treatments were blood transfusions to prevent stroke, the use of hydroxyurea
medication for preventing painful episodes, and prophylactic penicillin for blocking pneumococcal
disease in children [61, 68]. Nevertheless, the increased understanding of the pathophysiology of
SCD has led to four promising pharmacological interventions [47]:
i. Modifying the genotype: using allogeneic bone marrow transplants (i.e., stem cell transplant
coming from a donor) [69] or autologous hematopoietic stem cell transplant modifications (i.e.,
genetically modified patient’s own stem cells created by gene editing or gene therapies) [70].
ii. Reducing Hb S polymerization: Hb F has been known to have a major protective effect
in SCD because it not only dilutes the intracellular Hb S concentration, but also inhibits
sickling. Hydroxyurea medication induces fetal hemoglobin synthesis and has been proven
to have high efficacy and safety, particularly in the pediatric population (even though the
Chap. 2 Background 12
increase in Hb F is uneven and not equally present in all red blood cells). One of the newlyapproved anti-sickling agents, OxbrytaTM (Voxelotor/GBT440), increases the oxygen affinity
of the hemoglobin molecule to reduce Hb S polimerization [71].
iii. Targeting vasocclusion: Over the last 40 years, L-glutamine has been investigated as
a possible medication to decrease vasocclusion because increasing L-glutamine appears to
decrease endothelial adhesion of erythrocytes, but its mechanisms are still not fully understood
[72, 73]. Despite the concerns of an increased risk of mortality in more compromised patients,
it was approved by the FDA in 2017 but not by the European Medicine Agency [73, 74].
The only other FDA approved medication for vasocclusion is Crizanlizumab. Its mechanism
also blocks the adhesion of activated erythrocytes, neutrophils and platelets. Patients treated
with crizanlizumab reported significantly lower rates of sickle cell–related pain crises and
fewer adverse episodes [75].
iv. Decreasing inflammation: Different elements play a critical role in creating and sustaining
chronic inflammation that further harms organs, including white blood cells (i.e., immune
system response), endothelium, and coagulation factors. Different therapeutic approaches
have been proposed to reduce the inflammation in SCD patients that, include common
anticoagulants such as aspirin (with no major benefit over placebo) [76], to newer and sophisticated methods including Canakinumab (a monoclonal antibody that targets inflammatory
cytokines, thus, mitigating some of the inflammation in SCD) [77], and N-acetylcysteine
(commonly used in respiratory conditions proved to inhibit dense cell formation and reduced
oxidative stress) that have been through clinical trials but not yet FDA approved [78].
.
Chap. 2 Background 13
2.3 Cerebral hemodynamic response
2.3.1 Cerebral blood flow
The brain is one of the most active organs in the body and has major metabolic requirements of
oxygen and adenosine triphosphate (ATP). In a healthy adults, on average, the brain accounts for
2% of the total body weight, but it consumes close to 20% of the total resting O2 [66]. Interestingly,
in kids younger than ten years of age brain’s O2 consumption is close to 50% [79].
Additionally, the brain uses ATP obtained from extracellular compounds with a high-energy content,
such as glucose. However, storage of energy and oxygen in the brain is minimal. Consequently, to
preserve its functional and structural integrity, the brain is highly dependent on a continuous supply
of these compounds from the bloodstream, making it highly sensitive to minor disturbances, e.g.,
cerebral blood flow is interrupted completely during cardiac arrest. If blood flow is not restored
within 10 seconds, consciousness will be lost, possibly leading to white matter damage [66].
The brain is very sensitive to CBF fluctuations and can not tolerate any major drop in its
perfusion (see Table 2.3), e.g., in a healthy individual if CBF is reduced to half its normal rate
it will cause loss of consciousness. Nevertheless, there are physiological mechanisms to sustain
adequate CBF even when arterial pressure falls in times of stress for example, autoregulation.
Cerebral autoregulation is a homeostatic process that regulates CBF across a range of blood
pressures. At least four mechanisms are proposed for autoregulation [80]:
i. Myogenic: It is the ability of the vascular smooth muscle to constrict or dilate in response
to changes in transmural pressure [81]. For example, myogenic tone is produced when
an arteriole and small artery smooth muscle contract or relax in response to increased or
decreased pressure, respectively.
ii. Neurogenic: This mechanism manages cerebral vasoreactivity and mediates the diameter of
small to medium vessels. Neurons and some types of glia (namely astrocytes and microglia)
produce several neurotransmitters with vasoactive effects. Examples of vasodilators are
Chap. 2 Background 14
Figure 2.2: (A) Plot of CBF versus blood oxygen content for control subjects (solid dots), anemia
control subjects (e.g. thalassemia subjects, x symbols), and sickle cell disease (SCD, open circles).
The solid line is the best linear fit to the log transformed data. The light grey lines are the 95%
confidence intervals calculated from a historical cohort of older nonhemoglobinopathy patients. (B)
boxplot and linear correlations of oxygen delivery. Figures adapted from [1] and [2].
acetylcholine and nitric oxide whereas vasoconstrictors are serotonin and neuropeptide-Y
[82].
iii. Metabolic: This process that subserves autoregulation occurs in the microvasculature, and it
is subject to continuous changes in the local environment. Metabolic adjustments contribute
the most to regulation of the smaller vessels. For example, carbon dioxide (CO2) can increase
vasomotor responses (1 mmHg increment in partial CO2 corresponds to a ∼4% rise in CBF)
[83]. Similarly, an increment in the gradient of hydrogen ions (H+) will lead to vasodilation;
however, this mechanism has to be tightly regulated as it can cause acidification of the
environment leading to neuronal damage, e.g., in brain ischemia the effects of carbonic
anhydrase inhibitors can be protective as they seem to reduce H+ and thus maintain pH
homeostasis [84].
iv. Endothelial: The endothelium is a monolayer of specialized cells that forms the inner lining of blood vessels and controls several processes, such as the regulation of vascular tone,
inflammatory response, thrombosis, adhesion, and vascular permeability [85]. In regards
to CBF, the endothelium generates several vasoactive contents that are vasodilators (nitric
Chap. 2 Background 15
Figure 2.3: Axial view of whole brain CBF and oxygen delivery maps. The panels show the average
values for healthy controls (N=25) and SCD patients (N=32). The top color-bar represents absolute
perfusion in ml/100g/min, and the bottom oxygen delivery in milliliter of oxygen/100g brain tissue
per minute. Figure modified from [3].
oxide, endothelium-derived hyperpolarizing factor), vasoconstrictors (endothelin-1) or have
both effects (eicosanoids, a family of bioactive signaling lipids) [86].
In general, CBF is measured as the volume of blood supplied to a defined mass of tissue per
unit time, usually measured in mL/100g · min−1
[66]. According to positron emission tomography
(PET) studies in the human brain, the average CBF is approximately 57 mL/100g tissue/min (see
Table 2.3). Therefore, for an average brain of 1400g, the entire CBF will be closer to 640-800
ml/min, that is, approximately 15-20% of the total basal cardiac output [79, 66, 67].
It is also important to note that the grey matter is more active than the white, which explains
why there is a strong correlation between CBF and the total volume of grey matter [24, 87]. During
childhood, the ratio of grey to white matter is highest between the ages of 6 and 10 [88], but this
ratio gradually decreases as the brain undergoes synaptic pruning until early adulthood.
Chap. 2 Background 16
Figure 2.4: The plots display the median and interquartile ranges of CVRCBF in (A) grey and
(B) white matter, for patients with severe and mild SCD, as well as healthy controls. Lines
with asterisks indicate statistically significant differences that remain after multiple comparison
correction (p < 0.017), while diamonds represent statistically significant differences that did not
remain after correction. Figure modified from [4].
This is the reason children appear to have greater CBF than adults, peaking at around 6.5 years
of age (see Table 2.3) [1].
Knowing the CBF value, the oxygen delivered to the brain can be quantified as:
02 delivery = CBF · 02 content (ml/100g/min) (2.1)
where the oxygen content in the blood depends linearly of the hemoglobin value,
02 content = 1.34 · Hb · Ya + 0.003 · pO2 (mL O2/mL blood) (2.2)
and, Ya is the arterial saturation measured by pulse oximetry.
These equations demonstrate that hemoglobin levels are inversely correlated with CBF. In fact,
this behavior has been shown among healthy subjects (see Fig. 2.2-A). This figure further illustrates
that, in cases of chronic reduction in oxygen capacity, anemic patients exhibit an increase in CBF.
Chap. 2 Background 17
On a whole organ basis, this rise in CBF serves as a compensatory mechanism to maintain
overall resting oxygen delivery throughout the brain (see Fig. 2.2-B) [1, 2, 24]. Studies have shown
that individuals with SCD experience consistent oxygen supply to the grey matter, demonstrating
the inverse correlation between CBF and the blood’s ability to transport oxygen (see Fig. 2.3) [3].
This preserved oxygen delivery correlates with the absence of GM atrophy typically observed in
these patients.
Unlike other areas, compensatory vasodilation has not been observed in white matter, which
implies that the delivery of oxygen in WM is directly linked to hemoglobin concentration [3].
Fig. 2.3 shows that, despite the CBF increase in SCD patients, regions of deep WM do not receive
as much oxygen as those in healthy individuals. This reduction in WM oxygen delivery may be
due to accelerated geometric and morphological changes or the destruction of the microvascular
network within the WM, but the exact cause is still unknown. The reduction may increase with
age.
2.3.2 Cerebrovascular reserve
Cerebrovascular reserve (CVR), which is the brain’s ability to increase blood flow in response to
stress, i.e. periods of reduced oxygen [1], is used as a marker of microvascular health in various
conditions. CVR is measured as the change in CBF between baseline and the response to a
vasodilatory stimulus [89], such as intravenous pharmacological agents (e.g., acetazolamide (ACZ))
or respiratory challenges (e.g., carbon dioxide):
CV RCBF =
CBFpostACZ − CBFpreACZ
CBFpreACZ
100%. (2.3)
As previously stated, patients with anemia exhibit higher resting CBF to compensate for the lack
of oxygen and ensure adequate supply to the brain (see Fig 2.2.). However, this CBF increment can
significantly impede the body’s ability to respond to other hemodynamic stressors, such as sudden
changes in hemoglobin levels or low oxygen levels during sleep. This is the primary consequence of
Chap. 2 Background 18
the arterioles and capillaries are already maximally dilated. [90, 91].
According to previous research, patients with SCD have their highest CBF at around 150
ml/100g/min [92]. This means that, if patients with CBF of 120 ml/100g/min experience reduced
oxygen levels, they would only be able to increase their oxygen delivery by 25%. In comparison,
healthy people with resting CBF of 60 ml/100g/min could increase their oxygen delivery by 250%.
This discrepancy underlines the vulnerability of anemic patients to ischemic strokes, which can be
caused by elevated resting flow and reduced vascular reserves [1].
The severity of anemia plays a crucial role in determining an individual’s ability to cope with
stress. For example, Afzafali et. al. (2021) used time-encoded arterial spin labeling (ASL) scans
to quantify CBF in three groups: healthy individuals, severe, and mild SCD patients both before
and after using ACZ. This study revealed that, in patients with severe SCD, there was a significant
decrease in CVRCBF specifically in the GM but not in the WM regions, when compared to the
control group. Furthermore, among severe SCD patients, GM CVRCBF was lower compared to
those with mild SCD. It is worth noting that, while the statistical significance did not persist after
applying multiple comparison corrections, there is a clear pattern indicating decreased CVRCBF
in both the WM and the GM depending on the severity of anemia.
2.4 Watershed white matter areas
Cerebral blood flow is particularly important because any compromise, such as during a stroke or
a period of reduced oxygenation, causes great damage to the watershed (WS) areas of the brain.
This is due to their location between the major arteries.
There are two distinct WS areas in the brain: the cortical WS area, which lies between the cortical
territories of the anterior cerebral (ACA), middle cerebral (MCA), and posterior cerebral arteries
(PCA); the second area is the internal WS territory, which includes the white matter found slightly
above the lateral ventricle, from the deep to the superficial arterial systems of the MCA, and
Chap. 2 Background 19
between the MCA’s superficial system and the ACA’s [93].
In the context of SCD and thalassemia, two main types of strokes can occur: hemorrhagic and
ischemic. Hemorrhagic strokes result from bleeding in the brain while ischemic strokes are caused
by reduced blood flow to specific regions, which leads to hypoxia, and can additionally cause an
infarction, i.e., a complete blood blockage resulting in necrosis or cellular death [94].
Even in the absence of noticeable focal neurological symptoms, most of the cerebral MRI hyperintensities found in patients correspond with the presence of infarctions or ischemia, and they are
collectively named silent cerebral infarctions (SCIs) [95, 96].
In patients with SCD, SCIs can become evident as early as the sixth month of life. The
underlying mechanisms involve various factors associated with the polymerization of Hb S, including
the adhesion of sickled blood cells to leukocytes and the endothelium, activation of the endothelium,
hemolysis (the breakdown of red blood cells), inflammation, and hypercoagulation [97, 98]. Further
studies have shown that SCI prevalence increases over time, affecting approximately 25% of children
by the age of 6, 39% by age 18, and 53% by young adulthood [99]. Notably, no reports have
suggested a plateau in the prevalence of SCI as individuals progress through different stages of life.
Additionally, in thalassemia, there is a pathological tendency for blood clot formation, leading to
a hypercoagulable state [100, 101]. This state is considered a significant contributing factor in
the development of SCIs. Other factors, including the severity of the disease itself, iron chelation
therapy, the presence of associated comorbidities (e.g., hypertension, small vessel disease), and
individual variations [100, 102].
The vulnerability of watershed areas in the brain to SCIs can vary based on an individual’s
specific arterial anatomy. However, studies have indicated that deep white matter WS areas are
particularly susceptible. It has been reported that, in SCD patients, up to 90% of SCIs occur
in these regions, which comprise only 5.6% of the total brain volume [5]. Similar deep white
matter disarrangements have been reported in thalassemia patients [101, 103]. This highlights the
Chap. 2 Background 20
disproportionate impact of SCIs, and their prevalence also indicates a higher risk for overt stroke.
Additional risks factors for SCIs development include events associated with damage to the
cerebral blood vessels, such as high blood pressure and both intracranial and extracranial stenosis
[104], as well as, impaired cerebral oxygen delivery due to low baseline hemoglobin levels and acute
anemic events. In fact, a comprehensive study published by Ford et. al. (2018) examined 286
SCD children from the Silent Infarct Transfusion (SIT) Trial [105], and employed the expertise
of neuroradiology and neurology committees to meticulously outline and coregister each infarct to
a brain atlas. This process allowed for the creation of an infarct density map (see Fig. 2.5-A).
Interestingly, the infarct densities ranged from 1% to 18%, with the greatest densities falling in the
deep white matter of the frontal and parietal lobes.
Furthermore, this study aligned CBF maps (acquired with pseudo-continuous arterial spin
labeling) from an independent pediatric SCD cohort to the SIT Trial infarct density map. Fig.
2.5-B) shows the lowest CBF map (average <45 mL/100 g/min) overlaid onto the infarct density
map. This analysis reveals a significant overlap between regions of nadir CBF (located in the
watershed areas) and the highest infarct density, suggesting a strong correlation between reduced
blood flow and SCI occurrence [5].
2.5 MRI contributions to the characterization of chronic anemia
Advancements in in-vivo MRI research have provided significant breakthroughs in understanding
the mechanisms underlying neurological and neurocognitive impairments in patients with chronic
anemia. Qualitative MRI has revealed that conditions like overt strokes, SCIs, and vasculopathies
only represent a fraction of the brain abnormalities affecting these patients [106].
Quantitative techniques, however eliminate subjectivity through statistical inference, making
them more reliable in measuring parameters, such as brain volume, microstructural integrity,
structural and functional connectivity, perfusion, and oxygenation. Currently, many quantitative
Chap. 2 Background 21
Figure 2.5: (A) Damage to tissues caused by infarction exhibited densities ranging from 1% to
18%, with the most prominent concentrations observed in the deep white matter of the frontal and
parietal lobes. (B) Examination conducted on the lowest CBF (average of < 45 mL/100 g/min)
from an independent pediatric SCD cohort and superimposed onto the infarct density map of the
SIT Trial. The outcomes demonstrated that areas with nadir CBF, situated in the watershed
regions, strongly correlate with the highest infarct density. Figure adapted from [5].
techniques are still in the early stages of development. However, ongoing research aims to understand
the specificity, validity, and reproducibility of various imaging parameters [107]. As research
progresses, quantitative MRI is poised to provide deeper insights into the neurological and cognitive
complications of anemia patients, paving the way for more effective diagnostic and therapeutic
methods.
2.5.1 Contributions from T1 weighted
One sequence of MRI, T1-weighted (T1w), is a highly effective method used to accurately identify
anatomical locations and independently analyze a range of parameters, including gray and white
matter volume, subcortical volumes, cortical thickness, and cortical curvature. For gray matter,
Chap. 2 Background 22
previous research has shown that older children and adolescents with SCD may have thinning of
the cortex [108, 109] and decreased cerebral volume, either total or regional [110, 111], they also
have differences in many subcortical structures [112]. However, little is know about the effect of
thalassemias in gray matter.
When it comes to white matter, studies focusing on SCD patients (over nine years of age) consistently
show lower volume both, total and regional. These reductions have been observed in critical regions
of the brain, such as the corpus callosum, frontal, parietal, and temporal lobes, as well as the brain
stem and cerebellum [110, 111, 113]. Interestingly, the occipital region appeared to be relatively
unaffected [110]. Similar findings of reduced volume have been shown in beta thalassemia patients
[21].
In particular, our laboratory conducted a novel study using T1w to carefully analyze a group of
patients suffering from anemia, including those with thalassemia and SCD. Our findings revealed
that the volume of white matter in deep watershed areas was significantly reduced, and this decrease
was directly correlated with the severity of the anemia, regardless of the intrinsic nature of both
diseases. [103]
2.5.2 Diffusion MRI
Another MRI sequence used to study the brain microstructure is diffusion MRI (dMRI). This
technique uses magnetic resonance to detect contrasts based on the movement of water molecules
at the microscopic level [114]. The diffusion of water in living organisms is restricted by structures
such as cell membranes, making it possible to differentiate between tissues with varying cell densities
and those containing only fluid. Therefore, dMRI has the potential for inferring microstructural
properties of the brain tissue, such as its microstructural organization, complexity, and in the case
of WM, the degree of myelination [115].
In addition, dMRI’s motion sensitivity is direction-dependent, enabling it to identify the orientations
of fiber bundles, such as the WM fasciculi. Furthermore, it details the diffusion signal in the
Chap. 2 Background 23
WM and provides valuable information, such as intra-axonal volume (fiber density), areas with
crossing fibers and estimation of axon diameters. As a result, dMRI provides a unique insight
into microstructural details in-vivo and facilitates the quantitative assessment of both normal and
pathological tissue characteristics [116].
The diffusion signal can be modeled with multiple mathematical approximations that, in general,
can be divided into two main groups: those that focus on the signal representations and those that
provide biophysical models [7, 117]. In this work, we will focus on the three frameworks most
commonly used in clinical settings, which we used to study chronic anemia. The first two are
categorized as signal representation models and include the diffusion tensor imaging (DTI) and the
diffusion kurtosis imaging (DKI) frameworks. In particular, DTI estimates water movements in the
brain using a Gaussian approximation [118].
Among all the dMRI methods, it is the most frequently used, and it represents the anisotropy
and directionality of the water, assuming a single ellipse in a voxel [119]. On the other hand, DKI
takes into account the water’s non-Gaussian distribution [120, 121]. It can describe more complex
brain microstructures, including crossing fibers [122]. The third model, a relatively new biophysical
model, is called neurite orientation dispersion and density imaging (NODDI) [123, 8]. It is a threecompartment model representing neurite density, isotropic free water, and neurite dispersion (axons
and dendrites), providing more biophysical information on the WM structures.
2.5.2.1 Diffusion Tensor Imaging
The DTI model, proposed by Basser et al. in 1994 [118], characterizes the dMRI signal using a
second-order diffusion tensor, S(n, b). When employing Einstein’s summation convention, this DTI
model can be represented as:
S(n, b) = S0 e
−bninjDij (2.4)
Chap. 2 Background 24
where n = [n1, n2, n3] represents the direction of the diffusion gradient, the b coefficient represents
the diffusion weighting, and D is the diffusion tensor that encodes the information of the water’s
movement in the brain [122].
We can then write Eq. 2.4 in terms of D:
log
S(n, b)
S0
= −b ni nj Dij , (2.5)
in this form Dij is solved as a set of six linear equations. Then, D can be decomposed into its three
eigenvectors (e1, e2, and e3), and the corresponding eigenvalues (λ1, λ2, and λ3) [119]. Finally,
these eigenvalues van be used to calculate metrics that are independent of the gradient direction
that was applied initially, namely [124, 125]:
i. Fractional anisotropy (FA), which indicates the relatively unhindered diffusion of water
molecules in a specific direction, describing the diffusion directionality within an imaging
voxel. FA is unitless and its values range from 0 (perfect isotropic diffusion, the tensor will
be spherical) to 1 (complete anisotropic diffusion, tensor will be elongated with ellipsoidal
shape). Mathematically, FA is computed as:
F A =
s
1
2
(λ1 − λ2)
2 + (λ1 − λ3)
2 + (λ2 − λ3)
2
λ
2
1 + λ
2
2 + λ
2
3
, (2.6)
ii. Mean diffusivity (MD), which is a measure of the degree of diffusion and computed as the
mean of the tensor’s eigenvalues:
MD =
λ1 + λ2 + λ3
3
, (2.7)
making MD independent of direction.
Chap. 2 Background 25
iii. Axial diffusivity (AD), which is often referred as parallel diffusivity because AD describes
the diffusion rate along the primary axis (or eigenvalue) of diffusion:
AD = λ1 , (2.8)
this is considered the water moving parallel to the axons.
iv. Radial diffusivity (RD), which reflects the perpendicular diffusivity, i.e., the average
diffusivity along the other two minor axes:
RD =
λ2 + λ3
2
. (2.9)
It is important to note that MD, AD, RD are measured in measured in mm2/s.
2.5.2.2 Interpretation of DTI Measures
DTI has been widely employed across the medical field, owing to its remarkable ability to discern
subtle differences in the microstructural composition of cellular membranes through water diffusion.
Specifically, water movement is directly impacted by changes in the space between membrane layers:
an increase|decrease in space leads to elevated|reduced diffusivity.
This characteristic has made DTI an invaluable technique for detecting variations in tissue composition
at the microscopic level. It is critical, however, to carefully evaluate changes in all the tensor
metrics. For example, conditions that affect WM often lead to decreased anisotropy, i.e., decreased
FA and increased MD caused by increments in RD, AD, or both. Consequently, a comprehensive
evaluation and understanding of diffusion tensor changes could help to differentiate between stages
of a particular disease and elucidate differences between neurological diseases.
The diffusion tensor has shown diverse associations with the WM changes observed in histopathological studies and mouse models in neurological conditions that are relatively well understood.
These associations can be relevant to the understanding of chronic anemia:
Chap. 2 Background 26
1. Demyelination: refers to the loss of myelin while maintaining the relative integrity of axons.
This condition arises from diseases that inflict damage upon myelin sheaths or the cells
responsible for their formation [126]. In the central nervous system demyelinating diseases
can be categorized based on their pathogenesis into distinct groups: those resulting from
inflammatory processes, viral origins, acquired metabolic abnormalities, hypoxic-ischemic
conditions, and demyelination induced by focal compression. Interestingly, many demyelination
studies with disease-related processes have found less diffusion anisotropy when axons are less
myelinated.
(a) Inflammatory demyelination: One of the most common diseases that falls into this
category is multiple sclerosis. It is thought to be caused by the interaction of multiple
genetic and environmental factors, and it is described as an autoimmune disorder where
the myelin is mistaken for a foreign body and targeted by immune cells. The immune
system produces excess amounts of inflammatory markers called cytokines that in turn
damage the myelin sheath and lead to demyelination [127]. In regards to the DTI metrics,
typically increased MD and decreased FA in the normal appearing WM (NAWM) of
different brain regions, including the corpus callosum, have been detected in MS [128,
129, 130]. Additionally, in a 2-year longitudinal study, decreased FA and increased
RD mostly in the CC were observed [131]. In contrast, no changes in diffusivity were
observed in the NAWM of MS over 2–4 years [132, 133].
(b) Hypoxic–ischaemic demyelination: In extreme cases, when exposed to hypoxia or
ischemia, brain tissue tends to undergo necrosis rather than demyelination. However,
the myelinating oligodendrocytes may bear the primary impact of hypoxic or ischemic
damage under certain conditions. Examples of this are seen in patients with severe
cerebral small vessel disease (CSVD) [126]. Its effects are observed in arterioles, capillaries,
and small veins that provide blood supply to the white matter and deep structures within
the brain [134]. CSVD is a chronic and progressive vascular condition, and various studies
using different DTI approaches have revealed a reduction in FA along with an increase
in MD in the normal appearing white matter of these patients [135, 136, 137].
Chap. 2 Background 27
2. Dysmyelination: The term describes the presence of malformed and faulty myelin sheaths,
which is different from the destruction of previously healthy myelin seen in demyelination
[138]. These disorders are also known as leukodystrophies, with metachromatic leukodystrophy
(MLD) being the most common example. They comprise a group of genetic disorders characterized by the accumulation of lipids (fatty substances such as oils and waxes) and other
storage compounds within cells situated in the white matter of the central nervous system
and peripheral nerves. This accumulation causes disruption in the growth and formation of
the myelin sheath [139]. A DTI study in MLD patients by Van Rappaard et. al. (2018) [140],
showed that before receiving any treatment, patients throughout the WM had decreased
FA while MD and RD increased. These findings were also observed in the corpus callosum
(affected early in the disease) and even included the NAWM. Interestingly, AD was decreased
in a large part of the WM, including the corpus callosum. The authors concluded that since
both animal and human studies have shown a correlation between increased RD and myelin
loss, their results of increased RD in WM also suggest myelin loss in MLD patients that goes
in line with those of histopathological findings [141].
2.5.2.3 DTI findings in Chronic Anemia
Using DTI metrics few research groups have studied chronic anemia. In general, SCD patients
have exhibited altered metrics compared with healthy controls. Especially, reduced fractional
anisotropy (FA, loss of the main diffusion directionality of water) has been reported in various
WM regions [142] including, the corpus callosum (CC) [20, 143, 144, 145, 146], the internal capsule
[143], the corticospinal tract [144], and along the deeper white matter [143, 145, 146]. In contrast,
increased mean diffusivity (MD, average diffusion of water) was found across frontal and parietal
lobes [20, 146, 147], CC [20, 146, 147], and subcortical structures [20, 146, 147], and higher radial
diffusivity (RD, magnitude of water diffusing perpendicular to WM) in the corticospinal tract
[20, 147], anterior thalamic radiations [147], the internal capsule and frontal and parietal lobes
[20, 147].
Chap. 2 Background 28
Nevertheless, no changes in axial diffusivity (AD, magnitude of water movement parallel to
WM) [20, 147], even though higher AD has been correlated with lower Hb [148]. In the case of
non-SCD anemias, to our knowledge (at this point in time) there is one single report looking at
DTI metrics when comparing β-thalassemia patients with CTL individuals that did not find any
statistical differences between them [23].
2.5.2.4 Diffusion Kurtosis Imaging
Kurtosis is a statistical measure used in probability to determine how far a distribution differs from
a Gaussian distribution (normal bell-shaped curve) [6]. In the brain, the intrinsic tissue structure
creates barriers and compartments that restrict water movements, leading to non-Gaussian diffusion
of water. DKI is an attempt to capture this non-Gaussian distribution of water and has been used
as a marker of tissue heterogeneity [149], Fig. 2.6. In particular, DKI uses diffusion-weighted MRI
to estimate diffusional kurtosis with standard pulse sequences and modest acquisition times that
can be acquired in clinical MRI scanners [120].
Similar to DTI that depends on the difussion tensor D (see Eq. 2.4), the degree of nonGaussianity is convenionally quantified by the diffusional kurtosis tensor W. Particularly, the DKI
model can be written in terms of both D (units on the order of 10−3mm2/s) and W (unitless) as
cumulants of the DWI signal up to second order in b, Fig. 2.6. Therefore, using a similar notation
as Eq. 2.5, DKI is express as [121]:
log
S(n, b)
S0
= −b ni nj Dij +
b
2 D
2
6
ni nj nk nl Wijkl , (2.10)
where b is the diffusion weighting (b-value), ni denotes the i-th component measured in direction ˆn,
Dij is the i, j element of the D tensor (symmetric, with rank 2), and Wijkl is the i, j, k, l element
of the W tensor (symmetric, with rank 4).
The DKI model in Eq. 2.10 can also be solved by a set of linear equations to obtain: the
six independent parameters of D and fifteen independent parameters of W [121, 122]. Some of
Chap. 2 Background 29
Figure 2.6: (A) Illustration of gaussian (upper panel) and non-gaussian (lower panel) diffusion
of water in different environments. (B) This diagram shows the kurtosis shape (yellow ring)
distribution in relation to the diffusion ellipsoid (tensor) model. D1, D2 and D3 are the eigenvectors
(with their respective lambda values). W1, W2 and W3 are the kurtosis values along principle
directions of diffusion ellipsoid. Pictures adapted from [6].
the intrinsic differences with the typical DTI protocol is that In addition to 15 different gradient
directions to resolve the anisotropic information of W, the DKI model requires at least three bvalues (these can include signals for b=0 in addition to two non-zero b-values).
Moreover, like DTI, measures invariant to rotation can be established based on the kurtosis
tensor [122]:
i. Fractional anisotropy kurosis (KFA), which similarly to how FA quantifies the anisotropy
of the diffusion tensor, KFA quantifies lower to higher kurtosis tensor anisotropy in a range
between 0 and 1, and it is quantified as:
KF A =
||W − MKT I
(4)||F
||W||F
, (2.11)
Chap. 2 Background 30
where the Frobenius norm is identified as ||...||F , I
(4) is the isotropic tensor (fully symmetric
of rank 4), and MKT is the mean kurtosis tensor defined by
MKT =
1
5
T r(W)
=
1
5
(W1111 + W2222 + 2W3333 + 2W1122 + 2W2233) .
(2.12)
ii. Mean kurosis (MK), which is characterized as the mean of directional kurtosis coefficients
across all spatial directions. It is formulated through the surface integral:
MK =
1
4π
Z
dΩnK(n) . (2.13)
Sampling from the fitted diffusion and kurtosis tensors yields the directional kurtosis, K(n),
for a specific direction n as:
K(n) = MD2
D(n)
2
ni nj nk nlWijkl , (2.14)
and
D(n) = ni nj Dij . (2.15)
iii. Radial kurosis (RK), which is the mean of directional kurtosis across all directions perpendicular to the primary direction of fibers, aligning with the main direction of the diffusion
tensor, e1:
RK =
1
2π
Z
dΩθ K(θ) δ(θ · e1) . (2.16)
iv. Axial kurosis (AK), which represents the directional kurtosis along the primary direction
of highly aligned structures.:
AK = K(e1) . (2.17)
Chap. 2 Background 31
Figure 2.7: (A) Cross-sectional electron microscopy image of a white matter bundle. The blue
regions depicts axons while the myelin (for a specif axon) is highlighted with green. In addition,
yellow is used to illustrate CSF and the free water in the brain. (B) Shows a schematic illustration
of the three compartments modeled by NODDI: neurite desity index (NDI), free water fraction
(diso), and the orientation dispersion index (ODI). The arrows show some of the constrains and
assumption to fit NODDI. (C) Diagram of fibers orientation dispersion where dec,k
is the water
diffusivity inside each axon, dec,k and dec,⊥ denote the local extra-axonal diffusivities, and d’ec,k
and d’ec,⊥ are the apparent extra-axonal diffusivities. Figures adapted from [7, 8].
2.5.2.5 Neurite Orientation Dispersion and Density Imaging
As discussed before, DTI and DKI are signal representation models designed to improve sensitivity
to target pathologies, with the caveat that these frameworks come with no obvious biological
interpretation. Conversely, biophysical models like NODDI [123] offer better explanations for
observed abnormalities in tissue properties by increasing the specificity of their parameters, making
them clinically meaningful [8].
NODDI is a three-compartment model designed to distinguished between: intra-neurite (ic
compartment), extra-neurite(ec compart) and CSF (iso compartment) environments. Each of these
micro-structural compartments affect water diffusion and they generate an independent normalized
MR signal (see Fig. 2.7). The complete expression for the normalized signal A can be formulated
as:
A = (1 − νiso) (νicAic + (1 − νic)Aec) + νisoAiso , (2.18)
Chap. 2 Background 32
where Aic, Aec and Aiso are the normalized signal for each compartment. Similarly, νic, νic, and
νic represent the volume fraction per compartment.
In particular, each compartmental tissue and their respective dMRI signal can be further described
as follows [123]:
i. Intra-neurite model, which refers to the area confined only by the membrane of neurites.
This compartment represents axons and possibly dendritic processes, whose area is depicted as
a collection of sticks to capture the significantly constrained diffusion across neurites and the
unhindered diffusion along them. Such sticks are essentially cylinders of zero radius (diffusion
in the perpendicular direction is negligible, dic,⊥ ≈ 0) [150], whose orientation distribution
can range from highly parallel to highly dispersed. The mathematical expression for Aic, the
normalized signal for the intra-cellualr compartment is [151]:
Aic =
Z
s
2
e
−b dic,k (q·n)
2
f(n) dn , (2.19)
where f(n) dn gives the probability of finding sticks along orientation n, b is the b-value and
q is the gradient direction of the DWI signal. Additionally, e
−b dic,k (q·n)
2
measures the signal
attenuation caused by unhindered diffusion along a stick with intrinsic diffusivity dic,k
(i.e.,
parallel diffusivity intra-neurite compartment) and orientation n.
ii. Extra-neurite model, which in the white matter denotes the compartmental area surrounding
neurites (but not including them), occupied by different types of glial cells (microglial,
astrocytes and oligodendrocytes), in the gray matter denotes the cell bodies (somas), and
other structures within the brain such as ependymal cells, extra-cellular matrices, and vascular
structures [8]. In this compartment, the diffusion of water is impeded by neurites but not
entirely restricted, hence it is represented using a simple Gaussian diffusion model. The
equation for Aec is:
log Aec = −b q
T
Z
s
2
f(n) Dec(n) dn
q , (2.20)
Chap. 2 Background 33
where n is the principal diffusion direction that characterizes the cylindrical symmetric tensor
Dec(n) (extra-cellular diffusion) that comprises the diffusion coefficients dec,k
(ec-parallel
diffusion) and dec,⊥ (ec-perpendicular diffusion) to n [151].
iii. CSF compartment, which refers to the free water compartment and is represented by the
isotropic diffusion:
Aiso = e
(−b diso)
, (2.21)
where diso =3 µm2/ms.
In NODDI, the orientation distribution function is restricted to an axially symmetric distribution
defined by a singular parameter called the Watson distribution. Moreover, to ensure stable estimation,
the subsequent constraints are enforced (see Fig. 2.7):
• The ec- and ic- parallel diffusion are equal:
dic,k = dec,k = 1.7 µm2
/ms , (2.22)
the choice of this fixed value was based on the minimization of the fitting residuals for voxels
in the corpus callosum made by Alexander et al., in 2010 [152].
• The extracellular space’s tortuosity indicates the resistance encountered by the diffusion
process due to a geometrically complex medium, as opposed to an obstacle-free environment.
Based on the tortuosity model as proposed by Szafer et al., 1995 [153]:
dec,⊥ = (1 − νic) dec,k
. (2.23)
In order to analyze the microstructure within a voxel, NODDI relies on three scalar parameters
[7, 8, 151]:
Neurite density index (NDI), which is also referred as intra-neurite fraction (νic), quantifies
the packing density of axons or dendrites
Chap. 2 Background 34
Orientation dispersion index (ODI), which assesses the orientation coherence of neurites. It
ranges from 0, that will indicate perfectly aligned straight fibers, to 1 indicates complete
isotropy.
Free water fraction (diso), which assesses the orientation coherence of neurites. It is also
denoted in the literature as FWF, or isotropic fraction (FISO).
Chapter 3
Chronic anemia: The effects on the
connectivity of white matter
“The brain is a world consisting of a number of
unexplored continents and great stretches of
unknown territory.”
— Santiago Ram´on y Cajal
3.1 Introduction
Chronic anemia (CA) is a condition in which the number of erythrocytes or hemoglobin (Hb)
concentration is lower than expected and incapable of meeting physiological needs [11]. Worldwide,
the prevalence of anemia is very high, affecting around 1.93 billion people and causing more
significant disability than asthma, diabetes, and cardiovascular disease combined [9]. Tissue oxygen
consumption is heterogeneous and organ-specific. The brain is one of the organs with higher
metabolic demand that receives preferential blood flow under acute circumstances [12]. As a result,
neurons are specifically sensitive to hypoxia [154]. CA causes reduced oxygenation in the brain,
leading to hypoxia, neuroinflammation, and white matter (WM) remodeling [13].
35
Chap. 3 Chronic anemia and the connectivity of WM 36
CA is a standard clinical feature seen in patients with hemoglobinopathies [14], mainly represented
by qualitative disorders in Hb structure (e.g., sickle cell disease, SCD) and quantitative disorders
of Hb synthesis (e.g., thalassemia syndromes, non-SCD) [15]. Hemoglobinopathies have been
associated with gray matter (GM) [18, 19, 112, 155] and WM alterations [20, 103, 147, 156, 148, 157],
cerebral vasculopathies [22, 23, 158], and changes in cerebral blood flow [4, 24, 159, 160, 161, 162],
thus serving as a model for the cerebral changes caused by CA [103, 157].
Studying SCD and non-SCD patients with a wide range of Hb values and genetic predisposition
to anemia simultaneously allows the characterization of the effects of CA in isolation from sickle
Hb [24, 157]. In this context, novel structural magnetic resonance imaging (MRI) research done
in our laboratory, where SCD and non-SCD patients were simultaneously analyzed, demonstrated
that WM volume was diffusely lower in deep, watershed areas proportional to anemia severity
regardless of Hb genotype [103]. This relationship between anemia and WM volume was confirmed
in a repeated analysis with a restricted population consisting of patients with beta-thalassemia [21].
We hypothesized that CA causes similar damaging effects and changes in structural connectivity
of WM in patients with hemoglobinopathies. Furthermore, the damage is driven by hyperemia and
not by the intrinsic pathophysiology of these hemoglobinopathies. To characterize alterations in the
WM, we performed a diffusion tensor imaging (DTI) analysis and quantified the average fractional
anisotropy (FA, overall directionality of water diffusion) along the pathways connecting every pair
of regions of interest (ROIs) defined by an anatomical atlas. This approach covers all the WM
bundles in the brain and not merely the main association fascicles [163].
Diffusion MRI (dMRI) is used to study in-vivo WM microstructure and allows quantitative
characterization of healthy and diseased tissue. The most widely used dMRI technique is DTI,
despite various limitations. Its derived metrics like FA are potential biomarkers of brain abnormalities
in patients with neurodegenerative diseases. For example, DTI analysis already provided evidence
of the relationship between WM microstructure and markers of anemia severity, such as oxygen
Chap. 3 Chronic anemia and the connectivity of WM 37
saturation level and Hb value [147]. Additionally, SCD patients have shown lower FA in the
corticospinal tract and cerebellum [147], across the internal capsule [156], the corpus callosum
[156, 148, 144, 146], and in the deep WM [146]. Decrement of FA was also observed in major
WM tracts in CA patients, regardless of the anemia subtype, and correlated significantly with the
neurocognitive decline observed in the CA population [144].
3.2 Materials and methods
3.2.1 Participation criteria
All participants in this study were part of a larger project on Sickle Cell Disease at Children’s
Hospital Los Angeles that its Institutional Review Board approved. Each participant was recruited
with informed consent. We collected MRI data and blood tests on patients, including non-SCD,
SCD, and healthy controls matched by sex and age. The accepted age range was between 10
to 50 years old. Eligibility criteria included patients with SCD diagnosis (Hb SS, Hb SC, Hb
Sβ0, and Hb Sβ
+ genotypes), patients with chronic anemia diagnosis (beta-thalassemia major,
beta-thalassemia intermedia, hemoglobin H-constant spring, congenital dyserythropoietic anemia,
spherocytosis anemia, and autoimmune hemolytic anemia) and healthy controls (mainly recruited
from family members of SCD patients to match race and ethnicity between groups).
The exclusion criteria disqualified those patients with previous overt stroke, acute chest syndrome,
pain crisis hospitalization (within one month), and pregnant candidates. Similarly, individuals with
prior history of neurologic insults, developmental delay, or chronic medical conditions that require
regular medical care or medications and pregnant candidates would be ineligible as healthy controls.
We followed the standard guidelines and regulations for MRI safety and exclusion criteria. On
the same day, each participant completed the MRI examination without sedating medications, and
we collected vital signs and blood samples.
Chap. 3 Chronic anemia and the connectivity of WM 38
3.2.2 Laboratory markers
To account for the similarities and differences in CA pathophysiology, all participants enrolled
in our study underwent a thorough examination of their blood samples. Complete blood count,
reticulocyte total, and quantitative Hb electrophoresis percentages of Hb S, Hb A, Hb A2, hemoglobin
F, etc.) were analyzed in the clinical laboratory. Additional surrogates for hemolysis like lactate
dehydrogenase (LDH) and plasma-free Hb levels were also quantified
3.2.3 Image acquisition
The MRI data were acquired on a 3T Philips Achieva scanner using an 8-channel head coil for each
participant. The structural 3D T1-weighted (T1-w) sequence specification was TR/TE = 8.3/3.8
ms, SENSE = 2, and isotropic voxel size of 1 mm3
. In addition, a single-shell dMRI sequence
was acquired with TR/TE = 6,700/86 ms; isotropic voxel size of 2.5 mm3
; 30 diffusion-encoding
directions at b-value = 1,000 s/mm2 and one b-value = 0 s/mm2 using a single-shot echo-planar
imaging sequence.
3.2.4 Post-processing
The Brain extraction and parcellation from the T1-weighted (T1-w) images were processed with
BrainSuite (https://brainsuite.org, v19b). Specifically, BrainSuite’s Cortical Surface Extraction
(CSE) tool was used to perform skull stripping [164], tissue classification, including partial volume
fraction of voxels identified as WM, GM, and CSF [165], topological corrections [166], and delineation
of the inner/outer cortex. In addition, BrainSuite’s Surface Volume Registration (SVReg) tool
[167, 168] performed anatomical co-registration to the BCI-DNI anatomical atlas [168], and brain
segmentation.
The dMRI data were corrected for localized geometric distortions to enable accurate multimodal analysis. Each subject’s motion and eddy current-induced distortions were corrected with
FSL’s eddy module [169, 170]. Using BrainSuite’s Diffusion Pipeline (BDP), we registered the dMRI
Chap. 3 Chronic anemia and the connectivity of WM 39
to the T1-w data, followed by susceptibility-distortion correction based on the inverse contrast
normalization [171].
3.2.5 Diffusion modeling
Using the well-known tensor equation, we calculated the fractional anisotropy maps in BDP. To
render more accurate tractography in the WM, we also computed in BDP the orientation density
functions (ODFs). In particular, the ensemble average propagator response function optimized
(ERFO) uses machine learning and linear estimation theory to optimize ODF accuracy for arbitrary
q-space sampling schemes. It has shown advantages over other methods [172]. Furthermore, ERFO
can model single-shell (and multi-shell) data and has the capacity of rendering crossing fibers with
the most negligible false positives [173].
Whole-brain deterministic fiber tracking, based on quantitative anisotropy [174], and visualization
were performed with the DSI Studio tractography package (http://dsi-studio.labsolver.org).
The sections of tracks entering cortical GM or subcortical regions were excluded to avoid partial
volume effects. Afterward, detail connectivity analysis of fiber bundles connecting two ROIs
(previously labeled on the T1-w structural images) was implemented with the TractConnect Matlab
package (https://neuroimage.usc.edu/neuro/Resources/TractConnect). Specifically, TractConnect uses filtered tracks connecting two ROIs to define a volumetric white matter surface
(WMS) and projects it into the FA maps (Figure 3.1). Then FA values within WMSs were averaged.
For each individual, the average FA values were used as the elements of the connectivity
matrix. 88 ROIs from the BCI-DNI atlas were used: 66 cortical regions, 14 subcortical, corpora
quadrigemina, mammillary bodies, brainstem, and cerebellum.
Overall, this modeling method offers higher sensitivity and specificity to detect not only regional
differences in WM microstructure (like voxel-wise analysis would do) but along the connectivity
pathways, and it is robust to some of the commonly criticized featured of DTI [175, 176]: the
Chap. 3 Chronic anemia and the connectivity of WM 40
inability to render crossing fibers and to define connectivity between ROIs by streamlining counting.
The first was overcome by using ERFO to model diffusion ODFs and the latter by defining the
WMSs and characterizing these “connections” with the mean FA value.
3.2.6 Statistical analysis
Figure 3.1: FA analysis in WMS
based on the coregistration of the
parceled T1-w with dMRI maps.
This coregistration allows mapping
the connectivity tracks between two
ROIs to the FA map. The exact
process is repeated to create a
connectivity matrix for each subject.
For each element of the connectivity matrix (upper
triangular part), Figure 3.1, the FA differences between
groups were modeled using multiple linear regression analysis
after controlling for logarithm of age (log-age), sex, and
group. The logarithm of age was used because brain
maturational effects are nonlinear with age in adolescents
and young adults [177, 178]. Finally, the results were also
corrected for multiple comparisons using the False Discovery
Rate (FDR) to adjust the correspondent p-values [179] with
a 20% acceptance rate.
A similar analysis was performed using a permutation
analysis using Manly’s method [180, 181], which was also
FDR corrected [182] with the same threshold. Given the
complexity of the data, it was not possible to guarantee all
the assumptions of linear modeling. Consequently, we also
chose to model the WMSs using nonparametric permutation
analysis. Overlapping between the two methods provided an
additional confidence level in our results.
Chap. 3 Chronic anemia and the connectivity of WM 41
For completeness, we tested the possible contribution of changes in FA caused by monthly
transfusions and LDH values in patients. For this, we only ran a multiple linear regression analysis
controlling also for log-age, sex, and group. All the statistical analyses were calculated using the
R statistical package [183].
3.3 Results
3.3.1 Demographics
In this analysis we considered 19 clinically asymptomatic SCD patients (age = 22.4 ± 7.8 years;
Hb = 10.1 ± 2.1 g/dL; F = 9 patients), 15 non-SCD anemic patients (age = 22.4 ± 4.8 years; Hb
= 10 ± 2.8 g/dL; F = 8 patients) and 23 control subjects (age = 21.3 ± 6 years; Hb = 13.3 ±
1.2 g/dL; F = 14 individuals). The age range for all the participants was 11.2 to 35.8 years. All
demographics are reported in Table 3.1.
The breakdown of the race (and ethnicity) for control subjects was 17 African-American (nonHispanic) and 5 White (Hispanic) individuals. SCD patients included 17 African-American (nonHispanic) and 2 White (Hispanic) patients. The non-SCD group consisted of 8 Asian (nonHispanic), 5 White (non-Hispanic), and 2 White (Hispanic) patients.
For the SCD group, the genotypes were 12 Hb SS and 7 Hb SC patients. Because of the specific
matches between control and SCD, 9 of the control subjects were identified with sickle cell trait
having hemoglobin AS (Hb AS). Previous work in our laboratory suggests that the Hb AS subtype
does not alter normal cerebral blood flow (CBF) regulation and balance of oxygen supply and
demand Field (22), making Hb AS carriers good candidates for control subjects.
The specific anemias in the non-SCD group consisted of 7 patients with beta-thalassemia major,
3 beta-thalassemia intermedia, 2 hemoglobin H-constant spring, 1 congenital dyserythropoietic
anemia, 1 spherocytosis anemia and 1 autoimmune hemolytic anemia.
Chap. 3 Chronic anemia and the connectivity of WM 42
Table 3.1: Demographics.1
CTL non-SCD SCD
CTL
vs.
nonSCD†
CTL
vs.
SCD†
nonSCD
vs.
SCD†
N 23 15 19 - - -
Age 21.3±5.9 22.4±4.8 21.8±8.3 0.88 0.97 0.96
Sex (F:M) 14:9 8:7 10:9 - - -
Transfused 0 9 5 - - -
Hemoglobin (g/dL) 13.2±1.2 10.3±1.7 10.2±2.1 ≤0.01 ≤0.01 0.96
Hematocrit (%) 39.8±3.6 32.2±5.82 28.7±5.3 ≤0.01 ≤0.01 0.10
WBC count (x103
) 5.6±1.7 7.0±3.1 9.3±4.3 0.33 ≤0.01 0.11
Reticulocytes (%) 1.2±0.5 3.1±3.1 7.8±3.5 0.07 ≤0.01 ≤0.01
Plasma-free Hb 6.7±5.2 22.3±23.8 21.8±19.6 ≤0.02 0.02 0.99
LDH 519.5±75.2 632.2±361.5 1008.5±573.7 0.64 ≤0.01 ≤0.01
Abs. neutrophil count 3.2±1.6 4.2±1.8 5.4±3.6 0.41 ≤0.01 0.37
Heart rate (min−1
) 70.5±20.4 79.0±13.4 81.7±13.9 0.28 0.08 0.88
SBP (mmHg) 116.1±9.2 113.9±9.1 115.5±12.6 0.79 0.98 0.89
DBP (mmHg) 66.1±9.5 60.3±9.3 63.7±7.3 0.13 0.65 0.51
O2 Sat. (%) 99.5±0.9 98.3±2.9 97.8±1.6 0.17 ≤0.01 0.66
Hemoglobin A (%) 82.2±17.9 90.7±7.9 23.4±32.4 - - -
Hemoglobin F (%) 0.7±2.4 2.9±4.1 8.8±9.5 - - -
Hemoglobin S (%) 14.1±18.1 0.0±0.0 50.9±26.0 - - -
1 Mean and standard deviation of demographic information and selected blood count measurements.
† Group comparison using one-way analysis of variance (ANOVA) result with Tukey-Kramer test
for multiple comparisons. Statistical significant values (p≤0.05) are color-coded as follows: green
color denoted comparison between CTL and non-SCD, red compares CTL and SCD, and blue
color non-SCD with SCD patients.
CTL: control group; non-SCD: non-sickle cell disease group; SCD: sickle cell disease group; LDH:
lactate dehydrogenase; SBP: systolic blood pressure; DBP: dystolic blood pressure.
Of the CA patients, 8 non-SCD (7 beta-thalassemia major and 1 congenital dyserythropoietic
anemia) and 5 SCD (Hb SS patients) were on monthly transfusions. The rest of the non-transfused
SS patients were prescribed hydroxyurea and had a mean hemoglobin F fraction of 18%. One patient
with SC was also taking hydroxyurea. At Children’s Hospital Los Angeles, it is recommended to
treat all pediatric patients, of all SCD genotypes, nine months and older with hydroxyurea unless
they have been placed on chronic transfusion [184, 185], as indicated by NIH guidelines [186].
Furthermore, as of 2,000, all SCD patients at our facility have received access to the transcranial
Doppler screening [184, 185, 186, 187].
Chap. 3 Chronic anemia and the connectivity of WM 43
Table 3.2: Results for ∗FA.1
ROI-1 ROI-2 CTL
∗FA
CA
∗FA r
†
∗F A,Hg T†
-statistic r
∗F A,Hg p
∗F A,Hg
CTL vs. non-SCD
R. caude
nucleus
R. middle
frontal gyrus 0.37 -0.57 -0.46 t(36)=3.1
p≤0.01 0.29 ≤0.01
R. thalamus R. middle
frontal gyrus 0.49 -0.70 -0.59 t(30)=3.9
p≤0.01 0.41 ≤0.01
R. thalamus R. amygdala 0.47 -0.65 -0.55 t(34)=3.8
p≤0.01 0.51 ≤0.01
L. thalamus L. gyrus rectus 0.40 -0.63 -0.50 t(34)=3.4
p≤0.01 0.37 ≤0.01
L. thalamus L. parahippocampal gyrus 0.40 -0.56 -0.48 t(34)=3.2
p≤0.01 0.32 0.06
R. superior
frontal gyrus
R. cingulate
gyrus
0.46 -0.73 -0.58 t(34)=4.2
p≤0.01 0.41 ≤0.01
R. transverse
frontal gyrus
R. subcallosal
gyrus
0.49 -0.68 -0.59 t(31)=4.0
p≤0.01 0.37 0.03
R. cingulate
gyrus
L. cingulate
gyrus
0.52 -0.72 -0.61 t(34)=4.5
p≤0.01 0.53 ≤0.01
L. cingulate
gyrus
L. pre-cuneus 0.45 -0.82 -0.61 t(32)=4.3
p≤0.01 0.39 0.02
L. middle
temporal gyrus
L. inferior
temporal gyrus -0.40 0.62 -0.50 t(34)=3.3
p≤0.01 -0.25 ≤0.01
CTL vs. SCD
L. thalamus L. parahippocampal gyrus 0.50 -0.55 -0.54 t(38)=3.9
p≤0.01 0.46 ≤0.01
R. gyrus rectus
L. middle
orbito-frontal
gyrus
0.63 -0.73 -0.68 t(37)=5.6
p≤0.01 0.35 0.03
R. middle
orbito-frontal
gyrus
L. middle
orbito-frontal
gyrus
0.48 -0.59 -0.54 t(36)=3.8
p≤0.01 0.28 0.09
L. middle
orbito-frontal
gyrus
R. subcallosal
gyrus
0.54 -0.63 -0.58 t(39)=4.5
p≤0.01 0.42 ≤0.01
L. middle
orbito-frontal
gyrus
L. subcallosal
gyrus
0.58 -0.71 -0.64 t(38)=5.1
p≤0.01 0.27 0.10
1 Connectivity between ROI-1 and ROI-2 that was statistically significant after FDR correction in
multilinear and permutation models controlling for the group, sex, and age (log-transformed).
The upper and lower sections of the table show the statistics for FA when comparing healthy
controls (CTL) with non-SCD and SCD patients. No connections were statistically significant
when comparing non-SCD with SCD patients.
∗ Mean group FA controlled for sex and age (log-transformed) along the volumetric white matter
surface connecting these ROIs. Standarized (unitless) values are shown.
† Point-biseral correleation coefficients and results of the unpared two-samples t-test on the ∗FA
values between groups.
Pearson correlation of ∗FA with Hb and the correspondent p-value is also displayed.
Chap. 3 Chronic anemia and the connectivity of WM 44
Table 3.3: Results of FA when controlling for transfusion status and LDH.1
ROI-1 ROI-2 T-statistic
Transfusion♠ LDH♦
CTL vs. non-SCD
R. caudate nucleus R. middle frontal gyrus - -
R. thalamus R. middle frontal gyrus t(30)=3.0 p≤0.01 t(30)=3.4 p≤0.01
R. thalamus R. amygdala - t(34)=3.5 p≤0.01
L. thalamus L. gyrus rectus - -
L. thalamus L. parahippocampal
gyrus
- -
R. superior frontal gyrus R. cingulate gyrus t(34)=2.9 p≤0.01 t(34)=3.5 p≤0.01
R. transverse frontal
gyrus
R. subcallosal gyrus - t(31)=4.1 p≤0.01
R. cingulate gyrus L. cingulate gyrus t(34)=2.7 p≤0.01 t(34)=4.0 p≤0.01
L. cingulate gyrus L. pre-cuneus t(32)=2.3 p≤0.01 t(32)=3.8 p≤0.01
L. middle temporal gyrus L. inferior temporal gyrus - t(34)=-3.2 p≤0.01
CTL vs. SCD
L. thalamus L. parahippocampal
gyrus
t(38)=3.9 p≤0.01 -
R. gyrus rectus L. middle orbito-frontal
gyrus
t(37)=5.6 p≤0.01 t(37)=4.1 p≤0.01
R. middle orbito-frontal
gyrus
L. middle orbito-frontal
gyrus
t(36)=3.8 p≤0.01 t(36)=2.8 p≤0.01
L. middle orbito-frontal
gyrus
R. subcallosal gyrus t(39)=4.5 p≤0.01 t(39)=3.2 p≤0.01
L. middle orbito-frontal
gyrus
L. subcallosal gyrus t(38)=5.1 p≤0.01 t(38)=4.1 p≤0.01
1 Connectivity between ROI-1 and ROI-2 that was statistically significant after FDR correction in
the multilinear model. For consistency same ROIs are displayed as in Table 3.2. The upper and
lower sections of the table show the statistics for FA when comparing healthy controls (CTL) with
non-SCD and SCD patients. Similar to Table 3.2, no conections were statistically significant when
comparing non-SCD with SCD patients.
♠ Unpaired two-sample t-test on FA values controlled for the group, sex, age (log-transformed), and
transfusion status.
♦ Unpaired two-sample t-test on FA values controlled for the group, sex, age (log-transformed), and
LDH values.
3.3.2 Laboratory comparisons
Laboratory and clinical markers are shown in Table 3.1. Hemoglobin (p = 0.96) and hematocrit
(p = 0.10) levels were not statistically different between CA groups, but both had statistically
significant lower values compared with healthy control (non-SCD, SCD: p ≤ 0.01). Furthermore,
the SCD population showed significantly higher levels of reticulocytes (p ≤ 0.01) and LDH (p
≤ 0.01) compared to both non-SCD and control, suggesting increased intravascular hemolysis.
However, plasma-free Hb was not different between CA types (p = 0.99). SCD patients had mildly
Chap. 3 Chronic anemia and the connectivity of WM 45
increased white cell counts with respect to control subjects (p ≤ 0.01) but not relative to the nonSCD anemic patients (p = 0.11). In the case of Hb electrophoresis, Hb S was highest for SCD
patients. Still, our control also exhibited a smaller percentage of Hb S because of the inclusion of
sickle trait subjects. SCD patients demonstrated the highest hemoglobin F (Hb F) concentration,
with intermediate levels observed in non-SCD patients. Most control subjects had no Hb F, but
one subject had 11.7% Hb F.
3.3.3 White matter connectivity
Overall, no statistically significant WMSs were found when comparing SCD with non-SCD patients.
10 WMSs in CTL vs. non-SCD and 5 WMSs in CTL vs. SCD analysis showed significant differences
in both multiple linear regression and permutation analysis (Table 3.2). ∗FA indicates the group
mean FA after controlling for log-age and sex using the multiple linear regression. The point-biserial
correlation coefficient, r∗F A,Gr, shows the direction and strength of the relationship between ∗FA
and the status of being anemic or not. A negative r∗F A,Gr value depicts higher ∗FA in control
individuals than CA patients. All the WMSs reported in Table 3.2 showed this behavior except for
one (left middle and inferior temporal gyrus) in healthy controls vs. non-SCD comparison. The
∗FA unpaired two-samples t-test statistic is also displayed for completeness, which agrees with the
multiple linear regression analysis.
When Hb was included in the mathematical models as a covariate, all the ∗FA differences
reported in Table 3.2 were no longer statistically significant. This suggests that many of the effects
reported are driven by the Hb differences between healthy controls and CA patients. To further
study the relationship between ∗FA and Hb, we calculated the Pearson correlation coefficient,
r∗F A,Hb, and their respective p-value, p∗F A,Hb for the WMSs reported in Table 3.2. ∗FA was
significantly correlated with Hb levels in 8 out of the 10 WMSs in the population consisting of
control and non-SCD analysis and in 3 out of the 5 WMSs in the control and SCD population.
Consequently, by calculating r2
∗F A,Hb, the proportion of variance in ∗FA explained by Hb, we
observed that in control and non-SCD for the WMSs reported in Table 3.2, Hb accounts for up to
Chap. 3 Chronic anemia and the connectivity of WM 46
Figure 3.2: 3D-rendering of left and right hemispheres of a representative subject, the white matter
surfaces (WMSs) where FA was controlled for age (log transformed), sex and group and it was
statistically significant in both statistical models (linear regression and permutation analysis) after
FDR correction. The specific regions of interest are listed on Table 3.2. Top Row: green WMSs,
comparison of healthy controls (CTL) with non-sickle cell anemia (non-SCD). Bottom row: red
WMSs, comparison of CTL with sickle cell anemia (SCD).
26% (right thalamus and right amygdala) of the variance in ∗FA, and up to 21% (left thalamus and
left parahippocampal gyrus) in the case of controls with SCD patients.
The spatial locations of the WMSs listed in Table 3.2 are 3D-rendered in the left and right
hemispheres of a representative subject (Figure 3.2). In these same WMSs, we saw a positive
correlation of ∗FA with Hb. Significant results were bilateral and generally symmetrical across
hemispheres. Interestingly, more WMSs survived for the non-SCD (mainly intrahemispheric and
along with watershed areas) than for the SCD (mainly interhemispheric) group compared with
healthy controls.
Chap. 3 Chronic anemia and the connectivity of WM 47
Table 3.3 shows the results of the multiple linear regression when adding in the model transfusion
status. For non-SCD patients, 4 out of 10 WMSs displayed in Table 3.2 were still statistically
significant, and for SCD patients, all the WMSs reported in Table 3.2 still appeared. When including
LDH as a marker of hemolysis in the mathematical model, 6 out of 10 WMSs were statistically
significant for non-SCD and 4 of 5 WMSs for SCD patients.
3.4 Discussion
We observed potential microstructural differences along WMSs in both groups of patients with CA
with and without SCD compared to controls. These results predominantly showed lower FA values
in CA patients, indicating the loss of coherence in the main diffusion direction, which could indicate
WM injury. Lower FA was highly associated with decreasing Hb levels revealing that the decreased
microstructural integrity found in CA patients is highly driven by chronic hypoxia.
Previous work in CA has shown that the whole brain increases cerebral blood flow (CBF)
to compensate for the loss of oxygen-carrying capacity [1, 2, 159, 161, 188]. This offset in CBF
preserves total resting oxygen delivery to the whole brain [1, 3, 24, 159], such that the correspondent
oxygen extraction fraction (OEF) from the cerebral cortex seems to be normal or even reduced
[162, 189, 190, 191, 192]. Although resting oxygen delivery is preserved in the cortex, the cerebral
vascular reserve is diminished, proportional to the resting hyperemia [4, 160], potentially leaving
the brain vulnerable to acute insults such as nighttime hypoxia, acute anemia, and fever. However,
while there have been some reports of reduced cortical and subcortical GM volumes [108, 110, 112]
in patients not receiving chronic transfusion or hydroxyurea treatment, total GM volume appears
preserved [110] in clinically asymptomatic SCD patients (adolescents and young adults) treated
with either therapy suggesting that cortical volume loss may not manifest until later in life.
However, regions of deep WM do not seem to have the same metabolic stability during periods
of ischemic stress. OEF is increased in the deep watershed areas [188], colocalizing with WM
injury patterns [3, 5, 158]. Chai et al. [3] showed that independently of disease state, CBF
Chap. 3 Chronic anemia and the connectivity of WM 48
and oxygen delivery to regions of deep WM and border zone regions are considerably smaller
than those measured elsewhere within the WM and GM. Furthermore, Wang et al. [146] showed
that elevated CBF can be associated with normal-appearing (i.e., infarct-free) WM disruption.
Inadequate resting oxygen delivery in the WM is further compounded by blunted cerebrovascular
reserve [4].
Thus, chronic hypoperfusion plays a role in the development of the entire WM damage phenotype,
including hyperintensities on the T2 FLAIR [3, 5, 157, 158, 193], reduced WM volume [21, 103, 111]
or changes in diffusion metrics [20, 147, 156, 148, 144, 146].
Based on this consideration, the damage observed in non-SCD patients compared with controls
followed an intuitive pattern located primarily in the frontal-parietal WM watershed areas (Table
3.2). Watershed areas are regions in the brain that sit in-between major cerebral arterial territories
and are the most susceptible to hypoxic-ischemic damage when a supply-demand mismatch occurs
in the cerebrovascular supply [3, 158, 194].
While most of the affected connections were unilateral, WMSs observed with lower FA in nonSCD patients appeared similarly distributed between the two hemispheres (Figure 3.2). These
results are consistent with the spatial patterns of lower WM volume associated with the severity
of anemia diffusely across frontal, parietal lobes, and temporal lobes especially in these watershed
areas [103, 111].
Given that most of the WMSs survived when controlling for a hemolysis marker (Table 3.3),
the results are also aligned to a model of global hypoxia that will usually cause diffuse, bilateral
brain injury as seen in patients in drowning accidents, cardiac arrest, or bilateral carotid stenosis,
in contrast to more localized and asymmetric injury patterns caused by embolic stroke [195].
Therefore, we suggest that the injury pattern in WM microstructure of non-SCD patients can
indicate global chronic hypoxia driven by anemia’s effect on the brain’s hemodynamics. A possible
Chap. 3 Chronic anemia and the connectivity of WM 49
explanation is that the vascular architecture providing blood perfusion to WM areas is the longpenetrating medullary arteries with poor collateralization. Consequently, WM is especially vulnerable
to hyperintensities development under focal ischemic events or periods of acute stress [196].
For SCD patients and healthy controls, three out of five connections crossed to the contralateral
side. Interhemispheric involvement is consistent with previous results from our laboratory showing
lower FA in the corpus callosum in CA patients (higher burden on SCD) [144]. There are similar
observations on SCD in studies performed in Tanzania [148], the United Kingdom [156], and
the United States [146]. Kawadler et al. [147] also showed associations between microstructural
properties in the corpus callosum with daytime oxygen saturation and Hb levels in SCD patients,
indicating that hypoxia related at least in part to low hemoglobin in SCD patients drives the WM
injury patterns.
Previous DTI studies in SCD patients have also reported widespread FA decrease in the WM
[20, 147, 156, 146]. Surprisingly, we did not observe this extent of systematic FA derangements.
This difference possibly reflects the variability in disease expression in our SCD cohort compared
to previous reports; 7 subjects had SC genotype, and 5 of the 12 SS patients were receiving
chronic transfusions. While SC and chronically transfused patients develop WM hyperintensities,
the phenotype of their WM disease is less severe than nontransfused SS and Sβ
0
[197], and may
even result from different mechanisms. Furthermore, the mean hemoglobin F fractions among the
nontransfused SS patients was 18%, suggesting good response to hydroxyurea. Our non-SCD and
SCD cohorts were matched for hemoglobin level, so one could reasonably have expected a similar
spectrum of disease. Nevertheless, this is a cross-sectional study of young adults, and there is a
possibility that exposure to severely reduced arterial oxygen content prior to treatment irreversibly
affected brain microstructure during brain development in transfusion-dependent non-SCD patients.
Additionally, Table 3.3 shows almost no contribution from monthly transfusions or LDH in the
WMSs found in SCD patients. Possibly, the distribution of WMS in SCD and non-SCD looks
Chap. 3 Chronic anemia and the connectivity of WM 50
substantially different due to uncontrolled confounders, such as chronic pain reported extensively
as a burden for SCD patients [198, 199].
In childhood, SCD patients might have an intermittent pain phenotype. Around 50% of the
cases evolve as a chronic pain syndrome in adulthood, with periods of lower and higher pain
correlated with the ongoing vaso-occlusion [200]. Table 3.2 shows a significant involvement of the
orbito-frontal gyrus, which has been implicated in the modulation of chronic pain [201, 202, 203] and
pain-related emotions [204]. Furthermore, functional imaging studies have shown that regions like
the thalamus and the parahippocampal gyrus, also depicted in Table 3.2, belong to the functional
pain network [205, 206]. In particular, the thalamus has been identified as a central region that
processes pain [206]. Anemia, by itself, is a robust biomarker of disease severity in SCD, so it is
not surprising that hemoglobin levels correlate with FA in pain circuits.
Several research groups have also shown neurocognitive decline in patients with CA, suggesting
a possible and early involvement of the brain even in the absence of overt strokes [207, 208].
Many significant WMSs were in the prefrontal cortex (Figure 3.2), where WM abnormality has
been associated with negative effects on neurocognitive function in CA patients [103, 156, 144].
Specifically, Chai et al. reported that lower FA in the corpus callosum was associated with lower
scores across nine neurocognitive measures. At the same time, Stotesbury et al. [156] found that
white matter microstructural properties were associated with processing speed, where FA was the
strongest predictor. Additional work in our laboratory has previously demonstrated that lower WM
volume predicted low matrix reasoning scores, a measure of executive function, in CA patients and
identified changes in resting-state fMRI activity in the orbitofrontal and subcallosal gyri [155].
Altogether, microstructural injury patterns indicated in CA patients driven by low Hb levels may
have cognitive consequences.
Chap. 3 Chronic anemia and the connectivity of WM 51
3.5 Limitations
The small sample size limited our study. All participants were part of a larger project on Sickle
Cell Disease at Children’s Hospital Los Angeles that involved various MRI protocols (75 minutes
of total scan time) and was not limited to the dMRI sequence (4 minutes). Our current approach
allowed us to include all WM tracks, but it required to have high-quality structural and diffusion
data, limiting us to a smaller sample size than described in Chai et al. [144]. While significant sex
differences have been previously indicated in the study of CA, we could not fully resolve sex diseasespecific differences. The use of data from multiple cohorts of CA, whose intrinsic pathophysiology
is different, weakens statistical power in the short term but opens the possibility to differentiate and
characterize the unique damage induced by individual hemoglobinopathies. In addition, the control
subjects were mainly recruited from family members of SCD patients, and they do not necessarily
represent the non-SCD population. Consequently, the statistical differences that we found in nonSCD patients (even when controlling for log-age and sex) could be affected by random effects.
Although our contemporaneous hematological investigations are a strength of this study, we did
not have previous hemoglobin or any oxygen saturation values. The use of chronic blood transfusion
therapy in some of our patients potentially represents a limitation on our findings because no single
hemoglobin level characterizes the hypoxic exposure. Furthermore, SCD patients are often placed
on chronic transfusions later in life (than non-SCD patients) and their current hemoglobin values do
not reflect their lifetime hypoxic exposure. Chronic transfusion is also a complicated therapeutic
yielding improvement in erythrocyte deformability and oxygen-carrying capacity but increased
viscosity in the microcirculation can potentially worsen the blood flow and oxygen delivery [209].
Given our sample size, it would be tough to accurately separate the rheologic and oxygen-carrying
capacity effects of red blood cells. Furthermore, the inclusion of transfusion status in the model
weakens the statistical power by adding an additional degree of freedom. In addition, we were
not able to assess any additional effect of low oxygen saturation on arterial oxygen content and
therefore hypoxic exposure.
Chap. 3 Chronic anemia and the connectivity of WM 52
The constraints associated with using single-shell diffusion images and simple tensor modeling
are well documented in the literature, and urge caution to draw firm conclusions from a single
tensor metric like FA. This work tried to address some of these limitations by using ERFO ODFs
to correctly render crossing fibers and creating WMSs to avoid characterizing connectivity by
streamlining counting. However, we recognize that the information provided by FA is limited and
that other methods like diffusion kurtosis imaging and neurite orientation dispersion have proven
to be more robust to some of the pitfalls of DTI and could provide a more biological explanation
of our current observations.
3.6 Conclusion
To characterize the effects of CA in white matter, mean FA along the WMSs (surface connecting
two ROIs) of chronic anemic patients with sickle and non-sickle anemias were compared with
healthy controls. This grouping allowed the isolation of sickle hemoglobin effects in our analysis.
Both CA cohorts showed localized FA differences along the WMSs of patients compared with
controls but did not show differences between them. However, non-SCD patients manifested bigger
systematic FA derangements in the watershed areas that were bilateral and spatially symmetrical.
These results suggested that the broad spectrum of variability in disease expression in our sickle cell
anemia cohort and uncontrolled confounders of mesostructure integrity affected our ability to detect
widespread WM abnormalities as proposed in the literature. Nevertheless, finding interhemispheric
WMSs affected in SCD aligns with previous literature reports showing decreased FA in the corpus
callosum in CA patients. Recognizing both the differences and the similitudes between CA patients
and the affliction that anemia causes in white matter may help develop earlier and more generalized
interventions to help overcome the anemia burden.
Chapter 4
Unraveling the link between chronic
anemia and white matter damage: A
comprehensive diffusion MRI analysis
“The task is not so much to see what no one has
yet seen, but to think what nobody has yet thought
about that which everybody sees.”
— Erwin Rudolf Josef Alexander Schr¨odinger
4.1 Introduction
Chronic anemia (CA) is a prevalent global health problem that affects approximately 1.92 billion
people [210]. It is linked to adverse health consequences, heightened morbidity and mortality
rates, and to significant healthcare and economic burdens [211]. CA is characterized by ongoing
shortages of red blood cells (erythrocytes) or lower hemoglobin (Hb) levels, which carry oxygen
in the bloodstream. The brain is one of the organs with high metabolic demand, and despite its
physiological adaptation to oxygen fluctuations, it is prone to hypoxic injuries.
53
Chap. 4 Chronic anemia and WM damage 54
CA can be caused by genetic mutations that affect erythrocyte’s production, structure, or
function. These disorders appear in childhood [36] and serve as models of chronic brain injury
caused by hypoxia. Some well-known examples include sickle cell disease (SCD, qualitative disorders
of Hb structure) and thalassemia syndromes (non-SCD, quantitative disorders of Hb synthesis) [15].
In particular, SCD occurs due to the presence of abnormal hemoglobin S (Hb S) [47], which causes
erythrocytes to deform into rigid crescent shapes and substantially reduces their life expectancy
[61]. SCD causes a broad range of repercussions on the brain, from cognitive disabilities to stroke,
caused by anemia, microvascular obstruction, chronic pain, and even secondary manifestations
of acute chest syndrome such as fat emboli [62, 184]. Thalassemias, however, result from a
globin chain subunit affected by a genetic defect (either α- or β-globin) [212], causing ineffective
erythropoiesis [49] and intramedullary hemolysis [213]. Central features of its pathophysiology that
impact the brain include anemia, dysregulation of iron homeostasis, and a chronic hypercoagulable
state [15, 212].
While early MRI studies in SCD emphasized cerebral cortex changes [18, 108, 109, 214, 155, 215],
more recent studies have demonstrated silent cerebral infarctions [5, 216, 217, 218, 219, 220], WM
loss [103, 110, 221], and abnormal perfusion in deep WM areas [5, 3, 96, 146]. Diffusion MRI is
particularly well suited to study the WM microstructure, facilitating the quantitative assessment
of both healthy and pathological tissue characteristics. Patients with SCD exhibit altered diffusion
tensor imaging (DTI) metrics compared with healthy controls (CTL), in particular, there was a
decrease in fractional anisotropy (FA, loss of the main diffusion directionality of water) in various
WM regions [142], including the corpus callosum [20, 143, 144, 145, 146], the internal capsule [143],
the corticospinal tract [144], and within the deeper white matter [143, 145, 146].
In contrast, there was an increase in mean diffusivity (MD, average diffusion of water) across the
frontal and parietal lobes [20, 142, 147], corpus callosum [20, 142, 147], and subcortical structures
[146, 147], and radial diffusivity (RD, magnitude of water diffusing perpendicular to WM) in the
corticospinal tract [20, 147], anterior thalamic radiations [147], the internal capsule and frontal and
Chap. 4 Chronic anemia and WM damage 55
parietal lobes [20, 147].
The extent to which WM integrity is affected in non-sickle anemias remains a topic of debate.
Patients with thalassemia have been found to experience similar WM loss to those with SCD [21],
with reductions in WM volume and cognitive function correlating with the severity of anemia,
regardless of hemoglobin genotype [103]. DTI analysis conducted by our laboratory also revealed a
comparable pattern to SCD, with decreased FA observed in major WM tracks such as the corpus
callosum and cortico-spinal tract [144]. Additionally, a different study found that non-SCD patients
exhibited more WM disarrangements than SCD patients, particularly in interhemispheric WM
bundles and watershed areas) [145]. In contrast, research conducted on Italian cohorts did not
reveal any differences in volumetric or DTI measurements between β-thalassemia patients and age
and sex-matched CTL individuals [23].
Thus, this study revisited the similarities and differences in WM integrity between sickle and
non-sickle anemia syndromes using MRI, cognitive testing, and phlebotomy. We extended previous
investigations through the analysis of multi-shell acquisitions such as diffusion kurtosis imaging
(DKI) and the multi-compartment model neurite orientation dispersion and density imaging -
(NODDI). By doing so, we sought to determine which changes could be attributed to anemia
alone, as opposed to disease-specific pathophysiology. Despite being popular in clinical settings,
DTI lacks specificity in the brain as it assumes that water motion is unhindered.
Cell structures hinder water molecules in the complex tissue environment of WM, formed by
axons, myelin sheaths, astrocytes, oligodendrocytes, and microglial cells [222]. DKI enhances the
traditional DTI framework by better-capturing tissue constraints, resulting in a more accurate
representation of white matter microstructure and its alterations. NODDI models three tissue
compartments (intra-, extra-neurite, and free water fraction) in a biologically informed manner.
This approach enables the individual estimation and analysis of various parameters. Our findings
Chap. 4 Chronic anemia and WM damage 56
contribute to the growing body of research on WM implications in CA and introduce potential
biomarkers for assessing CA-related damage.
4.2 Materials and methods
4.2.1 Participation criteria
We conducted this study at Children’s Hospital Los Angeles and the Amsterdam University Medical
Center in The Netherlands. Local Institutional Review Boards approved this study, and all
participants provided informed consent. We collected MRI data and blood tests from eligible
participants between the ages of 16 to 65 years old, including patients with non-SCD diagnosis
of β-thalassemia major, thalassemia intermedia, E-β thalassemia, hemoglobin H-constant spring,
α-thalassemia Hb H, sideroblastic anemia, SCD diagnosis of Hb SS, Hb Sβ
0
, Hb Sβ
+, and Hb SC
genotypes, as well as control subjects matched by sex and age to the SCD patients.
Exclusions were made for patients with previous overt stroke, acute chest syndrome, pain
crisis hospitalization (within one month), and pregnant candidates. Furthermore, we examined
patients undergoing chronic transfusion therapy, specifically up to two days before their scheduled
routine transfusions. Ineligible CTLs were individuals with a prior history of neurologic insults,
developmental delay, or chronic medical conditions that require regular medical care or medications,
and pregnant candidates.
Throughout the study, we strictly adhered to the established guidelines and regulations for MRI
safety, as well as the appropriate MRI exclusion criteria to ensure the well-being of all participants.
On the designated day, each participant underwent the MRI examination without the use of sedating
medications, ensuring that the data collected remained unaltered by any external factors.
Chap. 4 Chronic anemia and WM damage 57
4.2.2 Laboratory markers
Additionally, we recorded vital signs and obtained blood samples from each participant to gather
comprehensive information to capture both commonalities and distinctions in the pathophysiology
of chronic anemia. Blood parameters included hematocrit (percentage), erythrocytes (x106/µL),
Leukocytes (x103/µL), absolute reticular count (ARC, x103/µL), mean corpuscular volume (MCV,
fL) and quantitative Hb electrophoresis (proportions of Hb A, Hb F, hemoglobin S). We also
quantified lactate dehydrogenase (LDH, U/L) as an indicator of hemolysis.
4.2.3 Image acquisition
We collected MRI data for every participant using a 3T Philips Achieva scanner with a 32-channel
head coil. The structural 3D T1-weighted (T1-w) sequence had a specification of TR/TE = 8.3/3.8
ms, SENSE = 2, and isotropic voxel size of 1 mm3
. Moreover, we obtained a multi-shell dMRI
sequence with TR/TE = 5,300/89 ms; voxel size of 1.8x1.6x2 mm3
; 30 diffusion-encoding directions
at b-values = 1,000 and 2500 s/mm2 and three b-values = 0 s/mm2 using a multi-band SENSE
factor of three.
4.2.4 Post-processing
The T1-w images underwent a series of processing steps using BrainSuite(https://brainsuite.
org, v21a) to extract brain structures and perform tissue classification. BrainSuite’s Cortical
Surface Extraction (CSE) tool was utilized for skull stripping [164], tissue classification, and
determination of partial volume fractions for WM, gray matter (GM), and cerebrospinal fluid
(CSF) voxels [165]. Additionally, topological corrections and delineation of the inner and outer
cortex were performed. Furthermore, BrainSuite’s SVReg tool [167, 168] was used to perform
anatomical co-registration of the T1-w with the BCI-DNI anatomical atlas [168] and to segment
the brain.
Chap. 4 Chronic anemia and WM damage 58
For the dMRI data, a two-step preprocessing procedure was employed to ensure data quality
and enable precise multi-modal analysis. Firstly, DIPY was utilized to perform noise removal
[223] and Gibbs ringing correction [224]. Subsequently, FSL was employed to address distortions
induced by motion, susceptibility, and eddy currents [169, 170], further refining the dMRI data for
comprehensive analysis.
4.2.5 Diffusion modeling
We employed the DIPY libraries [225] to fit both the diffusion tensor and kurtosis models. Notably,
the diffusion tensor can be effectively separated from the effects of higher-order terms within DKI’s
tensor [226]. As a result, all standard diffusion tensor statistics (FA, MD, AD, and RD) can be
derived from the kurtosis fit and are anticipated to exhibit superior accuracy compared to the
regular tensor model [227]. This decoupling of the diffusion tensor within the kurtosis framework
provides a promising avenue for more precise diffusion-related metrics.
Additionally, we calculated the NODDI parameters using the NODDI MATLAB Toolbox [123]:
neurite density (NDI), orientation dispersion (ODI), and free water fraction (FISO).
Using tools from BrainSuite’s Diffusion Pipeline (BDP), we registered the tensor, kurtosis, and
NODDI maps to the T1-w data (susceptibility-distortion correction based on the inverse contrast
normalization [171]), followed by co-registration of these maps to the BCI-DNI atlas using SVReg
outputs.
4.2.6 Statistical analysis
We only analyzed the voxels identified as pure white matter in the BCI-DNI atlas to avoid partial
volume effects. Our diffusion analyses compared the CTL group with SCD and non-SCD patients,
but we did not compare the two patient groups. Figure 1 provides a visual representation of this
analysis. Given the complexity of the data, it was impossible to ensure that all assumptions of
linear modeling were met. Therefore, we used Manly’s method for permutation analysis to model
differences between groups after controlling for the logarithm of age (log-age) and sex. We calculated
the logarithm of age because brain maturational effects are nonlinear with age in adolescents and
Chap. 4 Chronic anemia and WM damage 59
young adults [180, 181]. All the results were adjusted for multiple comparisons using the false
discovery rate (FDR) and a 0.05 threshold acceptance rate was applied to the corresponding pvalues [179, 182].
To account for the effects of chronic anemia in all the tensor and kurtosis metrics, in addition to
log-age and sex, we also regressed the Hb values and then performed a permutation analysis. For
completeness, we also tested the potential contribution of Ferritin (ng/mL), Hb A (%), Hb F (%),
Leukocytes (x103/µL), MCV (fL), ARC (x103/µL), and LDH (U/L) values only for those maps
that had a residual signal after regressing Hb (g/dL). Each of these measurements was added one
at a time. All statistical analyses were conducted using MATLAB (R2022b).
4.3 Results
4.3.1 Demographics
In this analysis we considered 32 CTL subjects (age = 31.7 ± 8.5 years; Hb = 13.0 ± 1.6 g/dL;
F = 19 individuals), 20 clinically asymptomatic non-SCD patients (age = 30.6 ± 9.9 years; Hb =
9.6 ± 1.5 g/dL; F = 12 patients), and 76 clinically asymptomatic SCD patients (age = 29.5 ± 10.3
years; Hb = 9.8 ± 1.9 g/dL; F = 31 patients). The age range for all the participants was 16.5 to 63
years. The breakdown of ethnicity for CTL subjects was 61 African-Descendants, 7 Hispanic and
8 from other groups. For non-SCD|SCD we had 0|61 African-Descendants, 3|7 Hispanic and 17|8
from other ethnicities. All demographics are reported in Table 1.
The genotypes for six of the control subjects were identified with sickle cell trait having hemoglobin
AS (Hb AS). Findings from prior research indicated that the Hb AS subtype does not modify
regulation of cerebral blood flow (CBF) or the equilibrium between oxygen supply and demand
[24]. This observation positions carriers of Hb AS suitable candidates for control subjects.
Chap. 4 Chronic anemia and WM damage 60
Table 4.1: Demographics.1
CTL non-SCD SCD
CTL
vs
nonSCD†
CTL
vs
SCD†
nonSCD
vs
SCD†
N 32 20 76 - - -
Age 31.7±8.5 30.6±9.9 29.5±10.3 - - -
Sex 19F:13M 12F:8M 31D:45M - - -
Transfused 0 13 11 - - -
Hydroxyurea 0 0 40 - - -
Hemoglobin
(g/dL) 13.0±1.6 9.6±1.5 9.8±1.8 ≤ 0.01 ≤ 0.01 0.99
Hematocrit
(%) 38.1±8.2 26.4±11.7 28.4±5.4 ≤ 0.01 ≤ 0.01 0.54
Erythrocytes
(x106/µL) 4.5±0.4 3.9±0.7 3.4±1.0 0.07 ≤ 0.01 0.02
Leukocytes
(x103/µL) 5.9±1.7 6.9±2.7 8.0±3.4 0.43 ≤ 0.01 0.36
Mean
corpuscular
volume (fL)
87.5±8.2 77.8±8.8 86.2±14.6 0.02 0.86 0.02
Ferritin
(ng/mL) 53.7±53.1 1 1087±1249 - 706±1795 - 0.04 0.10 0.56
Reticulocytes
(%) 1.5±0.7 2 3.5±3.8 1 7.4±4.7 - 0.21 ≤ 0.01 ≤ 0.01
Absolute
reticular count
(x103/µL)
68.8±30.7 2 143.2±152 1 222±129.6 - 0.08 ≤ 0.01 0.02
Heart rate
(min−1
)
69.7±10.8 6 75.4±9.7 1 71.5±12.9 3 0.16 0.76 0.28
Systolic blood
pressure
(mmHg)
119.8±12.5 5 108.6±14 1 114.8±13.6 3 0.01 0.22 0.18
Diastolic blood
pressure
(mmHg)
75.4±11.2 5 67.1±10.1 1 67.7±9 3 0.01 ≤ 0.01 0.96
Lactose
dehydrogenase
(U/L)
462.7±262 3 497.7±322 - 533±318.3 12 0.91 0.56 0.89
Hemoglobin A2
(%) 2.7±0.4 1 2.3±1.2 3 4.8±1.4 - 0.32 ≤ 0.01 ≤ 0.01
Hemoglobin F
(%) 0.3±0.4 - 5.4±10.1 - 9±10.9 2 - - -
Hemoglobin S
(%) 6.7±14.6 5 0.0±0.0 - 69.8±22.2 1 - - -
PSI 0.15±0.9
M i s s i n g - v a l u e s
8 0.08±0.7
M i s s i n g - v a l u e s
6 -1.08±1.0
M i s s i n g - v a l u e s
14 0.96 ≤ 0.01 ≤ 0.01
1 Mean and standard deviation of demographic information and selected blood count measurements.
† Group comparison using one-way analysis of variance (ANOVA) result with Tukey-Kramer test for multiple
comparisons. Statistical significant values (p≤0.05) are displayed in bold letters.
Chap. 4 Chronic anemia and WM damage 61
The specific anemias in the non-SCD group consisted of 10 patients with β-thalassemia major,
2 β-thalassemia intermedia, 1 E-beta thalassemia, 2 Hb H constant spring, 4 α-thalassemia Hb H
and 1 sideroblastic anemia. For the SCD group, the genotypes were 46 Hb SS patients, 8 Hb Sβ
0
,
7 Hb Sβ
+, and 12 Hb SC.
Furthermore, of the CA patients, 13 non-SCD (10 β-thalassemia major, 1 β-thalassemia intermedia,
1 α-thalassemia Hb H, and 1 sideroblastic anemia) and 11 SCD (Hb SS) patients were on monthly
transfusions. Additionally, of the SCD group 26 of the non-transfused Hb SS, 3 Hb SC, 6 Hb
Sβ
+, and 5 Hb Sβ
0 patients were prescribed hydroxyurea and had a mean Hb F fraction of 13.5%.
Three patients with SC were also taking hydroxyurea. Children’s Hospital Los Angeles advises
the administration of hydroxyurea to all pediatric patients aged nine months and older, regardless
of their SCD genotypes, unless they are already undergoing chronic transfusion [184, 185, 187].
This approach aligns with the guidelines outlined by the National Institute of Health (NIH) [186].
Furthermore, as of the year 2000, all SCD patients at our facility have received access to the
transcranial Doppler screening [184, 185, 186, 187]. Very similar guidelines are followed at the
Amsterdam University Medical Centers in The Netherlands [228, 229].
4.3.2 DTI Results
The DTI results discussed in this study were adjusted for age and sex, and only statistically
significant voxels (p<0.05, after FDR correction) are displayed in the T-maps. The left-hand
columns of Fig. 4.1 and Fig. 4.2 show the maps for FA and MD, respectively. Similarly, the RD
and AD maps are displayed in Fig. A.1 and Fig. A.2. The top panels compare SCD patients
vs. CTL subjects, and the bottom panels show the non-SCD patients vs. CTL subjects. In all
these figures, the magma (red) and blue denote decreased and increased values in SCD or non-SCD
patients with respect to controls. For all the DTI metrics, no statistically significant results were
found when adding hemoglobin to the regression model.
Chap. 4 Chronic anemia and WM damage 62
Figure 4.1 (left three columns) shows the fractional anisotropy T-maps. Both SCD and non-SCD
patients exhibit a similar pattern of decreased FA around the watershed WM regions, including
the corpus callosum (mostly the genu) and internal capsule, as well as widespread WM areas closer
to the cerebral cortex (including areas in the prefrontal, temporal, parietal, and occipital lobes).
However, in SCD patients, Fig. 4.1 (left-superior panel), these areas appear wider and have smaller
T-values than those in non-SCD patients. Additionally, both groups display clusters of increased
FA, mainly around the watershed areas. However, the biggest clusters in SCD patients appear near
the right and left globus pallidus, right-superior, and left-posterior corona radiata.
Complementary to the information provided by the FA maps, Fig. 4.2 (left three columns) shows
the T-maps for the mean diffusivity when comparing SCD|non-SCD patients to CTL individuals.
For the most part the SCD and non-SCD maps showed increased MD not only in the watershed WM
regions, including the corpus callosum and the internal capsule, but also in wide-spread clusters near
the cerebral cortex. The non-SCD patients (see Fig. 4.2 bottom left) also showed a clear pattern
of increased MD around the cingulum (above the corpus callosum and near the hippocampus).
Interestingly, the watershead areas including the left|right posterior thalamic radiation and posterior
limb of the internal capsule, in the SCD group showed clusters of decreased MD.
Additionally, in both SCD and non-SCD patients, the watershed areas including the corpus
callosum, cingulum, external, and internal capsule, and cortical WM, Fig. A.1 (left three columns),
showed increased, symmetrical RD patterns. Interestingly, both groups have clusters of decreased
RD. SCD patients, in particular, Fig. A.1 (left-superior panel), show decreased RD around the
posterior limb of the internal capsule, posterior thalamic radiations, and globus pallidus.
Lastly, Fig. A.2 (left three columns) displays the T-maps of axial diffusivity that show a mixed
pattern of increased and decreased AD. Although the majority of the clusters represent increased
AD, those showing decrements had higher absolute values. And, in non-SCD patients, Fig. A.2
(left-inferior panel), most of the clusters of increased AD appear near the cortex.
Chap. 4 Chronic anemia and WM damage 63
Figure 4.1: FA and KFA: T-maps displaying voxels that were statistically significant (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The left three columns
reveal a mixture of decreased and increased fractional diffusivity (magma and blue color scales).
Very similar patterns are observed in the right three columns depicting the fractional anisotropy
kurtosis when controlling for logarithm of age and sex. No results were found when hemoglobin
was added to the statistical model. These findings strongly suggest that anemia drives the observed
FA’s and KFA’s decrements|increments in both the SCD and non-SCD cohorts.
Chap. 4 Chronic anemia and WM damage 64
Figure 4.2: MD and MK: T-maps displaying voxels that were statistically significant (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The left three columns
reveal an overall increase in mean diffusivity (blue color scale), while the middle three columns
depict an overall decrease in mean kurtosis (magma color scale) when controlling for logarithm of
age and sex. Additionally, the right three columns highlight statistically significant voxels when the
hemoglobin effects are also regressed out from the MK maps. Notably, the only remaining signal was
found in the MK model of SCD vs CTL, whereas non-SCD patients showed no statistical difference.
Moreover, in both cases, the MD maps showed no statistical differences when also removing the
effects of hemoglobin. These findings strongly suggest that anemia is the driving factor behind the
observed WM changes in the non-SCD cohort.
Chap. 4 Chronic anemia and WM damage 65
4.3.3 DKI results
As with the DTI, we controlled all the results of the kurtosis metrics for log-age and sex. Only those
statistically significant voxels (p<0.05, after FDR correction) are displayed in the T-maps in Figs.
4.1, 4.2 and Figs. A.1, A.2 (middle and right three columns). The magma (red) and blue denote
decreased and increased values in SCD or non-SCD patients with respect to controls. Interestingly,
when regressing the effects of Hb in the SCD vs. CTL comparison, several voxels in the MK (Fig.
4.2), RK (Fig. A.1), and AK (Fig. A.2) remained statistically significant (see right-superior panel
in all the figures). This was not true for KFA (Fig. 4.1). However, for the non-SCD vs control
group, none of the kurtosis metrics displayed statistical significance after removing the Hb effects.
These findings strongly suggest that Hb drives the changes in kurtosis highly for the non-SCD group.
The T-maps of the fractional kurtosis are presented in the right three columns of Figure 4.1,
while controlling for log age and sex. The KFA results for SCD and non-SCD patients exhibit
patterns comparable to their DTI counterparts, with the DKI’s clusters of voxels appearing wider
than those of DTI’s, mainly for voxels with decreased values. Generally, the patterns of decreased
KFA around the watershed WM regions, including the corpus callosum and the internal capsule,
as well as widespread WM areas closer to the cerebral cortex, are similar in CA patients. However,
in SCD, the areas seem wider in Fig. 4.1 (right-superior panel), with smaller T-values than those
of non-SCD patients. Both groups display clusters of increased KFA mainly around the watershed
areas, with SCD also showing the most prominent clusters of increased values.
In Fig. 4.2, the middle three columns show the mean kurtosis maps after adjusting for log age
and sex. In both SCD|non-SCD vs. CTL comparisons, the primary effect is a significant decrease
in MK that affects almost the entire WM. As MD increases in certain regions, the corresponding
MK will decrease as the water moves freely in all directions. Importantly, aside from the tensor
analogous clusters, we observed that MK decreased in many other areas as well. On the other
hand, the clusters showing increased MK are in good agreement with those that showed decreased
MD, primarily observed in the SCD vs. CTL, Fig. 4.2 (middle-superior panel). When we look
Chap. 4 Chronic anemia and WM damage 66
at the hemoglobin-regressed model, we see that the WM regions that are preserved are those that
appear with the highest-yellowish intensity on the log age and sex regression model in Fig. 4.2
(right-superior panel).
Notably, the results obtained from the MK maps go beyond the areas that have been previously
identified in the literature as being affected by ischemic processes, which are mostly limited to
the watershed regions. The findings indicate that the white matter of chronic anemic patients
undergoes changes in almost its entirety, including areas close to the cortex. This suggests that the
pathophysiology of the white matter in these patients is influenced by factors other than hypoxia.
In the middle three columns of Fig. A.1, a vast area with decreased RK values can be observed in
both SCD and non-SCD patients. These patterns are primarily symmetrical around the watershed
areas and close to the cortex resembling the MK changes observed in Fig. 4.2. Furthermore, both
groups exhibit clusters of increased RK, though non-SCD patients have fewer clusters. A closer
examination of the Hb-regressed model reveals that the preserved WM regions correspond to those
exhibiting the highest yellowish-intensity on the log age and sex regression model, which are clusters
predominantly contained within the watershed WM areas, Fig. A.1 (right-superior panel).
Lastly, the middle three columns of Fig. A.2 show T-maps for axial kurtosis that exhibit a clear
pattern of clusters with decreased AK values. It is important to note that areas with reduced AK
values are more prominent than those with increased AD values, Fig. A.2 (left three columns).
However, in both groups, there is a trend of increasing AK values, particularly along the watershed
WM areas, that resemble some of the clusters of decreased AD. Similar to the MK and RK cases,
the Hb-regressed model in AK maps shows that the remaining signal are those voxels with highestyellow intensity (i.e., those with the smallest AK values) in the log age and sex regression model.
Chap. 4 Chronic anemia and WM damage 67
Figure 4.3: PSI: Statistically significant correlation (p<0.05) between MK and PSI while controlling
for logarithm of age and sex. Positive correlation coefficients are represented on the blue scale, and
negative correlation is depicted on the magma scale. The left panel presents the results for SCD
patients and CTL individuals, whereas the right panel displays the results for non-SCD patients
along with CTL individuals.
4.3.3.1 Additional Blood Markers
After accounting for the influence of Hb on the MK, AK, and RK maps, it was evident that there was
still a discernible signal when comparing SCD patients to healthy controls (Fig. 4.2, right column).
To determine its source as well as its physiological principles, we serially integrated leukocytes, Hb
F, lactose dehydrogenase, and absolute reticular count (see Table 4.1) into the multi-linear model
(Fig. A.3). Leukocyte values did not discernibly alter the results (Fig. A.3, left two panels) and
including Hb F only marginally explained the residual differences between SCD and control subjects
(middle panel). However, upon incorporating lactate dehydrogenase or absolute reticulocyte count,
all MK differences between SCD and control disappeared (right two panels). Although hemoglobin
and markers of hemolysis are correlated in SCD, both factors were necessary to explain MK, AK,
and RK differences between SCD patients and healthy control subjects.
Chap. 4 Chronic anemia and WM damage 68
4.3.3.2 Processing Speed Index
In Table 4.1, the mean and standard deviation of the PSI index for the SCD, non-SCD, and
CTL groups are shown. A statistical significance was found only in the comparison between the
anemic SCD patients and the healthy CTL individuals. Figure 4.3 depicts the correlation between
the PSI and MK, with a blue scale indicating a direct relationship—where a decrease in MK is
associated with a decrease in PSI. The deeper white matter structures display the most consistent
positive correlation with the more negative T-scores observed in the MK maps shown in Figure 4.2.
Additionally, the relationship between MK and PSI is notably stronger in individuals with SCD
compared to those with non-SCD anemias, although the overall patterns remain similar.
4.3.4 NODDI results
To maintain consistency with DTI and DKI methodologies, all NODDI results were adjusted for
log-age and sex parameters. The T-maps presented in Fig. 4.4 and Fig. A.4 illustrate only those
voxels that reached statistical significance (p<0.05 following false discovery rate correction). In
these visual representations, red denotes decreased values, while blue indicates increased values in
patients with SCD or non-SCD patients compared to control subjects.
Furthermore, the analysis of hemoglobin’s effects during the comparison between SCD and
control groups revealed statistically significant voxels in the NDI, as depicted in the right-superior
panel of Fig. 4.4. In contrast, no significant results were observed for the ODI and FISO metrics
(refer to Fig. A.4). Additionally, none of the NODDI metrics for the non-SCD in comparison with
healthy controls exhibited statistical significance after accounting for hemoglobin effects (see Figs.
4.4 and A.4). These results indicate that hemoglobin levels predominantly influence the alterations
in NODDI metrics for the non-SCD group.
Chap. 4 Chronic anemia and WM damage 69
Figure 4.4: NDI: T-maps displaying voxels that were statistically significant (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The three columns
on the left reveal a mixture of decreased (mainly) and increased neural density index (magma and
blue color scales). In addition, when hemoglobin was added to the statistical model, those voxels
with the highest pink intensity (decreased values) were preserved only in SCD patients. These
findings strongly suggest that anemia drives the observed NDI decrements in both cohorts, but in
SCD, more factors influence the changes in NDI.
Chap. 4 Chronic anemia and WM damage 70
In Fig. 4.4, the right three columns present the T-maps of the neurite density index, while
controlling for log-age and sex. The patterns of increased and decreased NDI observed in patients
with SCD and non-SCD exhibit a significant resemblance to their MK counterparts, as illustrated in
Fig. 4.2 (middle three columns). Specifically, SCD patients demonstrate a marked reduction in NDI
values, accompanied by increased values predominantly located in the globus pallidus, thalamus,
and the posterior limb of the internal capsule. This trend parallels the MK results shown in Fig.
4.4 (left-superior panel).
Additionally, non-SCD patients also reveal a pronounced pattern of decreased NDI values, with
minimal to nonexistent occurrences of increased values, as depicted in Fig. 4.4 (left-inferior panel).
In the same figure, the right superior panel illustrates the results obtained after accounting for the
effects of Hb. Nearly all the bright yellow clusters identified in the log-age and sex model were
retained; however, almost all the increased NDI were insignificant. These maps demonstrated a
high level of similarity to their MK counterparts (right three columns of Fig. 4.2).
The findings regarding the orientation dispersion index are illustrated in Fig. A.4 (left three
columns). Comapred to the healthy control group, both patient cohorts demonstrated regions of
increased and decreased ODI throughout the white matter. Specifically, SCD patients displayed a
combination of elevated and diminished ODI values in areas adjacent to the watershed regions,
the internal capsule, and the corpus callosum, as depicted in Fig. A.4 (left-superior panel).
Furthermore, these patients presented decreased ODI clusters in the external capsule and the
coronal radiata on both hemispheres. In contrast, an increase in ODI values was observed in white
matter regions closer to the cerebral cortex.
Chap. 4 Chronic anemia and WM damage 71
In addition, the left-inferior panel of Fig. A.4 reveals that non-SCD patients generally exhibited
heightened ODI patterns around the prefrontal cortex, watershed areas, the cingulum, and the
corpus callosum, particularly in the genu and body regions, when compared to healthy controls.
Conversely, there were diminished ODI values noted in the superior corona radiata on both sides.
The isotropic fraction, also referred to as the free water fraction, is illustrated in the right
three columns of Fig. A.4. In comparison to the control group, patients diagnosed with SCD
exhibit significantly lower FISO values, not only in specific regions of the watershed but also across
extensive areas of the WM adjacent to the cortex. Notably, there exist clusters of increased FISO
values, primarily concentrated near the ventricles and within the bilateral posterior limb of the
internal capsule, as indicated in the right superior panel of Fig. A.4.
Lastly, it is observed that non-SCD patients have fewer clusters than their SCD counterparts,
as shown in the right inferior panel of Fig. A.4. The clusters found in non-SCD patients are more
broadly dispersed and generally show lower FISO values, with little to no FISO increases. The
reduced clusters are predominantly situated within the bilateral sagittal stratum, the posterior
thalamic radiations, and the anterior corona radiata.
4.4 Discussion
The distinctive novelty in our study lies in the utilization of DKI’s modeling of non-Gaussian water
diffusion and NODDI’s biological representations of the WM tissue to reveal three key findings in
chronic anemia:
• Patients with chronic anemia experience extensive changes in the white matter regardless of
the intrinsic cause of anemia. These changes go beyond what the literature has described
as damage caused by hypoxia [3, 5, 146], indicating that other neuropathic processes like
demyelination might be happening in the CA population.
Chap. 4 Chronic anemia and WM damage 72
• We found significant differences in the white matter of SCD patients (only) compared to
healthy controls when hemoglobin (the driving factor in anemia) was regressed out from the
kurtosis and NODDI metrics.
• Finally, we also identified blood markers that may explain further damage and processes
happening in the white matter of SCD patients that are not significantly present in non-SCD
patients, namely lactate dehydrogenase, hemoglobin A, and absolute reticular count.
This study also proved that the unique integration of more advanced diffusion models while
studying chronic anemia from SCD and non-SCD patients offers a more detailed and subtle understanding
of the microstructural alterations that unfold across the spectrum of CA.
4.4.1 WM alterations measured by dMRI models
4.4.1.1 Diffusion tensor in CA
Our research aligns with previous studies that have identified a decline in FA among SCD patients
in the corpus callosum, internal capsule, and deep watershed WM regions, Fig. 4.1 (left three
columns). However, our study expands on these findings by demonstrating that non-SCD patients
also display a similar pattern of reduced FA in these regions Fig. 4.1 (left three columns).
Additionally, both groups also exhibit numerous clusters of reduced FA in regions closer to the
cortex. Interestingly, we also identified clusters of increased FA primarily across the WM watershed
areas, suggesting that distinct processes may be taking place simultaneously in the WM.
These FA findings also revealed a spatial correlation to the MD maps, albeit without a precise
one-to-one match. Additionally, regions exhibiting reduced FA, indicative of diminished directional
coherence of water diffusion, correspond to those displaying elevated MD, indicative of an overall
increase in diffusion, and vice versa. Previous research has suggested that reduced FA and increased
MD are commonly found together in cases of demyelination, a condition that is a consequence of
chronic inflammation and hypoxic-ischemic processes, both of which are known to be prevalent in
our patient population. As such, our findings provide compelling evidence for the proposition that
Chap. 4 Chronic anemia and WM damage 73
regions exhibiting reduced|increased FA|MD are focal points of demyelination.
Notably, a combination of clusters with both decreased and increased MD is more prevalent
in the watershed regions of SCD patients, Fig. 4.2 (left-superior panel). This could indicate that
the WM in these patients might be undergoing more complex processes than solely demyelination.
Conversely, in these same areas, non-SCD patients exhibit a prevalence of clusters with increased
MD that also expand to areas closer to the gray matter, with minimal instances of decreased MD,
Fig. 4.2 (left-inferior panel).
Furthermore, upon comparing FA with RD results, a clear spatial overlap between decreased FA
and increased RD (increment in perpendicular diffusivity) is observed and the other way around.
This pattern is evident in both SCD and non-SCD patients (with increased RD accounting for most
of the changes observed in WM). This model of decrease|increase FA|RD has been also associated
with electrophysiological markers of demyelination [230]. Additionally, with respect to MD, RD
shows a positive correlation: areas of increased RD correspond to regions of increased MD, and
the opposite relation is also true. Especially in non-SCD patients, areas closer to the GM with
increased MD overlap with clusters of increased RD.
Interestingly, the comparison between FA and AD does not show a clear correspondence as
FA and RD maps (or even MD and RD) have. Specifically, the extent of reduced FA does not
correspond precisely with the extent of increased AD; instead, in these same regions, there is a mix
of decreased and increased AD clusters for both patient groups. However, regions with increased
FA have a better alignment with those of increased AD. Additionally, with respect to MD, AD does
not show a specific overlap or pattern of incremented|decreased values.
Putting this information together, we see that for the most part the DTI model in CA shows
reduced FA but increase in both RD and MD maps. This further supports our theory that CA
patients, for the most part, might be undergoing demyelination and possible processes of axonal
Chap. 4 Chronic anemia and WM damage 74
loss as this same relationship (between FA, MD and RD) has been described in human models of
demyelination. However, the results of the AD maps are slightly harder to interpret, and this is not
unique to the data presented here [231]. In some studies, AD and FA showed significant correlation
with axonal integrity [230]. However, some studies have also suggested that in pathologic changes
from acute to chronic stage result in AD becoming less informative over time. A clear example
where AD measurement will not be completely reliable is the case of inflammation (acute to chronic
sate) [232]: This process will lead to less tightly packed demyelinated axons and wider spaces in
between them. In this situation, DTI will have a harder time detecting and measuring AD within
these fiber bundles and this can be an explanation why we see do not see a clear pattern of increase.
4.4.1.2 Diffusion kurtosis in CA
Despite its frequent use in clinical settings, DTI lacks specificity as it assumes a Gaussian water
distribution in biological tissues. However, considering the highly complex in-vivo environment,
such as the white matter composed of axons and myelin sheets, the behavior of water molecules
deviates significantly from the well delineated bell-shaped model [6]. Diffusion kurtosis imaging
offers a promising solution that captures non-Gaussian effects to address this limitation. DKI is a
mathematical extension of DTI and provides analogous metrics to those previously discussed from
the tensor model, i.e., KFA, MK, RK, and AK [120, 122].
To our current knowledge, this is the first study that uses the DKI model to understand the
effects of chronic anemia in the WM derived from congenital diseases. However, in demyelination
disease models, findings from other research groups indicate that DKI might exhibit greater sensitivity than diffusion coefficients in detecting alterations in the water exchange rate between two
compartments [6, 233]. This change could be closely associated with myelin integrity and water
exchange between intra- and extra-axonal spaces [120]. Therefore, it has been proposed that RK
metric is more effective than RD metric in capturing the extensive effects in demyelination models,
where an increase in intra- and extra-axonal water exchange is known to occur [6]. Additionally,
decreased MK in the normal-appearing WM has also been associated with describing demyelinating
Chap. 4 Chronic anemia and WM damage 75
diseases [233].
In our case, our results show that there is a massive decrease in MK in almost the entire WM,
surpassing those areas found by the MD metric, Fig. 4.2. Additionally, there is nearly a one-to-one
correspondence between those regions depicted in the MK maps and those appearing in the RK
maps, Fig. A.1. These findings further support the idea that CA patients might be undergoing
constant injury to the WM in the form of demyelination. However, we argue that the white matter
could also be undergoing a process of “constant repair”, such that, despite its limitations not many
changes are being depicted in the tensor model [234, 235].
Furthermore, the well-established mouse model, i.e., the cuprizone mouse, has been used to
study the processes of both demyelination and remyelination in the corpus callosum [236]. This
model has been coupled with DTI and DKI analysis to understand better the association between
dMRI markers and the actual undergoing pathology. In 2014, Falangola et al. [237] published a
modified version of the cuprizone mouse model that induced chronic demyelinated lesions. This
model revealed that the corpus callosum showed an increase in MD, RD, and AD. At the same time,
there was a decrease in MK, RK, and AK values. These results can be explained by the breakdown
and loss of myelin, leading to a reduction in structural complexity and diffusion barriers in the
corpus callosum microenvironment. Our results show that in addition to MK and RK extensively
decreasing, the WM areas where AK also reduces are big. Interestingly, the RK and AK do not
have a one-to-one correspondence; however, they colocalize enough to assume the overall results
depict traces of undergoing demyelination.
4.4.1.3 Neurite orientation dispersion and density imaging in CA
NODDI is a model that uses a multi-shell protocol to provide a more precise characterization
of tissue microstructure. The model identifies three tissue compartments, each modeled in a
biologically informed manner: intra-, extra-neurite, and free water. Three metrics called NDI,
ODI, and FISO are calculated from the diffusion signal in these compartments. Particularly, two
Chap. 4 Chronic anemia and WM damage 76
of them, neurite density and orientation dispersion, have been strongly correlated with histological
measures [238, 239]. NDI aligns well with optical myelin staining intensity [240], while ODI closely
corresponds to quantitative Golgi analysis in the brain [241].
To our current knowledge, there is only one research paper by Stotesbury et al. (2018) [156]
that in a smaller cohort compared NDI and ODI in SCD patients with healthy individuals. The
study found that NDI decreases in SCD patients, but there is no change in ODI. Moreover, the
results were consistent with the patterns of demyelination predicted by the tensor model, which
showed a decrease in FA and an increase in RD, with no change in axial diffusivity. It is worth
noting that there are no publications that have used the NODDI model in non-SCD patients.
Additionally, in the cuprizone mouse model, which is a well-established demyelination model,
Wang et al. (2019) [239] identified two different stages of demyelination - acute demyelination (5
weeks cuprizone administration) and chronic demyelination (13 weeks cuprizone administration) - to
investigate the variations in ODI and NDI. Demyelination was confirmed in both acute and chronic
stages using immunostaining (Luxol Fast Blue to stain myelin). The results in the corpus callosum
were consistent for ODI after 5 weeks and 13 weeks of cuprizone administration. However, the
NDI was lower than the surrounding gray matter in acute demyelination but elevated in chronic
demyelination mice. Furthermore, measurements from 332 ROIs (166 in each hemisphere) from
three different dMRI protocols showed that the correlation between FA and ODI is generally
negative, similar to the correlation between MD and NDI.
Our recent research has revealed that the ODI maps exhibit both increase and decrease values,
as shown in the left three columns of Fig. A.4. Even though the FA maps in Fig. 4.1 (left
three columns) do not correspond one-to-one with the ODI maps, it is evident that there is a
negative correlation between the FA and ODI maps, particularly in the watershed areas and the
corpus callosum as shown in the demyelination mouse model [239]. These patterns were observed
in both patient cohorts; however, in non-SCD the predominant trend was an increase in ODI.
Chap. 4 Chronic anemia and WM damage 77
Although these values do not agree with the no significant changes in ODI described by Stotesbury
et al. (2018) [156], there is also no clear consensus in the literature about the direction of ODI in
demyelination models. ODI has been reported to increase, decrease, or show no significant changes
[8].
Furthermore, there is a better agreement in demyelination diseases that NDI is reduced in MRvisible lesions and, to a lesser degree, in normal-appearing WM, revealing prominent and widespread
abnormalities than FA [8, 242, 243]. In our NDI results, we observe extensive decrements covering
almost the entire WM, for the log age and sex regressed model, that also agree with our MK results
(even for the smaller clusters of increased values), see Figs. 4.2 and 4.4. Interestingly, we observed
the same patterns when also regressing the hemoglobin values. Additionally, in agreement with the
mouse literature, the NDI appears negatively correlated with the MD maps, see Figs. 4.2 and 4.4.
Lastly, the free water fraction (FISO) results are displayed in the right three columns of Figure
A.4. It is interesting to note that only in the case of SCD patients, the voxels with increased FISO
are also colocalized with most of the decreased MD, where we controlled only for sex and log age
(Figure 4.2, left superior panel). Additionally, the voxels with decreased FISO are also colocalized
with most of the decreased NDI (and also MK) maps, where we controlled for hemoglobin (Figure
4.4 and Figure 4.2, left superior panel). However, the results for non-SCD patients do not show
the same association as the SCD patients nor exhibit a clear association with the kurtosis or the
tensor metrics. This could indicate that the FISO metric captures intrinsic changes unique to SCD
patients.
4.4.2 Mechanisms of WM Injury
According to the literature, independently of the cause of chronic anemia, the whole brain increases
cerebral blood flow (CBF) to compensate for the loss of oxygen-carrying capacity [1, 2, 159, 161,
188]. This offset in CBF preserves total resting oxygen delivery to the whole brain [1, 3, 24, 161],
such that the correspondent oxygen extraction fraction (OEF) from the cerebral cortex seems to be
Chap. 4 Chronic anemia and WM damage 78
normal (or even reduced) [89, 162, 189, 190, 192]. Consequently, the overall GM volume seems to
remain unaffected [110] in clinically asymptomatic individuals with SCD, particularly adolescents
and young adults, who receive chronic transfusion or hydroxyurea treatments (although there have
been documented instances of diminished volumes in the cortex or subcortical regions among
patients who are not undergoing these treatments) [112, 244, 108]. This suggests that cortical
volume reductions may not become apparent until later stages of life.
Despite this compensation mechanism, the cerebral vascular reserve decreases proportionately
to the resting hyperemia. Hence, acute insults such as nighttime hypoxia, severe anemia, and fever
can significantly threaten the brain. In fact, it appears that regions within the deep WM do not
exhibit the same degree of metabolic stability when faced with periods of ischemic stress.
In CA, regardless of the disease state, CBF and oxygen delivery to deep WM regions and
border zones are notably lower than those measured in other WM and GM areas [3]. An impaired
cerebrovascular reserve further exacerbates this insufficient oxygen delivery at rest in the WM [4].
Specifically, there is an increase in OEF [188] corresponding to those areas of WM where injury
patterns have been reported [3, 5, 158]. Consequently, prolonged inadequate perfusion contributes
to the formation of the WM damage phenotype. This encompasses features such as T2 FLAIR
hyperintensities [3, 5, 157, 158, 193], diminished WM volume [21, 103, 110, 111], or alterations in
diffusion metrics [20, 143, 144, 145, 146]. Furthermore, it has also been proposed that increase in
CBF could be associated with impairments in WM even in the absence of infarcts [146].
In this study, we discovered a distinct pathophysiology that appears to be associated with
demyelination processes that have a significant impact on the white matter and do not seem to
be driven by hypoxia. This phenomenon was also recently described by Hazra et al. (2023) [245]
in a Townes’ sickle mice model. In this study, the authors discovered that white-matter injury
with demyelination of axons, detected by MRI, corresponded to neuroaxonal damage detected by
histology and cognitive dysfunction. In particular, they found that in SCD, astrocyte activation may
Chap. 4 Chronic anemia and WM damage 79
play a crucial role in developing neuroaxonal damage and cognitive dysfunction. This contribution
further links the neuroimaging findings to brain pathology and behavioral phenotype.
Additionally, the study conducted by Stotesbury et al. (2018) [156] on SCD patients provided
evidence that cognitive parameters are reduced due to diffuse changes in DTI and NODDI, regardless
of the presence of silent cerebral infarcts. The authors concluded that the loss of white matter
integrity is related to the degree of slower processing speed. However, until this point, this
phenomenon was not recognized as underlying demyelination.
The diffuse nature of the white matter disarrangements suggests that demyelination is most
likely driven by whole-brain neuroinflammation, not local neuroinflammation. An example of
general neuroinflammation has been observed in the prevalence of post-acute sequelae of SARSCoV2, also known as Long-COVID. This condition can result in persistent neuroinflammatory
changes and consequent dysregulation of neural cell types necessary for healthy cognitive function.
Survivors of SARS-Cov2 may experience symptoms related to the central nervous system, such as
fatigue, headache, sleep disturbance, inattention, forgetfulness, and brain fog [246, 247].
In a mouse model replicating Long-COVID symptoms [248], researchers found that neuroinflammatory alterations, specifically targeted microglial response in white matter, correlate with
hindered hippocampal neurogenesis, disrupted regulation of the oligodendroglial lineage (whose
functions include forming myelin sheaths that enwrap axons of the central nervous system), and
loss of myelin.
Notably, neuroinflammation plays a crucial role in the pathophysiology of Long-covid. It
has also been observed in other viral infections such as H1N1 influenza and is associated with
lasting changes following the use of certain drugs such as methotrexate (used for chemotherapy)
[249, 250]. In these other neuroinflammatory conditions, research has shown that reactive microglia
and brain-infiltrating macrophages hinder the mechanisms responsible for maintaining cellular
Chap. 4 Chronic anemia and WM damage 80
homeostasis and plasticity (generation of myelin-forming oligodendrocytes, plasticity of myelin, and
the generation of new neurons in the hippocampus) [249, 251, 252]. In this case, the dysregulation
is partly caused by local microglial cytokine secretion, leading to cognitive impairment and reduced
neurogenesis due to high circulating cytokine/chemokine levels, especially CCL11 [253]. In addition
to the direct effects of inflammatory mediators on the brain’s cellular plasticity, microglia also cause
neurotoxic astrocyte reactivity through cytokine signaling [250]. Astrocytes can take on various
reactive states [254, 255], leading to further pathophysiology and even oligodendrocytes and neuron
death [250].
In addition to the mechanisms previously described as consequences of “pure” inflammation,
in sickle cell disease, free heme, an iron-containing molecule found in the body when hemoglobin
breaks down, can be linked to further neurotoxicity [46]. When released from damaged red blood
cells or due to the breakdown of heme-containing proteins, free heme can lead to oxidative stress
and additionally increase inflammation, both of which can harm neurons in the brain [256]. This
also applies to Thalassemia syndromes, although the rate of blood cell destruction is not as rapid
as in SCD [257, 258].
4.5 Limitations
The data collected for this study was a collaborative effort between two research facilities Children’s
Hospital Los Angeles and the Amsterdam University Medical Center in The Netherlands. As such,
the sample population between centers was not homogenous, i.e., the SCD population of patients
in Los Angeles has an ethnicity of Hispanic and African descent. In contrast, the Amsterdam
population is only based on African descendants. This variability helps use hemoglobin as a
continuous regressor (patients have different degrees of anemia) but could also introduce other
factors for which this study did not account. For example, it has been shown that the severity,
complications, and comorbidities of SCD varies with races [259].
It should be noted that most of the control subjects were recruited from family members of SCD
patients. However, it is possible that these populations do not accurately represent the non-SCD
Chap. 4 Chronic anemia and WM damage 81
population. As a result, any statistical differences found in non-SCD patients (even when accounting
for factors such as age and sex) could potentially be influenced by other factors.
Although previous studies have indicated significant sex differences in the study of CA, we
could not fully resolve sex-specific disease differences due to the smaller size of the non-SCD
compared to the SCD group. For the same reason (sample size), we did not analyze direct
differences between SCD and non-SCD patients. The use of data from multiple cohorts of CA, whose
intrinsic pathophysiology is different, opens the possibility to differentiate and characterize the
unique damage induced by individual hemoglobinopathies and it would be of interest to statistically
quantify their differences.
4.6 Conclusion
We conducted a study to investigate the impact of chronic anemia on white matter. To achieve
this, we analyzed diffusion kurtosis and neurite orientation dispersion and density imaging, building
on previous research that used diffusion tensor imaging. Our study compared healthy individuals
with patients suffering from sickle and non-sickle anemias, allowing us to isolate the effects of sickle
hemoglobin in our analysis.
In the DKI model, we found massive effects of decreased MK, RK, and AK in all the white
matter that could be associated with underlying demyelination processes that are a consequence
of chronic anemia, regardless of whether patients are classified as SCD or non-SCD. Additionally,
the NODDI model showed similar effects with an extensive decrease in NDI, further supporting the
idea of underlying demyelination. In particular, all the observed WM disagreements were present
not only in the watershed areas but also in areas closer to the gray matter; in both cases, bilateral
and spatially symmetrical.
Furthermore, we identified significant differences in the WM of only SCD patients when hemoglobin
(the driving factor in anemia) was regressed out from both the kurtosis and NODDI metrics. This
Chap. 4 Chronic anemia and WM damage 82
model might delineate other afflictions to the WM caused by unique factors of sickle cell disease.
This information expands our current understanding of the affliction that chronic anemia produces
in the white matter.
Chapter 5
Conclusion of PhD Work
“One never notices what has been done; one can
only see what remains to be done.”
— Maria Salomea Sk lodowska-Curie
This dissertation examines the impact of chronic anemia, resulting from abnormalities in hemoglobin synthesis or structure, on developing hypoxia, neuroinflammation, and the subsequent
remodeling of white matter. This research demonstrates that the damage to WM attributable
to chronic anemia extends significantly beyond the watershed regions typically associated with
hypoxic injury.
In Chapter Two, we provided a medical definition of anemia and discussed the pathophysiology
of chronic anemia from the perspective of hereditary conditions. Specifically, we outlined the clinical
descriptions, the complications, and therapeutic approaches for thalassemia syndromes and sickle
cell disease. Furthermore, we discussed the dependency of the oxygen delivered to the brain on the
circulating blood flow, leading to the discussion of the cerebral mechanisms of autoregulation. In
CA patients, it has been observed that the overall oxygen delivery to the brain remains preserved
due to increased blood flow circulating through the brain. The brain possesses an impressive
capacity to augment blood flow in response to decreased oxygen levels, a phenomenon referred
to as cerebrovascular reactivity. In cases of severe anemia, however, this capacity is significantly
83
Chap. 4 Chronic anemia and WM damage 84
impaired, rendering the brain more vulnerable to injuries, particularly in the WM watershed areas.
Due to their anatomical positioning, these regions are the first to be adversely affected when oxygen
supply is insufficient. Consequently, researchers have long posited that, in chronic anemia, most, if
not all, of the white matter injuries are confined to the watershed regions.
To deepen our understanding of white matter injuries related to hypoxia, we focused on analyzing
diffusion MRI data from CA patients and compared it to data from healthy controls. We specifically
selected dMRI because it is particularly effective in investigating abnormalities in white matter.
In Chapter Three, we calculated the average fractional anisotropy along pathways connecting
two specific regions of interest based on the BCI-DNI brain atlas from BrainSuite. Our findings
showed that patients with anemia have lower FA values, which suggested a loss of coherence in the
predominant direction of diffusion. This implies injury on the WM tracks connecting such regions.
We also found a positive relationship between FA levels and hemoglobin concentrations, meaning
that chronically low oxygen levels greatly affect the WM microstructural integrity. Additionally,
our analysis indicated that non-SCD patients have more significant reductions in FA, especially in
intrahemispheric WM pathways and watershed areas. In contrast, SCD patients primarily exhibit
reductions in interhemispheric connectivity.
Furthermore, we seek to identify alterations in WM that can be directly linked to anemia,
distinguishing them from changes attributable to disease-specific pathophysiological processes. In
Chapter Four, we built upon earlier studies by conducting analyses using multi-shell imaging
techniques, such as diffusion kurtosis imaging, which effectively captures the constraints imposed by
tissue architecture, and the neurite orientation dispersion and density imaging model, which offers
a biologically informed understanding of WM tissue. These innovative methodologies facilitated
voxel-wise analyses, allowing for a comparison between patients with CA and healthy control
subjects, resulting in the identification of three significant findings:
• Patients diagnosed with chronic anemia demonstrate significant changes in WM, regardless
of the underlying causes of their condition, including both sickle cell anemia and non-sickle
Chap. 4 Chronic anemia and WM damage 85
cell anemia. The primary determinant of WM damage is the severity of the anemia. These
alterations extend beyond the damage typically associated with hypoxia, and the observed
distribution and diffusion characteristics strongly suggest the occurrence of demyelination.
• Our analysis demonstrated a significant correlation between processing speed index and MK
in regions predominantly affected by WM abnormalities. Although MK presented with diffuse
anomalies, the correlations between MK and PSI were most pronounced in watershed areas
and key anatomical structures, such as the corpus callosum, cingulum, and central regions
of the frontal, temporal, parietal, and occipital lobes. These regions play a crucial role in
processing speed performance. The observed associations may be attributed, in part, to their
specific anatomical localization, as well as the degree of demyelination, given that myelin
sheath thickness is associated with nerve conduction velocity.
• We have identified blood markers of lactate dehydrogenase and absolute reticulocyte count to
elucidate additional damage and processes occurring in the WM of patients with SCD that
are not significantly observed in non-SCD patients. Deep WM hypoxia is more pronounced
in SCD patients than in other forms of anemia, even when accounting for hemoglobin levels.
Intravascular hemolysis is notably more significant in individuals with SCD, which correlates
with adverse cardiovascular, cerebrovascular, and overall health outcomes. The products of
hemolysis are well known to have pro-inflammatory properties through various biochemical
mechanisms and may also contribute to demyelination. While the proposed mechanisms
remain speculative, they are consistent with findings indicating that hemoglobin is the strongest
predictor of demyelination.
While this dissertation demonstrates the novel observation of diffuse demyelination, it is critical to
realize that these WM changes are superimposed on well-documented chronic ischemic insults in
deep watershed areas. Thus, our work does not claim that the diffuse demyelination observed in
this cohort is more important than the known ischemic white matter disease previously documented
in chronically anemic subjects; it only occurs in parallel and represents a distinct pathophysiology.
We hope this research will encourage further investigation of neuroinflammatory mediators, with
Chap. 4 Chronic anemia and WM damage 86
the expectation that such insights will contribute to the development of preventive treatments for
patients suffering from chronic anemia.
Bibliography
[1] A. M. Bush, M. T. Borzage, S. Choi, L. V´aclav˚u, B. Tamrazi, A. J. Nederveen, T. D. Coates,
and J. C. Wood, “Determinants of resting cerebral blood flow in sickle cell disease,” American
Journal of Hematology, vol. 91, no. 9, pp. 912–917, 2016.
[2] C. Vu, A. Bush, S. Choi, M. Borzage, X. Miao, A. J. Nederveen, T. D. Coates, and J. C. Wood,
“Reduced global cerebral oxygen metabolic rate in sickle cell disease and chronic anemias,”
American Journal of Hematology, vol. 96, no. 8, pp. 901–913, 2021.
[3] Y. Chai, A. M. Bush, J. Coloigner, A. J. Nederveen, B. Tamrazi, C. Vu, S. Choi, T. D.
Coates, N. Lepore, and J. C. Wood, “White matter has impaired resting oxygen delivery in
sickle cell patients,” American Journal of Hematology, vol. 94, no. 4, pp. 467–474, 2019.
[4] L. Afzali-Hashemi, K. P. A. Baas, A. Schrantee, B. F. Coolen, M. J. P. van Osch, S. M. Spann,
E. Nur, J. C. Wood, B. J. Biemond, and A. J. Nederveen, “Impairment of cerebrovascular
hemodynamics in patients with severe and milder forms of sickle cell disease,” Frontiers in
Physiology, vol. 12, 2021.
[5] A. L. Ford, D. K. Ragan, S. Fellah, M. M. Binkley, M. E. Fields, K. P. Guilliams, H. An,
L. C. Jordan, R. C. McKinstry, J.-M. Lee, and M. R. DeBaun, “Silent infarcts in sickle cell
disease occur in the border zone region and are associated with low cerebral blood flow,”
Blood, vol. 132, no. 16, pp. 1714–1723, 10 2018.
[6] A. J. Steven, J. Zhuo, and E. R. Melhem, “Diffusion kurtosis imaging: An emerging
technique for evaluating the microstructural environment of the brain,” American Journal
of Roentgenology, vol. 202, no. 1, pp. W26–W33, 2014.
[7] I. O. Jelescu and M. D. Budde, “Design and validation of diffusion mri models of white
matter,” Frontiers in physics, vol. 5, p. 61, 2017.
[8] K. Kamiya, M. Hori, and S. Aoki, “Noddi in clinical research,” Journal of Neuroscience
Methods, vol. 346, p. 108908, 2020.
[9] N. J. Kassebaum, “The global burden of anemia,” Hematology and oncology clinics of North
America, vol. 30, no. 2, pp. 247–308, 2016.
[10] WHO, “The global prevalence of anaemia in 2011,” Geneva: World Health Organization,
2015.
[11] C. M. Chaparro and P. S. Suchdev, “Anemia epidemiology, pathophysiology, and etiology in
low- and middle-income countries,” Annals of the New York Academy of Sciences, vol. 1450,
no. 1, pp. 15–31, 2019.
87
Bibliography 88
[12] A. K. Y. Tsui, P. A. Marsden, C. D. Mazer, J. G. Sled, K. M. Lee, R. M. Henkelman, L. S.
Cahill, Y.-Q. Zhou, N. Chan, E. Liu, and G. M. T. Hare, “Differential hif and nos responses
to acute anemia: defining organ-specific hemoglobin thresholds for tissue hypoxia,” American
Journal of Physiology-Regulatory, Integrative and Comparative Physiology, vol. 307, no. 1, pp.
13–25, 2014.
[13] F. J. Wolters, H. I. Zonneveld, S. Licher, L. G. Cremers, M. K. Ikram, P. J. Koudstaal,
M. W. Vernooij, and M. A. Ikram, “Hemoglobin and anemia in relation to dementia risk and
accompanying changes on brain mri,” Neurology, vol. 93, no. 9, pp. 917–926, 2019.
[14] M. Angastiniotis, A. Eleftheriou, R. Galanello, C. Harteveld, M. Petrou, and J. TraegerSynodinos, “Chapter 2 epidemiology of hemoglobinopathies,” in Prevention of Thalassaemias
and Other Haemoglobin Disorders: Volume 1, 2nd ed. Principles. Nicosia: Thalassaemia
International, 2013.
[15] A. Wahed and A. Dasgupta, “Chapter 4 hemoglobinopathies and thalassemias,” in
Hematology and Coagulation. San Diego: Elsevier, 2015, pp. 55–80.
[16] A. B. Payne, J. M. Mehal, C. Chapman, D. L. Haberling, L. C. Richardson, C. J. Bean, and
W. C. Hooper, “Trends in sickle cell disease–related mortality in the united states, 1979 to
2017,” Annals of Emergency Medicine, vol. 76, no. 3, pp. 28–36, 2020.
[17] B. Modell, M. Darlison, H. Birgens, H. Cario, P. Faustino, P. C. Giordano, B. Gulbis,
P. Hopmeier, D. Lena-Russo, L. Romao, and E. Theodorsson, “Epidemiology of haemoglobin
disorders in europe: an overview,” Scandinavian Journal of Clinical and Laboratory
Investigation, vol. 67, no. 1, pp. 39–70, 2007.
[18] M. Case, H. Zhang, J. Mundahl, Y. Datta, S. Nelson, K. Gupta, and B. He, “Characterization
of functional brain activity and connectivity using eeg and fmri in patients with sickle cell
disease,” NeuroImage: Clinical, vol. 14, pp. 1–17, 2017.
[19] Z. Metafratzi, M. I. Argyropoulou, D. N. Kiortsis, C. Tsampoulas, N. Chaliassos, and S. C.
Efremidis, “T2 relaxation rate of basal ganglia and cortex in patients with beta-thalassaemia
major,” The British Journal of Radiology, vol. 74, no. 881, pp. 407–410, 2001.
[20] B. Sun, R. C. Brown, L. Hayes, T. G. Burns, J. Huamani, D. J. Bearden, and R. A. Jones,
“White matter damage in asymptomatic patients with sickle cell anemia: Screening with
diffusion tensor imaging,” AJNR Am. J. Neuroradiol., vol. 33, no. 11, p. 2043–2049, 2012.
[21] S. Choi, R. M. Leahy, and J. C. Wood, “Lower white matter volume in beta-thalassemia
associated with anemia and cognitive performance,” American journal of hematology, vol. 95,
no. 6, p. 144—146, June 2020.
[22] A. Usmani and R. F. Machado, “Vascular complications of sickle cell disease,” Clinical
hemorheology and microcirculation, vol. 68, pp. 205–221, 2018.
[23] A. G. Russo, S. Ponticorvo, I. Tartaglione, M. Caiazza, D. Roberti, A. Elefante, M. Casale,
R. Di Concilio, A. Ciancio, E. De Michele, A. Canna, M. Cirillo, S. Perrotta, F. Esposito, and
R. Manara, “No increased cerebrovascular involvement in adult beta-thalassemia by advanced
mri analyses,” Blood Cells, Molecules, and Diseases, vol. 78, pp. 9–13, 2019.
Bibliography 89
[24] M. T. Borzage, A. M. Bush, S. Choi, A. J. Nederveen, L. V´aclav˚u, T. D. Coates, and J. C.
Wood, “Predictors of cerebral blood flow in patients with and without anemia,” Journal of
Applied Physiology, vol. 120, no. 8, pp. 976–981, 2016.
[25] C. Baldwin, J. Pandey, and O. Olarewaju, in Hemolytic Anemia. Treasure Island, Florida:
StatPearls Publishing, 2023.
[26] M. J. Cascio and T. G. DeLoughery, “Anemia: Evaluation and diagnostic tests,” Medical
Clinics of North America, vol. 101, no. 2, pp. 263–284, 2017.
[27] WHO, “Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity,”
Geneva: World Health Organization, 2011.
[28] M. Johnson-Spear and R. Yip, “Hemoglobin difference between black and white women with
comparable iron status: justification for race-specific anemia criteria,” The American Journal
of Clinical Nutrition, vol. 60, no. 1, pp. 117–121, 1994.
[29] E. Beutler and J. Waalen, “The definition of anemia: what is the lower limit of normal of the
blood hemoglobin concentration?” Blood, vol. 107, no. 5, pp. 1747–1750, 2006.
[30] C. H. H. Le, “The prevalence of anemia and moderate-severe anemia in the us population
(nhanes 2003-2012),” PLOS ONE, vol. 11, no. 11, pp. 1–14, 2016.
[31] M. A. Atkinson, M. L. Melamed, J. Kumar, C. N. Roy, I. Miller, Edgar R., S. L. Furth, and
J. J. Fadrowski, “Vitamin d, race, and risk for anemia in children,” The Journal of Pediatrics,
vol. 164, no. 1, pp. 153–158, 2014.
[32] “Anemia or iron deficiency,” https://www.cdc.gov/nchs/fastats/anemia.html, accessed: 2023-
03-28.
[33] V. Jansen, “Diagnosis of anemia: A synoptic overview and practical approach,” Transfusion
and Apheresis Science, vol. 58, no. 4, pp. 375–385, 2019.
[34] B. J. Gordon, P. DeSaix, E. Johnson, J. E. Johnson, O. Korol, D. H. Kruse, B. Poe, J. A.
Wise, M. Womble, and K. A. Young, in Anatomy and Physiology. Houston, Texas: OpenStax
College Rice University, 2022.
[35] M. Badireddy and K. M. Baradhi, in Chronic Anemia. Treasure Island, Florida: StatPearls
Publishing, 2022.
[36] M. M. Okam, “Anemia and hemoglobinopathies,” in The Brigham Intensive Review of
Internal Medicine. Oxford University Press, 07 2014.
[37] J. C. Sun and H. V. Joffe, “Chapter 9 - anemia,” in The most common inpatient problems in
internal medicine. Philadelphia: W.B. Saunders, 2007, pp. 179–198.
[38] V. G. Sankaran and S. H. Orkin, “The switch from fetal to adult hemoglobin,” Cold Spring
Harbor Perspectives in Medicine, vol. 3, no. 1, 2013.
[39] L. Manca and B. Masala, “Disorders of the synthesis of human fetal hemoglobin,” IUBMB
Life, vol. 60, no. 2, pp. 94–111, 2008.
Bibliography 90
[40] R. S. Franco, “Measurement of red cell lifespan and aging,” Transfusion medicina and
hemotherapy, vol. 39, no. 5, p. 302–307, 2012.
[41] M. qing Lei, L. feng Sun, X. sheng Luo, X. yang Yang, F. Yu, X. xia Chen, and Z. ming
Wang, “Distinguishing iron deficiency anemia from thalassemia by the red blood cell lifespan
with a simple co breath test: a pilot study,” Journal of Breath Research, vol. 13, no. 2, p.
026007, mar 2019.
[42] P. Gillette, J. Manning, and A. Cerami, “Increased survival of sickle-cell erythrocytes after
treatment in vitro with sodium cyanate,” Proceedings of the National Academy of Sciences
of the United States of America, vol. 68, no. 11, p. 2791—2793, November 1971.
[43] P. F. Milner and S. Charache, “Life span of carbamylated red cells in sickle cell anemia,” The
Journal of clinical investigation, vol. 52, no. 12, pp. 3161–3171, December 1973.
[44] P. R. McCurdy and A. S. Sherman, “Irreversibly sickled cells and red cell survival in sickle
cell anemia: a study with both df32p and 51cr,” The American Journal of Medicine, vol. 54,
no. 2, pp. 253–258, 1978.
[45] R. P. Rother, L. Bell, P. Hillmen, and M. T. Gladwin, “The Clinical Sequelae of Intravascular
Hemolysis and Extracellular Plasma HemoglobinA Novel Mechanism of Human Disease,”
JAMA, vol. 293, no. 13, pp. 1653–1662, 2005.
[46] O. T. Gbotosho, M. G. Kapetanaki, and G. J. Kato, “The worst things in life are free: The
role of free heme in sickle cell disease,” Frontiers in Immunology, vol. 11, 2021.
[47] G. Salinas Cisneros and S. L. Thein, “Recent advances in the treatment of sickle cell disease,”
Frontiers in Physiology, vol. 11, 2020.
[48] A. J. Marengo-Rowe, “The thalassemias and related disorders,” Baylor University Medical
Center Proceedings, vol. 20, no. 1, pp. 27–31, 2007.
[49] A. Zivot, J. M. Lipton, A. Narla, and L. Blanc, “Erythropoiesis: insights into pathophysiology
and treatments in 2017,” Molecular Medicine, vol. 24, no. 1, 2018.
[50] M. J. Herbert L and J. Campbell, “Alpha and beta thalassemia,” Am. Fam. Physician, vol. 80,
no. 4, pp. 339–344, 2009.
[51] G. Lucarelli, A. Isgr`o, P. Sodani, and J. Gaziev, “Hematopoietic stem cell transplantation
in thalassemia and sickle cell anemia,” Cold Spring Harbor Perspectives in Medicine, vol. 2,
no. 5, 2012.
[52] J. V. Raja, M. A. Rachchh, and R. H. Gokani, “Recent advances in gene therapy for
thalassemia,” Journal of Pharmacy and Bioallied Sciences, vol. 4, no. 3, pp. 194–201, 2012.
[53] A. A. Thompson, M. C. Walters, J. Kwiatkowski, J. E. Rasko, J.-A. Ribeil, S. Hongeng,
E. Magrin, G. J. Schiller, E. Payen, M. Semeraro, D. Moshous, F. Lefrere, H. Puy, P. Bourget,
A. Magnani, L. Caccavelli, J.-S. Diana, F. Suarez, F. Monpoux, V. Brousse, C. Poirot,
C. Brouzes, J.-F. Meritet, C. Pondarr´e, Y. Beuzard, S. Chr´etien, T. Lefebvre, D. T. Teachey,
U. Anurathapan, P. J. Ho, C. von Kalle, M. Kletzel, E. Vichinsky, S. Soni, G. Veres,
Bibliography 91
O. Negre, R. W. Ross, D. Davidson, A. Petrusich, L. Sandler, M. Asmal, O. Hermine,
M. De Montalembert, S. Hacein-Bey-Abina, S. Blanche, P. Leboulch, and M. Cavazzana,
“Gene therapy in patients with transfusion-dependent β-thalassemia,” New England Journal
of Medicine, vol. 378, no. 16, pp. 1479–1493, 2018.
[54] I. Motta, R. Bou-Fakhredin, A. T. Taher, and M. D. Cappellini, “Beta thalassemia: New
therapeutic options beyond transfusion and iron chelation,” Drugs, vol. 80, no. 11, pp. 1053–
1063, 2020.
[55] R. Feng, T. Mayuranathan, P. Huang, P. A. Doerfler, Y. Li, Y. Yao, J. Zhang, L. E. Palmer,
K. Mayberry, G. E. Christakopoulos, P. Xu, C. Li, Y. Cheng, G. A. Blobel, M. C. Simon,
and M. J. Weiss, “Activation of γ-globin expression by hypoxia-inducible factor 1α,” Nature,
vol. 610, no. 7933, pp. 783–790, 2020.
[56] FDA, “Reblozyl for injection, for subcutaneous use, initial u.s. approval: 2019,” https://
www.accessdata.fda.gov/drugsatfda docs/label/2019/761136lbl.pdf, accessed: 2023-04-08.
[57] S. Carrancio, J. Markovics, P. Wong, J. Leisten, P. Castiglioni, M. C. Groza, H. K. Raymon,
C. Heise, T. Daniel, R. Chopra, and V. Sung, “An activin receptor iia ligand trap promotes
erythropoiesis resulting in a rapid induction of red blood cells and haemoglobin,” British
Journal of Haematology, vol. 165, no. 6, pp. 870–882, 2014.
[58] D. Vela, “Hepcidin, an emerging and important player in brain iron homeostasis,” Journal of
Translational Medicine, vol. 16, no. 1, p. 25, 2018.
[59] C. Casu, E. Nemeth, and S. Rivella, “Hepcidin agonists as therapeutic tools,” Blood, vol. 131,
no. 16, pp. 1790–1794, 2018.
[60] C. Camaschella, A. Nai, and L. Silvestri, “Iron metabolism and iron disorders revisited in the
hepcidin era,” Haematologica, vol. 105, no. 2, pp. 260–272, 2020.
[61] M. H. Steinberg, M. H. Steinberg, B. G. Forget, D. R. Higgs, and D. J. Weatherall, SICKLE
CELL DISEASE, 2nd ed. Cambridge University Press, 2009, p. 435–436.
[62] A. Brunson, A. Lei, A. S. Rosenberg, R. H. White, T. Keegan, and T. Wun, “Increased
incidence of vte in sickle cell disease patients: risk factors, recurrence and impact on
mortality,” British Journal of Haematology, vol. 178, no. 2, pp. 319–326, 2017.
[63] R. E. Ware, M. de Montalembert, L. Tshilolo, and M. R. Abboud, “Sickle cell disease,” The
Lancet, vol. 390, no. 10091, pp. 311–323, 2017.
[64] E. Nader, M. Romana, and P. Connes, “The red blood cell—inflammation vicious circle in
sickle cell disease,” Frontiers in Immunology, vol. 11, 2020.
[65] F. Padilla, P. Bromberg, and W. Jensen, “The sickle-unsickle cycle: A cause of cell
fragmentation leading to permanently deformed cells,” Blood, vol. 41, no. 5, pp. 653–660,
1973.
[66] A. R. Zazulia, J. Markham, and W. J. Powers, “4 - cerebral blood flow and metabolism in
human cerebrovascular disease,” in Stroke, 5th ed. Saint Louis: W.B. Saunders, 2011, pp.
44–67.
Bibliography 92
[67] L. Sokoloff, “Chapter 2: The brain as a chemical machine,” in Neuronal-Astrocytic
Interactions, ser. Progress in Brain Research. Elsevier, 1992, vol. 94, pp. 19–33.
[68] K. Wailoo, “Sickle cell disease — a history of progress and peril,” New England Journal of
Medicine, vol. 376, no. 9, pp. 805–807, 2017.
[69] E. Gluckman, B. Cappelli, F. Bernaudin, M. Labopin, F. Volt, J. Carreras, B. Pinto Sim˜oes,
A. Ferster, S. Dupont, J. de la Fuente, J.-H. Dalle, M. Zecca, M. C. Walters, L. Krishnamurti,
M. Bhatia, K. Leung, G. Yanik, J. Kurtzberg, N. Dhedin, M. Kuentz, G. Michel, J. Apperley,
P. Lutz, B. Neven, Y. Bertrand, J. P. Vannier, M. Ayas, M. Cavazzana, S. Matthes-Martin,
V. Rocha, H. Elayoubi, C. Kenzey, P. Bader, F. Locatelli, A. Ruggeri, and M. Eapen, “Sickle
cell disease: an international survey of results of hla-identical sibling hematopoietic stem cell
transplantation,” Blood, vol. 129, no. 11, pp. 1548–1556, 2017.
[70] E. B. Esrick and D. E. Bauer, “Genetic therapies for sickle cell disease,” Seminars in
Hematology, vol. 55, no. 2, pp. 76–86, 2018, sickle Cell Disease- Unanswered Questions and
Future Directions in therapy.
[71] E. Vichinsky, C. C. Hoppe, K. I. Ataga, R. E. Ware, V. Nduba, A. El-Beshlawy, H. Hassab,
M. M. Achebe, S. Alkindi, R. C. Brown, D. L. Diuguid, P. Telfer, D. A. Tsitsikas,
A. Elghandour, V. R. Gordeuk, J. Kanter, M. R. Abboud, J. Lehrer-Graiwer, M. Tonda,
A. Intondi, B. Tong, and J. Howard, “A phase 3 randomized trial of voxelotor in sickle cell
disease,” New England Journal of Medicine, vol. 381, no. 6, pp. 509–519, 2019.
[72] C. T. Quinn, “L-glutamine for sickle cell anemia: more questions than answers,” Blood, vol.
132, no. 7, pp. 689–693, 2018.
[73] Y. Niihara, S. T. Miller, J. Kanter, S. Lanzkron, W. R. Smith, L. L. Hsu, V. R. Gordeuk,
K. Viswanathan, S. Sarnaik, I. Osunkwo, E. Guillaume, S. Sadanandan, L. Sieger, J. L. Lasky,
E. H. Panosyan, O. A. Blake, T. N. New, R. Bellevue, L. T. Tran, R. L. Razon, C. W. Stark,
L. D. Neumayr, and E. P. Vichinsky, “A phase 3 trial of l-glutamine in sickle cell disease,”
New England Journal of Medicine, vol. 379, no. 3, pp. 226–235, 2018.
[74] E. M. Agency, “Withdrawal assessment report: Xyndari,” https://www.ema.europa.eu/
en/documents/withdrawal-report/withdrawal-assessment-report-xyndari en.pdf, accessed:
2023-04-24.
[75] K. I. Ataga, A. Kutlar, J. Kanter, D. Liles, R. Cancado, J. a. Friedrisch, T. H. Guthrie,
J. Knight-Madden, O. A. Alvarez, V. R. Gordeuk, S. Gualandro, M. P. Colella, W. R. Smith,
S. A. Rollins, J. W. Stocker, and R. P. Rother, “Crizanlizumab for the prevention of pain
crises in sickle cell disease,” New England Journal of Medicine, vol. 376, no. 5, pp. 429–439,
2017.
[76] J. W. R. Sins, D. J. Mager, S. C. A. T. Davis, B. J. Biemond, and K. Fijnvandraat,
“Pharmacotherapeutical strategies in the prevention of acute, vaso-occlusive pain in sickle
cell disease: a systematic review,” Blood Advances, vol. 1, no. 19, pp. 1598–1616, 2017.
[77] D. C. Rees, Y. Kilinc, S. Unal, C. Dampier, B. S. Pace, B. Kaya, S. Trompeter, I. Odame, J. N.
Mahlangu, S. Unal, J. Brent, R. Grosse, B. R. Fuh, B. P. Inusa, A. Koren, C. Levin, S. Mortier,
Bibliography 93
E. McNamara, Y. Li, and S. J. Oliver, “Double-blind, randomized study of canakinumab
treatment in pediatric and young adult patients with sickle cell anemia,” Blood, vol. 134, no.
Supplement 1, pp. 615–615, 2019.
[78] E. Nur, D. P. Brandjes, T. Teerlink, H.-M. Otten, R. P. J. Oude Elferink, F. Muskiet, L. M.
Evers, H. ten Cate, B. J. Biemond, A. J. Duits, and J.-J. B. Schnog, “N-acetylcysteine reduces
oxidative stress in sickle cell patients,” Annals of Hematology, vol. 91, no. 7, pp. 1097–1105,
2012.
[79] D. D. Clarke and L. Sokoloff, “Chapter 31. circulation and energy metabolism of the brain,”
in Basic Neurochemistry: Molecular, Cellular and Medical Aspects, 6th ed. Philadelphia:
Lippincott-Raven, 1999.
[80] W. M. Armstead, “Cerebral blood flow autoregulation and dysautoregulation,” Anesthesiology
Clinics, vol. 34, no. 3, pp. 465–477, 2016.
[81] A. Silverman and N. H. Petersen, in Cerebral Autoregulation. Treasure Island, Florida:
StatPearls Publishing, 2023.
[82] E. Hamel, “Perivascular nerves and the regulation of cerebrovascular tone,” Journal of Applied
Physiology, vol. 100, no. 3, pp. 1059–1064, 2006.
[83] M. Yoshihara, K. Bandoh, and A. Marmarou, “Cerebrovascular carbon dioxide reactivity
assessed by intracranial pressure dynamics in severely head injured patients”, journal =
”journal of neurosurgery,” vol. 82, no. 3, pp. 386 – 393, 1995.
[84] I. Bulli, I. Dettori, E. Coppi, F. Cherchi, M. Venturini, L. Di Cesare Mannelli, C. Ghelardini,
A. Nocentini, C. T. Supuran, A. M. Pugliese, and F. Pedata, “Role of carbonic anhydrase
in cerebral ischemia and carbonic anhydrase inhibitors as putative protective agents,”
International Journal of Molecular Sciences, vol. 22, no. 9, 2021.
[85] S. Godo and H. Shimokawa, “Endothelial functions,” Arteriosclerosis, Thrombosis, and
Vascular Biology, vol. 37, no. 9, pp. 108–114, 2017.
[86] J. W. Ashby and J. J. Mack, “Endothelial control of cerebral blood flow,” The American
Journal of Pathology, vol. 191, no. 11, pp. 1906–1916, 2021.
[87] N. Forkert, M. Li, R. Lober, and K. Yeom, “Gray matter growth is accompanied by increasing
blood flow and decreasing apparent diffusion coefficient during childhood,” American Journal
of Neuroradiology, vol. 37, no. 9, pp. 1738–1744, 2016.
[88] E. Courchesne, H. J. Chisum, J. Townsend, A. Cowles, J. Covington, B. Egaas, M. Harwood,
S. Hinds, and G. A. Press, “Normal brain development and aging: Quantitative analysis at
in vivo mr imaging in healthy volunteers,” Radiology, vol. 216, no. 3, pp. 672–682, 2000.
[89] M. R. Juttukonda and M. J. Donahue, “Neuroimaging of vascular reserve in patients with
cerebrovascular diseases,” NeuroImage, vol. 187, pp. 192–208, 2019.
[90] L. V´aclav˚u, B. N. Meynart, H. J. Mutsaerts, E. T. Petersen, C. B. Majoie, E. T. VanBavel,
J. C. Wood, A. J. Nederveen, and B. J. Biemond, Haematologica, no. 4, pp. 690–699, 2019.
Bibliography 94
[91] F. Kirkham, D. Hewes, M. Prengler, A. Wade, R. Lane, and J. Evans, The Lancet, vol. 357,
no. 9269, pp. 1656–1659, 2001.
[92] I. Prohovnik, A. Hurlet-Jensen, R. Adams, D. D. Vivo, and S. G. Pavlakis, “Hemodynamic
etiology of elevated flow velocity and stroke in sickle-cell disease,” Journal of Cerebral Blood
Flow & Metabolism, vol. 29, no. 4, pp. 803–810, 2009.
[93] I. Momjian-Mayor and J.-C. Baron, “The pathophysiology of watershed infarction in internal
carotid artery disease,” Stroke, vol. 36, no. 3, pp. 567–577, 2005.
[94] A. Torvik, “The pathogenesis of watershed infarcts in the brain.” Stroke, vol. 15, no. 2, pp.
221–223, 1984.
[95] R. J. Adams, F. T. Nichols, V. McKie, K. McKie, P. Milner, and T. E. Gammal, “Cerebral
infarction in sickle cell anemia,” vol. 38, no. 7, pp. 1012–1012, 1988.
[96] H. Stotesbury, J. M. Kawadler, P. W. Hales, D. E. Saunders, C. A. Clark, and F. J.
Kirkham, “Vascular instability and neurological morbidity in sickle cell disease: An integrative
framework,” Frontiers in Neurology, vol. 10, 2019.
[97] D. Zhang, C. Xu, D. Manwani, and P. S. Frenette, “Neutrophils, platelets, and inflammatory
pathways at the nexus of sickle cell disease pathophysiology,” Blood, vol. 127, no. 7, pp.
801–809, 02 2016.
[98] R. P. Hebbel, R. Osarogiagbon, and D. Kaul, “The endothelial biology of sickle cell disease:
Inflammation and a chronic vasculopathy,” Microcirculation, vol. 11, no. 2, pp. 129–151, 2004.
[99] F. J. Kirkham and I. A. Lagunju, “Epidemiology of stroke in sickle cell disease,” Journal of
Clinical Medicine, vol. 10, no. 18, 2021.
[100] A. Kanavaki, K. Spengos, M. Moraki, P. Delaporta, C. Kariyannis, I. Papassotiriou, and
A. Kattamis, “Serum levels of s100b and nse proteins in patients with non-transfusiondependent thalassemia as biomarkers of brain ischemia and cerebral vasculopathy,”
International Journal of Molecular Sciences, vol. 18, no. 12, 2017.
[101] M. Hashemieh and N. Jafari, “Vascular brain damage in thalassemia syndrome: An emerging
challenge: Stroke in thalassemia syndrome,” Iranian Journal of Child Neurology, vol. 16,
no. 1, pp. 19–29, 2022.
[102] I. Pazgal, E. Inbar, M. Cohen, O. Shpilberg, and P. Stark, “High incidence of silent cerebral
infarcts in adult patients with beta thalassemia major,” Thrombosis Research, vol. 144, pp.
119–122, 2016.
[103] S. Choi, S. H. O’Neil, A. A. Joshi, J. Li, A. M. Bush, T. D. Coates, R. M. Leahy, and J. C.
Wood, “Anemia predicts lower white matter volume and cognitive performance in sickle and
non-sickle cell anemia syndrome,” American Journal of Hematology, vol. 94, pp. 1055–1065,
2019.
[104] M. Arkuszewski, J. Krejza, R. Chen, R. Ichord, J. L. Kwiatkowski, M. Bilello, R. Zimmerman,
K. Ohene-Frempong, and E. R. Melhem, “Sickle cell anemia: Intracranial stenosis and silent
Bibliography 95
cerebral infarcts in children with low risk of stroke,” Advances in Medical Sciences, vol. 59,
no. 1, pp. 108–113, 2014.
[105] J. F. Casella, A. A. King, B. Barton, D. A. White, M. J. Noetzel, R. N. Ichord, C. Terrill,
D. Hirtz, R. C. McKinstry, J. J. Strouse, T. H. Howard, T. D. Coates, C. P. Minniti, A. D.
Campbell, B. A. Vendt, H. Lehmann, and M. R. DeBaun, “Design of the silent cerebral
infarct transfusion (sit) trial,” Pediatric Hematology and Oncology, vol. 27, no. 2, pp. 69–89,
2010.
[106] H. Stotesbury, J. M. Kawadler, D. E. Saunders, and F. J. Kirkham, “Mri detection of brain
abnormality in sickle cell disease,” Expert Review of Hematology, vol. 14, no. 5, pp. 473–491,
2021.
[107] V. Gulani and N. Seiberlich, “Quantitative mri: Rationale and challenges,” in Quantitative
Magnetic Resonance Imaging, ser. Advances in Magnetic Resonance Technology and
Applications, N. Seiberlich, V. Gulani, F. Calamante, A. Campbell-Washburn, M. Doneva,
H. H. Hu, and S. Sourbron, Eds. Academic Press, 2020, vol. 1, pp. 37–51.
[108] J. A. Kim, J. Leung, J. P. Lerch, and A. Kassner, “Reduced cerebrovascular reserve is
regionally associated with cortical thickness reductions in children with sickle cell disease,”
Brain Research, vol. 1642, pp. 263–269, 2016.
[109] G. R. Kirk, M. R. Haynes, S. Palasis, C. Brown, T. G. Burns, M. McCormick, and R. A.
Jones, “Regionally Specific Cortical Thinning in Children with Sickle Cell Disease,” Cerebral
Cortex, vol. 19, no. 7, pp. 1549–1556, 2008.
[110] S. Choi, A. M. Bush, M. T. Borzage, A. A. Joshi, W. J. Mack, T. D. Coates, R. M. Leahy,
and J. C. Wood, “Hemoglobin and mean platelet volume predicts diffuse t1-mri white matter
volume decrease in sickle cell disease patients,” NeuroImage: Clinical, vol. 15, pp. 239–246,
2017.
[111] T. Baldeweg, A. M. Hogan, D. E. Saunders, P. Telfer, D. G. Gadian, F. Vargha-Khadem,
and F. J. Kirkham, “Detecting white matter injury in sickle cell disease using voxel-based
morphometry,” Annals of Neurology, vol. 59, no. 4, pp. 662–672, 2006.
[112] J. M. Kawadler, J. D. Clayden, F. J. Kirkham, T. C. Cox, D. E. Saunders, and C. A. Clark,
“Subcortical and cerebellar volumetric deficits in paediatric sickle cell anaemia,” British
Journal of Haematology, vol. 163, no. 3, pp. 373–376, 2013.
[113] S. Hamdule, M. K¨olbel, H. Stotesbury, R. Murdoch, J. D. Clayden, S. Sahota, A. M. Hood,
C. A. Clark, and F. J. Kirkham, “Effects of regional brain volumes on cognition in sickle cell
anemia: A developmental perspective,” Frontiers in Neurology, vol. 14, 2023.
[114] J. S. W. Campbell and G. B. Pike, “Diffusion magnetic resonance imaging,” in Encyclopedia
of Biomedical Engineering. Elsevier, 1992, pp. 505–518.
[115] K. Kamagata, C. Andica, A. Kato, Y. Saito, W. Uchida, T. Hatano, M. Lukies, T. Ogawa,
H. Takeshige-Amano, T. Akashi, A. Hagiwara, S. Fujita, and S. Aoki, “Diffusion magnetic
resonance imaging-based biomarkers for neurodegenerative diseases,” International Journal
of Molecular Sciences, vol. 22, no. 10, 2021.
Bibliography 96
[116] B. Bodini and O. Ciccarelli, “Chapter 11 - diffusion mri in neurological disorders,” in Diffusion
MRI (Second Edition), second edition ed., H. Johansen-Berg and T. E. Behrens, Eds. San
Diego: Academic Press, 2014, pp. 241–255.
[117] S. N. Jespersen, “White matter biomarkers from diffusion mri,” Journal of Magnetic
Resonance, vol. 291, pp. 127–140, 2018.
[118] P. J. Basser, J. Mattiello, and D. LeBihan, “Mr diffusion tensor spectroscopy and imaging,”
Biophysical journal, vol. 66, no. 1, pp. 259–267, 1994.
[119] S. Mori and J. Zhang, “Principles of diffusion tensor imaging and its applications to basic
neuroscience research,” Neuron, vol. 51, no. 5, pp. 527–539, 2006.
[120] J. H. Jensen and J. A. Helpern, “Mri quantification of non-gaussian water diffusion by kurtosis
analysis,” NMR in Biomedicine, vol. 23, no. 7, pp. 698–710, 2010.
[121] A. Chuhutin, B. Hansen, and S. N. Jespersen, “Precision and accuracy of diffusion kurtosis
estimation and the influence of b-value selection,” NMR in Biomedicine, vol. 30, no. 11, p.
e3777, 2017.
[122] R. N. Henriques, M. M. Correia, M. Marrale, E. Huber, J. Kruper, S. Koudoro, J. D. Yeatman,
E. Garyfallidis, and A. Rokem, “Diffusional kurtosis imaging in the diffusion imaging in
python project,” Frontiers in Human Neuroscience, vol. 15, 2021.
[123] H. Zhang, T. Schneider, C. A. Wheeler-Kingshott, and D. C. Alexander, “Noddi: Practical
in vivo neurite orientation dispersion and density imaging of the human brain,” NeuroImage,
vol. 61, no. 4, pp. 1000–1016, 2012. [Online]. Available: https://www.sciencedirect.com/
science/article/pii/S1053811912003539
[124] A. L. Alexander, J. E. Lee, M. Lazar, and A. S. Field, “Diffusion tensor imaging of the brain,”
Neurotherapeutics, vol. 4, no. 3, pp. 316–329, 2007.
[125] P. J. Basser and C. Pierpaoli, “Microstructural and physiological features of tissues elucidated
by quantitative-diffusion-tensor mri,” Journal of Magnetic Resonance, vol. 213, no. 2, pp.
560–570, 2011.
[126] S. Love, “Demyelinating diseases,” vol. 59, no. 11, pp. 1151–1159, 2006.
[127] C. S. Clements, H. H. Reid, T. Beddoe, F. E. Tynan, M. A. Perugini, T. G. Johns, C. C. A.
Bernard, and J. Rossjohn, “The crystal structure of myelin oligodendrocyte glycoprotein, a
key autoantigen in multiple sclerosis,” Proceedings of the National Academy of Sciences, vol.
100, no. 19, pp. 11 059–11 064, 2003.
[128] P. Preziosa, M. A. Rocca, S. Mesaros, E. Pagani, T. Stosic-Opincal, K. Kacar, M. Absinta,
D. Caputo, J. Drulovic, G. Comi, and M. Filippi, “Intrinsic damage to the major white
matter tracts in patients with different clinical phenotypes of multiple sclerosis: A voxelwise
diffusion-tensor mr study,” Radiology, vol. 260, no. 2, pp. 541–550, 2011.
[129] T. Sigal, M. Shmuel, D. Mark, H. Gil, and A. Anat, “Diffusion tensor imaging of corpus
callosum integrity in multiple sclerosis: Correlation with disease variables,” Journal of
Neuroimaging, vol. 22, no. 1, pp. 33–37, 2012.
Bibliography 97
[130] M. Kolasa, U. Hakulinen, A. Brander, S. Hagman, P. Dastidar, I. Elovaara, and M.-L.
Sumelahti, “Diffusion tensor imaging and disability progression in multiple sclerosis: A 4-
year follow-up study,” Brain and Behavior, vol. 9, no. 1, p. e01194, 2019.
[131] D. Harrison, B. Caffo, N. Shiee, J. Farrell, P.-L. Bazin, S. Farrell, J. Ratchford, P. Calabresi,
and D. Reich, “Longitudinal changes in diffusion tensor–based quantitative mri in multiple
sclerosis,” Neurology, vol. 76, no. 2, pp. 179–186, 2011.
[132] W. Rashid, A. Hadjiprocopis, G. Davies, C. Griffin, D. Chard, M. Tiberio, D. Altmann,
C. Wheeler-Kingshott, D. Tozer, A. Thompson et al., “Longitudinal evaluation of clinically
early relapsing-remitting multiple sclerosis with diffusion tensor imaging,” Journal of
neurology, vol. 255, pp. 390–397, 2008.
[133] D. Ontaneda, K. Sakaie, J. Lin, X.-F. Wang, M. J. Lowe, M. D. Phillips, and R. J. Fox,
“Measuring brain tissue integrity during 4 years using diffusion tensor imaging,” American
Journal of Neuroradiology, vol. 38, no. 1, pp. 31–38, 2017.
[134] J. Chojdak- Lukasiewicz, E. Dziadkowiak, A. Zimny, and B. Paradowski, “Cerebral small
vessel disease: A review,” Advances in Clinical and Experimental Medicine, vol. 30, no. 3,
pp. 349–356, 2021.
[135] E. A. Zeestraten, P. Benjamin, C. Lambert, A. J. Lawrence, O. A. Williams, R. G. Morris,
T. R. Barrick, and H. S. Markus, “Application of diffusion tensor imaging parameters to
detect change in longitudinal studies in cerebral small vessel disease,” PLOS ONE, vol. 11,
no. 1, pp. 1–16, 01 2016.
[136] M. Pasi, I. W. van Uden, A. M. Tuladhar, F.-E. de Leeuw, and L. Pantoni, “White matter
microstructural damage on diffusion tensor imaging in cerebral small vessel disease,” Stroke,
vol. 47, no. 6, pp. 1679–1684, 2016.
[137] M. M. D’Souza, S. Gorthi, K. Vadwala, R. Trivedi, C. Vijayakumar, P. Kaur, and S. Khushu,
“Diffusion tensor tractography in cerebral small vessel disease: correlation with cognitive
function,” The Neuroradiology Journal, vol. 31, no. 1, pp. 83–89, 2018.
[138] G. Jean Harry and A. D. Toews, “Chapter 4 - myelination, dysmyelination, and
demyelination,” in Handbook of Developmental Neurotoxicology, W. Slikker and L. W. Chang,
Eds. San Diego: Academic Press, 1998, pp. 87–115.
[139] N. I. of Neurological Disorders and Stroke, “Metachromatic leukodystrophy,” https://www.
ninds.nih.gov/health-information/disorders/metachromatic-leukodystrophy, accessed: 2023-
11-20.
[140] D. F. van Rappard, J. J. Boelens, and N. I. Wolf, “Metachromatic leukodystrophy: Disease
spectrum and approaches for treatment,” Best Practice & Research Clinical Endocrinology
and Metabolism, vol. 29, no. 2, pp. 261–273, 2015.
[141] P. Morell and H. Jurevics, “Origin of cholesterol in myelin,” Neurochemical research, vol. 21,
pp. 463–470, 1996.
Bibliography 98
[142] T. C. d. M. Costa, R. Chiari-Correia, C. E. G. Salmon, L. G. Darrigo-Junior, C. E. S. Grecco,
F. Pieroni, J. T. B. Faria, A. B. P. Stracieri, J. B. Dias, D. A. de Moraes et al., “Hematopoietic
stem cell transplantation reverses white matter injury measured by diffusion-tensor imaging
(dti) in sickle cell disease patients,” Bone Marrow Transplantation, vol. 56, no. 11, pp. 2705–
2713, 2021.
[143] A. Balci, S. Karazincir, Y. Beyoglu, C. Cingiz, R. Davran, E. Gali, E. Okuyucu, and
E. Egilmez, “Quantitative brain diffusion-tensor mri findings in patients with sickle cell
disease,” AJR Am J Roentgenol, vol. 198, no. 5, pp. 1167–1174, 2012.
[144] Y. Chai, C. Ji, J. Coloigner, S. Choi, M. Balderrama, C. Vu, B. Tamrazi, T. Coates, J. C.
Wood, S. H. O’Neil, and N. Lepore, “Tract-specific analysis and neurocognitive functioning
in sickle cell patients without history of overt stroke,” Brain and Behavior, vol. 11, no. 3, p.
e01978, 2021.
[145] C. Gonz´alez-Zacar´ıas, S. Choi, C. Vu, B. Xu, J. Shen, A. A. Joshi, R. M. Leahy, and J. C.
Wood, “Chronic anemia: The effects on the connectivity of white matter,” Frontiers in
Neurology, vol. 13, p. 894742, 2022.
[146] Y. Wang, S. Fellah, M. E. Fields, K. P. Guilliams, M. M. Binkley, C. Eldeniz, J. S. Shimony,
M. Reis, K. D. Vo, Y. Chen, J. M. Lee, H. An, and A. L. Ford, “Cerebral oxygen metabolic
stress, microstructural injury, and infarction in adults with sickle cell disease,” Neurology,
vol. 97, no. 9, pp. 902–912, 2021.
[147] J. M. Kawadler, F. J. Kirkham, J. D. Clayden, M. J. Hollocks, E. L. Seymour, R. Edey,
P. Telfer, A. Robins, O. Wilkey, S. Barker, T. C. Cox, and C. A. Clark, “White matter
damage relates to oxygen saturation in children with sickle cell anemia without silent cerebral
infarcts,” Stroke, vol. 46, no. 7, pp. 1793–1799, 2015.
[148] M. Jacob, H. Stotesbury, J. M. Kawadler, W. Lapadaire, D. E. Saunders, R. Z. Sangeda,
C. Chamba, R. Kazema, J. Makani, F. J. Kirkham, and C. A. Clark, “White matter integrity
in tanzanian children with sickle cell anemia: A diffusion tensor imaging study,” Stroke,
vol. 51, no. 4, pp. 1166–1173, 2020.
[149] E. X. Wu and M. M. Cheung, “Mr diffusion kurtosis imaging for neural tissue
characterization,” NMR in Biomedicine, vol. 23, no. 7, pp. 836–848, 2010.
[150] C. D. Kroenke, J. J. Ackerman, and D. A. Yablonskiy, “On the nature of the naa diffusion
attenuated mr signal in the central nervous system,” Magnetic Resonance in Medicine: An
Official Journal of the International Society for Magnetic Resonance in Medicine, vol. 52,
no. 5, pp. 1052–1059, 2004.
[151] H. Zhang, P. L. Hubbard, G. J. Parker, and D. C. Alexander, “Axon diameter mapping in
the presence of orientation dispersion with diffusion mri,” Neuroimage, vol. 56, no. 3, pp.
1301–1315, 2011.
[152] D. C. Alexander, P. L. Hubbard, M. G. Hall, E. A. Moore, M. Ptito, G. J. Parker, and T. B.
Dyrby, “Orientationally invariant indices of axon diameter and density from diffusion mri,”
NeuroImage, vol. 52, no. 4, pp. 1374–1389, 2010.
Bibliography 99
[153] A. Szafer, J. Zhong, and J. C. Gore, “Theoretical model for water diffusion in tissues,”
Magnetic Resonance in Medicine, vol. 33, no. 5, pp. 697–712, 1995.
[154] M. E. Watts, R. Pocock, and C. Claudianos, “Brain energy and oxygen metabolism: Emerging
role in normal function and disease,” Frontiers in Molecular Neuroscience, vol. 11, 2018.
[155] J. Coloigner, Y. Kim, A. Bush, S. Choi, M. C. Balderrama, T. D. Coates, S. H. O’Neil,
N. Lepore, and J. C. Wood, “Contrasting resting-state fmri abnormalities from sickle and
non-sickle anemia,” PLoS One, vol. 12, no. 5, 2017.
[156] H. Stotesbury, F. J. Kirkham, M. K¨olbel, P. Balfour, J. D. Clayden, S. Sahota, S. Sakaria,
D. E. Saunders, J. Howard, R. Kesse-Adu, B. Inusa, M. Pelidis, S. Chakravorty, D. C. Rees,
M. Awogbade, O. Wilkey, M. Layton, C. A. Clark, and J. M. Kawadler, “White matter
integrity and processing speed in sickle cell anemia,” Neurology, vol. 90, no. 5, 2018.
[157] M. E. Fields, A. E. Mirro, M. M. Binkley, K. P. Guilliams, J. B. Lewis, S. Fellah, Y. Chen,
M. L. Hulbert, H. An, A. L. Ford, and J. M. Lee, “Cerebral oxygen metabolic stress is
increased in children with sickle cell anemia compared to anemic controls,” American journal
of hematology, vol. 97, no. 6, 2022.
[158] K. P. Guilliams, M. E. Fields, D. K. Ragan, Y. Chen, C. Eldeniz, M. L. Hulbert, M. M.
Binkley, J. N. Rhodes, J. S. Shimony, R. C. McKinstry, K. Vo, H. An, J. M. Lee, and A. L.
Ford, “Large-vessel vasculopathy in children with sickle cell disease: A magnetic resonance
imaging study of infarct topography and focal atrophy,” Pediatric neurology, vol. 69, pp.
49–57, 2017.
[159] S. Gevers, A. J. Nederveen, K. Fijnvandraat, S. M. van den Berg, P. van Ooij, D. F. Heijtel,
H. Heijboer, P. J. Nederkoorn, M. Engelen, M. J. van Osch, and C. B. Majoie, “Arterial spin
labeling measurement of cerebral perfusion in children with sickle cell disease,” Journal of
magnetic resonance imaging : JMRI, vol. 35, no. 4, pp. 779–787, 2012.
[160] P. D. Kosinski, P. L. Croal, J. Leung, S. Williams, I. Odame, G. M. Hare, M. Shroff, and
A. Kassner, “The severity of anaemia depletes cerebrovascular dilatory reserve in children
with sickle cell disease: a quantitative magnetic resonance imaging study,” British journal of
haematology, vol. 176, no. 2, pp. 280–287, 2017.
[161] K. P. Guilliams, M. E. Fields, D. K. Ragan, C. Eldeniz, M. M. Binkley, Y. Chen, L. S.
Comiskey, A. Doctor, , M. L. Hulbert, J. S. Shimony, K. Vo, R. C. McKinstry, H. An, J. M.
Lee, and A. L. Ford, “Red cell exchange transfusions lower cerebral blood flow and oxygen
extraction fraction in pediatric sickle cell anemia,” Blood, vol. 131, no. 9, pp. 1012–1021, 2018.
[162] M. R. Juttukonda, M. J. Donahue, S. L. Waddle, L. T. Davis, C. A. Lee, N. J. Patel,
S. Pruthi, A. A. Kassim, and L. C. Jordan, “Reduced oxygen extraction efficiency in sickle
cell anemia patients with evidence of cerebral capillary shunting,” Journal of cerebral blood
flow and metabolism : official journal of the International Society of Cerebral Blood Flow and
Metabolism, vol. 41, no. 3, pp. 546–560, 2021.
[163] E. Mandonnet, S. Sarubbo, and L. Petit, “Response: Commentary: The nomenclature of
human white matter association pathways: Proposal for a systematic taxonomic anatomical
classification,” Frontiers in neuroanatomy, vol. 13, p. 91, 2019.
Bibliography 100
[164] S. Sandor and R. Leahy, “Surface-based labeling of cortical anatomy using a deformable
atlas,” IEEE transactions on medical imaging, vol. 16, no. 1, pp. 41–54, 1997.
[165] D. W. Shattuck, S. R. Sandor-Leahy, K. A. Schaper, D. A. Rottenberg, and R. M. Leahy,
“Magnetic resonance image tissue classification using a partial volume model,” NeuroImage,
vol. 13, no. 5, pp. 856–876, 2001.
[166] D. W. Shattuck and R. M. Leahy, “Automated graph-based analysis and correction of cortical
volume topology,” IEEE transactions on medical imaging, vol. 20, no. 11, pp. 1167–1177, 2001.
[167] A. A. Joshi, D. W. Shattuck, P. M. Thompson, and R. M. Leahy, “Surface-constrained
volumetric brain registration using harmonic mappings,” IEEE transactions on medical
imaging, vol. 26, no. 12, pp. 1657–1669, 2007.
[168] A. A. Joshi, S. Choi, Y. Liu, M. Chong, G. Sonkar, J. Gonzalez-Martinez, D. Nair, J. L.
Wisnowski, J. P. Haldar, D. W. Shattuck, H. Damasio, and R. M. Leahy, “A hybrid highresolution anatomical mri atlas with sub-parcellation of cortical gyri using resting fmri,”
Journal of neuroscience methods, vol. 374, p. 109566, 2022.
[169] J. L. R. Andersson, M. S. Graham, E. Zsoldos, and S. N. Sotiropoulos, “Incorporating outlier
detection and replacement into a non-parametric framework for movement and distortion
correction of diffusion mr images,” NeuroImage, vol. 141, pp. 556–572, 2016.
[170] J. L. R. A. S. N. Sotiropoulos, “An integrated approach to correction for off-resonance effects
and subject movement in diffusion mr imaging,” NeuroImage, vol. 125, pp. 1063–1078, 2016.
[171] C. Bhushan, J. P. Haldar, S. Choi, A. A. Joshi, D. W. Shattuck, and R. M. Leahy, “Coregistration and distortion correction of diffusion and anatomical images based on inverse
contrast normalization,” NeuroImage, vol. 115, pp. 269–280, 2015.
[172] D. Varadarajan and J. P. Haldar, “A theoretical signal processing framework for linear
diffusion mri: Implications for parameter estimation and experiment design,” NeuroImage,
vol. 161, pp. 206–218, 2017.
[173] D. Varadarajan and J. Haldar, “Towards optimal linear estimation of orientation distribution
functions with arbitrarily sampled diffusion mri data,” 2018, proceedings. IEEE International
Symposium on Biomedical Imaging.
[174] F. C. Yeh, T. D. Verstynen, Y. Wang, J. C. Fern´andez-Miranda, and W. Y. I. Tseng,
“Deterministic diffusion fiber tracking improved by quantitative anisotropy,” PLoS One,
vol. 8, no. 11, p. 80713, 2013.
[175] D. K. Jones, T. R. Kn¨osche, and R. Turner, “White matter integrity, fiber count, and other
fallacies: the do’s and don’ts of diffusion mri,” NeuroImage, vol. 73, pp. 239–254, 2013.
[176] J. M. Soares, P. Marques, V. Alves, and N.Sousa, “A hitchhiker’s guide to diffusion tensor
imaging,” Frontiers in neuroscience, vol. 7, p. 31, 2013.
[177] C. Lebel and C. Beaulieu, “Longitudinal development of human brain wiring continues from
childhood into adulthood,” The Journal of neuroscience : the official journal of the Society
for Neuroscience, vol. 31, no. 30, pp. 10 937–10 947, 2011.
Bibliography 101
[178] Y. S. Chang, J. P. Owen, N. J. Pojman, T. Thieu, P. Bukshpun, M. L. J. Wakahiro, J. I.
Berman, T. P. L. Roberts, S. S. Nagarajan, E. H. Sherr, and P. Mukherjee, “White matter
changes of neurite density and fiber orientation dispersion during human brain maturation,”
PLoS ONE, vol. 10, no. 6, p. 0123656, 2015.
[179] D. Yekutieli and Y. Benjamini, “Resampling-based false discovery rate controlling multiple
test procedures for correlated test statistics,” Journal of Statistical Planning and Inference,
vol. 82, no. 1, pp. 171–196, 1999.
[180] B. F. J. Manly, “Randomization and regression methods for testing for associations with
geographical, environmental and biological distances between populations,” Population
Ecology, vol. 28, no. 2, pp. 201–218, 1986.
[181] B. F. Manly, Randomization, Bootstrap and Monte Carlo Methods in Biology. Chapman
and Hall-CRC, 2007.
[182] A. M. Winkler, G. R. Ridgway, M. A. Webster, S. M. Smith, and T. E. Nichols, “Permutation
inference for the general linear model,” NeuroImage, vol. 92, pp. 381–397, 2014.
[183] R. C. Team, “R: A language and environment for statistical computing,” Vienna: R
Foundation for Statistical Computing, 2020.
[184] M. H. Steinberg, “Management of sickle cell disease,” N Engl J Med, vol. 340, pp. 1021–1030,
1999.
[185] P. L. Kavanagh, P. G. Sprinza, S. R. V. H. Bauchner, and C. J. Wang, “Management of
children with sickle cell disease: A comprehensive review of the literature,” Pediatrics, vol.
128, no. 6, pp. e1552–e1574, 2011.
[186] B. P. Yawn, G. R. Buchanan, A. N. Afenyi-Annan, S. K. Ballas, K. L. Hassell, A. H. James,
L. Jordan, S. M. Lanzkron, R. Lottenberg, W. J. Savage, P. J. Tanabe, R. E. Ware, M. H.
Murad, J. C. Goldsmith, E. Ortiz, R. Fulwood, A. Horton, and J. John-Sowah, “Management
of sickle cell disease: Summary of the 2014 evidence-based report by expert panel members,”
JAMA, vol. 312, pp. 1033–1048, 2014.
[187] H. Enninful-Eghan, R. H. Moore, R. Ichord, K. Smith-Whitley, and J. L. Kwiatkowski,
“Transcranial doppler ultrasonography and prophylactic transfusion program is effective in
preventing overt stroke in children with sickle cell disease,” J. Pediatrics, vol. 157, pp. 479–
484, 2010.
[188] M. E. Fields, K. P. Guilliams, D. K. Ragan, M. M. Binkley, C. Eldeniz, Y. Chen, M. L,
Hulbert, R. C, McKinstry, J. S. Shimony, K. D. Vo, A. Doctor, H. An, and J.-M. L. Andria
L. Ford, “Regional oxygen extraction predicts border zone vulnerability to stroke in sickle
cell disease,” Neurology, vol. 90, no. 13, p. 1134–1144, 2018.
[189] A. M. Bush, T. D. Coates, and J. C. Wood, “Diminished cerebral oxygen extraction and
metabolic rate in sickle cell disease using t2 relaxation under spin tagging mri, journal =
Magn Reson Med., volume = 80, pages = 294-303, year = 2018.”
Bibliography 102
[190] P. L. Croal, J. Leung, C. L. Phillips, M. G. Serafin, and A. Kassner, “Quantification of
pathophysiological alterations in venous oxygen saturation: A comparison of global mr
susceptometry techniques,” Magnetic Resonance Imaging, vol. 58, pp. 18–23, 2019.
[191] M. R. Juttukonda, M. J. Donahue, L. T. Davis, M. C. Gindville, C. A. Lee, N. J. Patel,
A. A. Kassim, S. Pruthi, J. Hendrikse, and L. C. Jordan, “Preliminary evidence for cerebral
capillary shunting in adults with sickle cell anemia,” J Cereb Blood Flow Metab., vol. 39, pp.
1099–1010, 2019.
[192] W. Li, X. Xu, P. Liu, J. J. Strouse, J. F. Casella, H. Lu, P. C. van Zijl, and Q. Qin,
“Quantification of whole-brain oxygenation extraction fraction and cerebral metabolic rate of
oxygen consumption in adults with sickle cell anemia using individual t2-based oxygenation
calibrations,” Magnetic Resonance in Medicine, vol. 83, no. 3, pp. 1066–1080, 2020.
[193] K. M. Musallam, A. T. Taher, and M. K. E. A. Rachmilewitz, “Cerebral infarction in βthalassemia intermedia: Breaking the silence,” Thrombosis Research, vol. 130, pp. 695–702,
2012.
[194] A. H. Ropper, J. Klein, M. A. Samuels, and S. Prasad, Adams and Victor’s Principles of
Neurology. N.Y: McGraw-Hill Education LLC, 2019.
[195] L. Caplan, Stroke. Cary: Oxford University Press, Incorporated, 2010.
[196] K. Sam, J. Conklin, K. R. Holmes, O. Sobczyk, J. Poublanc, A. P. Crawley, D. M. Mandell,
L. Venkatraghavan, J. Duffin, J. A. Fisher, S. E. Black, and D. J. Mikulis, “Impaired
dynamic cerebrovascular response to hypercapnia predicts development of white matter
hyperintensities,” NeuroImage: Clinical, vol. 11, pp. 796–801, 2016.
[197] C. T. Quinn, “Minireview: Clinical severity in sickle cell disease: the challenges of definition
and prognostication,” Exp Biol Med., vol. 241, pp. 679–688, 2016.
[198] E. M. Novelli and M. T. Gladwin, “Crises in sickle cell disease,” Chest, vol. 149, pp. 1082–
1093, 2016.
[199] D. S. Darbari, V. A. Sheehan, and S. K. Ballas, “The vaso-occlusive pain crisis in sickle cell
disease: Definition, pathophysiology, and management,” Eur J Haematol., vol. 105, no. 3, pp.
237–246, 2020.
[200] J. J. Field, “Five lessons learned about long-term pain management in adults with sickle cell
disease,” Hematology, vol. 2017, no. 1, pp. 406–411, 2017.
[201] Y. H. Dang, Y. Zhao, B. Xing, X. J. Zhao, F. Q. Huo, J. S. Tang, C. L. Qu, and T. Chen,
“The role of dopamine receptors in ventrolateral orbital cortex-evoked anti-nociception in a
rat model of neuropathic pain,” Neuroscience, vol. 4, no. 2010, pp. 1872–1880, 2010.
[202] E. Vachon-Presseau, P. T´etreault, B. Petre, L. Huang, S. E. Berger, S. Torbey, A. T. Baria,
A. R. Mansour, J. A. Hashmi, J. W. Griffith, E. Comasco, T. J. Schnitzer, M. N. Baliki,
and A. V. Apkarian, “Corticolimbic anatomical characteristics predetermine risk for chronic
pain,” Brain, vol. 139, no. 7, pp. 1958–1970, 2016.
Bibliography 103
[203] W. Y. Ong, C. S. Stohler, and D. R. Herr, “Role of the prefrontal cortex in pain processing,”
Molecular Neurobiology, vol. 56, pp. 1137–1166, 2019.
[204] H.-Y. Sheng, S.-S. Lv, Y.-Q. Cai, W. Shi, W. Lin, T.-T. Liu, N. Lv, H. Cao, L. Zhang,
and Y.-Q. Zhang, “Activation of ventrolateral orbital cortex improves mouse neuropathic
pain–induced anxiodepression,” JCI Insight, vol. 5, p. 133625, 2020.
[205] R. F. Smallwood, A. R. Laird, A. E. Ramage, A. L. Parkinson, J. Lewis, Daniel, J. Clauw,
D. A. Williams, T. Schmidt-Wilcke, M. J. Farrell, S. B. Eickhoff, and D. A. Robin, “Structural
brain anomalies and chronic pain: A quantitative meta-analysis of gray matter volume,”
Journal of Pain, vol. 14, no. 7, pp. 663–675, 2013.
[206] B. Zhang, M. Jung, Y. Tu, R. Gollub, C. Lang, A. Ortiz, J. Park, G. Wilson, J. Gerber,
I. Mawla, S.-T. Chan, A. Wasan, R. Edwards, J. Lee, V. Napadow, T. Kaptchuk, B. Rosen,
and J. Kong, “Identifying brain regions associated with the neuropathology of chronic low
back pain: a resting-state amplitude of low-frequency fluctuation study,” Br J Anaesth, vol.
123, pp. 303–311, 2019.
[207] L. D. Berkelhammer, A. L. Williamson, S. D. Sanford, C. L. Dirksen, W. G. Sharp, A. S.
Margulies, and R. A. Prengler, “Neurocognitive sequelae of pediatric sickle cell disease: A
review of the literature,” Child Neuropsychology, vol. 13, no. 2, pp. 120–131, 2007.
[208] M. S. Elalfy, R. H. Aly, H. Azzam, K. Aboelftouh, R. H. Shatla, M. Tarif, M. Abdatty,
and R. M. Elsayed, “Neurocognitive dysfunction in children with beta thalassemia major:
psychometric, neurophysiologic and radiologic evaluation,” Hematology, vol. 22, no. 10, pp.
617–622, 2017.
[209] J. A. Detterich, “Simple chronic transfusion therapy, a crucial therapeutic option for sickle
cell disease, improves but does not normalize blood rheology: What should be our goals
for transfusion therapy?” Clinical hemorheology and microcirculation, vol. 68, no. 2-3, pp.
173–186, 2018.
[210] G. . A. Collaborators et al., “Prevalence, years lived with disability, and trends in anaemia
burden by severity and cause, 1990–2021: findings from the global burden of disease study
2021,” The Lancet Haematology, vol. 10, no. 9, pp. 713–734, 2023.
[211] S. Safiri, A.-A. Kolahi, M. Noori, S. A. Nejadghaderi, N. Karamzad, N. L. Bragazzi, M. J.
Sullman, M. Abdollahi, G. S. Collins, J. S. Kaufman et al., “Burden of anemia and its
underlying causes in 204 countries and territories, 1990–2019: results from the global burden
of disease study 2019,” Journal of hematology & oncology, vol. 14, no. 1, pp. 1–16, 2021.
[212] H. L. Muncie Jr and J. S. Campbell, “Alpha and beta thalassemia,” American family
physician, vol. 80, no. 4, pp. 339–344, 2009.
[213] A. Kattamis, J. L. Kwiatkowski, and Y. Aydinok, “Thalassaemia,” The lancet, vol. 399, no.
10343, pp. 2310–2324, 2022.
[214] R. G. Steen, T. Emudianughe, M. Hunte, J. Glass, S. Wu, X. Xiong, and W. E. Reddick,
“Brain volume in pediatric patients with sickle cell disease: evidence of volumetric growth
delay?” American Journal of Neuroradiology, vol. 26, no. 3, pp. 455–462, 2005.
Bibliography 104
[215] W. T. Zempsky, M. C. Stevens, J. P. Santanelli, A. M. Gaynor, and S. Khadka, “Altered
functional connectivity in sickle cell disease exists at rest and during acute pain challenge,”
The Clinical Journal of Pain, vol. 33, no. 12, pp. 1060–1070, 2017.
[216] M. R. DeBaun, F. D. Armstrong, R. C. McKinstry, R. E. Ware, E. Vichinsky, and F. J.
Kirkham, “Silent cerebral infarcts: a review on a prevalent and progressive cause of neurologic
injury in sickle cell anemia,” Blood, The Journal of the American Society of Hematology, vol.
119, no. 20, pp. 4587–4596, 2012.
[217] J. M. Kawadler, C. A. Clark, R. C. McKinstry, and F. J. Kirkham, “Brain atrophy in
paediatric sickle cell anaemia: findings from the silent infarct transfusion (sit) trial,” British
Journal of Haematology, vol. 177, no. 1, pp. 151–153, 2017.
[218] D. S. Darbari, O. Eigbire-Molen, M. R. Ponisio, M. V. Milchenko, M. J. Rodeghier, J. F.
Casella, R. C. McKinstry, and M. R. DeBaun, “Progressive loss of brain volume in children
with sickle cell anemia and silent cerebral infarct: A report from the silent cerebral infarct
transfusion trial,” American journal of hematology, vol. 93, no. 12, pp. 406–408, 2018.
[219] J. N. Brewin, H. Rooks, K. Gardner, H. Senior, M. Morje, H. Patel, D. Calvet, P. Bartolucci,
S.-L. Thein, S. Menzel et al., “Genome wide association study of silent cerebral infarction in
sickle cell disease (hbss and hbsc),” haematologica, vol. 106, no. 6, p. 1770, 2021.
[220] M. E. Houwing, R. L. Grohssteiner, M. H. Dremmen, F. Atiq, W. M. Bramer, A. P. de Pagter,
C. M. Zwaan, T. J. White, M. W. Vernooij, and M. H. Cnossen, “Silent cerebral infarcts in
patients with sickle cell disease: a systematic review and meta-analysis,” BMC medicine,
vol. 18, pp. 1–17, 2020.
[221] R. Jones, M. J. Donahue, L. T. Davis, S. Pruthi, S. L. Waddle, C. Custer, N. J. Patel,
M. R. DeBaun, A. A. Kassim, M. Rodeghier et al., “Silent infarction in sickle cell disease is
associated with brain volume loss in excess of infarct volume,” Frontiers in Neurology, vol. 14,
2023.
[222] D. Purves, G. J. Augustine, D. Fitzpatrick, W. C. Hall, A.-S. LaMantia, J. O. McNamara,
and L. E. White, “Neuroscience. 4th,” Sunderland, Mass: Sinauer. xvii, vol. 857, p. 944,
2008.
[223] S. Fadnavis, A. Chowdhury, J. Batson, P. Drineas, and E. Garyfallidis, “Patch2self denoising
of diffusion mri with self-supervision and matrix sketching,” bioRxiv, 2022.
[224] E. Kellner, B. Dhital, V. G. Kiselev, and M. Reisert, “Gibbs-ringing artifact removal based on
local subvoxel-shifts,” Magnetic resonance in medicine, vol. 76, no. 5, pp. 1574–1581, 2016.
[225] E. Garyfallidis, M. Brett, B. Amirbekian, A. Rokem, S. Van Der Walt, M. Descoteaux,
I. Nimmo-Smith, and D. Contributors, “Dipy, a library for the analysis of diffusion mri
data,” Frontiers in neuroinformatics, vol. 8, 2014.
[226] J. Veraart, D. H. Poot, W. Van Hecke, I. Blockx, A. Van der Linden, M. Verhoye, and
J. Sijbers, “More accurate estimation of diffusion tensor parameters using diffusion kurtosis
imaging,” Magnetic resonance in medicine, vol. 65, no. 1, pp. 138–145, 2011.
Bibliography 105
[227] R. Neto Henriques, H. Ferreira, and M. Correia, “Diffusion kurtosis imaging of the healthy
human brain,” Master’s thesis, Program in Biomedical Engineering, Departamento de F´ısica,
2012.
[228] M. Houwing, P. De Pagter, E. Van Beers, B. Biemond, E. Rettenbacher, A. Rijneveld,
E. Schols, J. Philipsen, R. Tamminga, K. F. van Draat et al., “Sickle cell disease: clinical
presentation and management of a global health challenge,” Blood reviews, vol. 37, p. 100580,
2019.
[229] Y. Daniel, J. Elion, B. Allaf, C. Badens, M. J. Bouva, I. Brincat, E. Cela, C. Coppinger,
M. De Montalembert, B. Gulbis et al., “Newborn screening for sickle cell disease in europe,”
International Journal of Neonatal Screening, vol. 5, no. 1, p. 15, 2019.
[230] M. Kronlage, K. Pitarokoili, D. Schwarz, T. Godel, S. Heiland, M.-S. Yoon, M. Bendszus, and
P. B¨aumer, “Diffusion tensor imaging in chronic inflammatory demyelinating polyneuropathy:
diagnostic accuracy and correlation with electrophysiology,” Investigative radiology, vol. 52,
no. 11, pp. 701–707, 2017.
[231] P. J. Winklewski, A. Sabisz, P. Naumczyk, K. Jodzio, E. Szurowska, and A. Szarmach,
“Understanding the physiopathology behind axial and radial diffusivity changes—what do
we know?” Frontiers in neurology, vol. 9, p. 92, 2018.
[232] R. Naismith, J. Xu, N. Tutlam, K. Trinkaus, A. Cross, and S.-K. Song, “Radial diffusivity
in remote optic neuritis discriminates visual outcomes,” Neurology, vol. 74, no. 21, pp. 1702–
1710, 2010.
[233] M. Yoshida, M. Hori, K. Yokoyama, I. Fukunaga, M. Suzuki, K. Kamagata, K. Shimoji,
A. Nakanishi, N. Hattori, Y. Masutani et al., “Diffusional kurtosis imaging of normalappearing white matter in multiple sclerosis: preliminary clinical experience,” Japanese
journal of radiology, vol. 31, pp. 50–55, 2013.
[234] R. J. Franklin and C. Ffrench-Constant, “Remyelination in the cns: from biology to therapy,”
Nature Reviews Neuroscience, vol. 9, no. 11, pp. 839–855, 2008.
[235] T. W. Chapman, G. E. Olveda, X. Bame, E. Pereira, and R. A. Hill, “Oligodendrocyte
death initiates synchronous remyelination to restore cortical myelin patterns in mice,” Nature
Neuroscience, vol. 26, no. 4, pp. 555–569, 2023.
[236] C. Guglielmetti, J. Veraart, E. Roelant, Z. Mai, J. Daans, J. Van Audekerke, M. Naeyaert,
G. Vanhoutte, R. Delgado y Palacios, J. Praet, E. Fieremans, P. Ponsaerts, J. Sijbers, A. Van
der Linden, and M. Verhoye, “Diffusion kurtosis imaging probes cortical alterations and white
matter pathology following cuprizone induced demyelination and spontaneous remyelination,”
NeuroImage, vol. 125, pp. 363–377, 2016.
[237] M. F. Falangola, D. N. Guilfoyle, A. Tabesh, E. S. Hui, X. Nie, J. H. Jensen, S. V.
Gerum, C. Hu, J. LaFrancois, H. R. Collins, and J. A. Helpern, “Histological correlation of
diffusional kurtosis and white matter modeling metrics in cuprizone-induced corpus callosum
demyelination,” NMR in Biomedicine, vol. 27, no. 8, pp. 948–957, 2014.
Bibliography 106
[238] I. Timmers, A. Roebroeck, M. Bastiani, B. Jansma, E. Rubio-Gozalbo, and H. Zhang,
“Assessing microstructural substrates of white matter abnormalities: A comparative study
using dti and noddi,” PLOS ONE, vol. 11, no. 12, pp. 1–15, 12 2016.
[239] N. Wang, J. Zhang, G. Cofer, Y. Qi, R. J. Anderson, L. E. White, and G. Allan Johnson,
“Neurite orientation dispersion and density imaging of mouse brain microstructure,” Brain
Structure and Function, vol. 224, pp. 1797–1813, 2019.
[240] S. N. Jespersen, C. R. Bjarkam, J. R. Nyengaard, M. M. Chakravarty, B. Hansen,
T. Vosegaard, L. Østergaard, D. Yablonskiy, N. C. Nielsen, and P. Vestergaard-Poulsen,
“Neurite density from magnetic resonance diffusion measurements at ultrahigh field:
comparison with light microscopy and electron microscopy,” Neuroimage, vol. 49, no. 1, pp.
205–216, 2010.
[241] S. N. Jespersen, L. A. Leigland, A. Cornea, and C. D. Kroenke, “Determination of axonal and
dendritic orientation distributions within the developing cerebral cortex by diffusion tensor
imaging,” IEEE transactions on medical imaging, vol. 31, no. 1, pp. 16–32, 2011.
[242] S. Collorone, N. Cawley, F. Grussu, F. Prados, F. Tona, A. Calvi, B. Kanber, T. Schneider,
L. Kipp, H. Zhang et al., “Reduced neurite density in the brain and cervical spinal cord in
relapsing–remitting multiple sclerosis: A noddi study,” Multiple Sclerosis Journal, vol. 26,
no. 13, pp. 1647–1657, 2020.
[243] S. De Santis, M. Bastiani, A. Droby, P. Kolber, F. Zipp, E. Pracht, T. Stoecker, S. Groppa,
and A. Roebroeck, “Characterizing microstructural tissue properties in multiple sclerosis with
diffusion mri at 7 t and 3 t: the impact of the experimental design,” Neuroscience, vol. 403,
pp. 17–26, 2019.
[244] R. S. Mackin, P. Insel, D. Truran, E. P. Vichinsky, L. D. Neumayr, F. Armstrong, J. I. Gold,
K. Kesler, J. Brewer, and M. W. Weiner, “Neuroimaging abnormalities in adults with sickle
cell anemia,” Neurology, vol. 82, no. 10, pp. 835–841, 2014.
[245] R. Hazra, H. Pu, L. M. Foley, L. Little-Ihrig, T. K. Hitchens, S. Ghosh, S. F. Ofori-Acquah,
X. Hu, and E. M. Novelli, “White-matter abnormalities and cognitive dysfunction are linked
to astrocyte activation in sickle mice,” PNAS Nexus, vol. 2, no. 5, p. 149, 2023.
[246] J. H. Becker, J. J. Lin, M. Doernberg, K. Stone, A. Navis, J. R. Festa, and J. P. Wisnivesky,
“Assessment of Cognitive Function in Patients After COVID-19 Infection,” JAMA Network
Open, vol. 4, no. 10, pp. e2 130 645–e2 130 645, 10 2021.
[247] A. Nath, “Long-haul covid,” pp. 559–560, 2020.
[248] A. Fern´andez-Casta˜neda, P. Lu, A. C. Geraghty, E. Song, M.-H. Lee, J. Wood, M. R. O’Dea,
S. Dutton, K. Shamardani, K. Nwangwu, R. Mancusi, B. Yal¸cın, K. R. Taylor, L. AcostaAlvarez, K. Malacon, M. B. Keough, L. Ni, P. J. Woo, D. Contreras-Esquivel, A. M. S. Toland,
J. R. Gehlhausen, J. Klein, T. Takahashi, J. Silva, B. Israelow, C. Lucas, T. Mao, M. A.
Pe˜na-Hern´andez, A. Tabachnikova, R. J. Homer, L. Tabacof, J. Tosto-Mancuso, E. Breyman,
A. Kontorovich, D. McCarthy, M. Quezado, H. Vogel, M. M. Hefti, D. P. Perl, S. Liddelow,
R. Folkerth, D. Putrino, A. Nath, A. Iwasaki, and M. Monje, “Mild respiratory covid can
Chap. 4 Chronic anemia and WM damage 107
cause multi-lineage neural cell and myelin dysregulation,” Cell, vol. 185, no. 14, pp. 2452–
2468.e16, 2022.
[249] E. M. Gibson, S. Nagaraja, A. Ocampo, L. T. Tam, L. S. Wood, P. N. Pallegar, J. J. Greene,
A. C. Geraghty, A. K. Goldstein, L. Ni et al., “Methotrexate chemotherapy induces persistent
tri-glial dysregulation that underlies chemotherapy-related cognitive impairment,” Cell, vol.
176, no. 1, pp. 43–55, 2019.
[250] S. A. Liddelow, K. A. Guttenplan, L. E. Clarke, F. C. Bennett, C. J. Bohlen, L. Schirmer,
M. L. Bennett, A. E. M¨unch, W.-S. Chung, T. C. Peterson et al., “Neurotoxic reactive
astrocytes are induced by activated microglia,” Nature, vol. 541, no. 7638, pp. 481–487, 2017.
[251] A. C. Geraghty, E. M. Gibson, R. A. Ghanem, J. J. Greene, A. Ocampo, A. K. Goldstein,
L. Ni, T. Yang, R. M. Marton, S. P. Pa¸sca et al., “Loss of adaptive myelination contributes
to methotrexate chemotherapy-related cognitive impairment,” Neuron, vol. 103, no. 2, pp.
250–265, 2019.
[252] M. L. Monje, H. Toda, and T. D. Palmer, “Inflammatory blockade restores adult hippocampal
neurogenesis,” Science, vol. 302, no. 5651, pp. 1760–1765, 2003.
[253] S. A. Villeda, J. Luo, K. I. Mosher, B. Zou, M. Britschgi, G. Bieri, T. M. Stan, N. Fainberg,
Z. Ding, A. Eggel et al., “The ageing systemic milieu negatively regulates neurogenesis and
cognitive function,” Nature, vol. 477, no. 7362, pp. 90–94, 2011.
[254] P. Hasel, I. V. Rose, J. S. Sadick, R. D. Kim, and S. A. Liddelow, “Neuroinflammatory
astrocyte subtypes in the mouse brain,” Nature neuroscience, vol. 24, no. 10, pp. 1475–1487,
2021.
[255] J. S. Sadick, M. R. O’Dea, P. Hasel, T. Dykstra, A. Faustin, and S. A. Liddelow, “Astrocytes
and oligodendrocytes undergo subtype-specific transcriptional changes in alzheimer’s disease,”
Neuron, vol. 110, no. 11, pp. 1788–1805, 2022.
[256] R. P. Hebbel, J. Eaton, M. Balasingam, M. H. Steinberg et al., “Spontaneous oxygen radical
generation by sickle erythrocytes.” The Journal of clinical investigation, vol. 70, no. 6, pp.
1253–1259, 1982.
[257] E. Fibach and E. Rachmilewitz, “The role of oxidative stress in hemolytic anemia,” Current
molecular medicine, vol. 8, no. 7, pp. 609–619, 2008.
[258] R. Advani, E. Rubin, N. Mohandas, and S. Schrier, “Oxidative red blood cell membrane
injury in the pathophysiology of severe mouse beta-thalassemia,” 1992.
[259] A. Pokhrel, A. Olayemi, S. Ogbonda, K. Nair, and J. C. Wang, “Racial and ethnic differences
in sickle cell disease within the united states: From demographics to outcomes,” European
Journal of Haematology, vol. 110, no. 5, pp. 554–563, 2023.
Appendix A
Supplementary information for
Chapter 4
108
Appendix A 109
Figure A.1: RD and RK: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The left three
columns reveal a mixture of decreased and increased radial diffusivity (blue and magma color scales).
However, only smaller cluster of decreased RD are present in the non-SCD vs CTL comparison. The
middle three columns depict an overall decrease in radial kurtosis when controlling for logarithm
of age and sex. Additionally, the right three columns highlight statistically significant voxels when
the hemoglobin effects are also regressed out from the RK maps. The only remaining signal was
found in the RK model of SCD vs CTL, whereas non-SCD patients showed no statistical difference.
Moreover, in both cases, the RD maps showed no statistical differences when also removing the
effects of hemoglobin. These findings strongly suggest that anemia drives the observed RD’s
increments and RK’s decrements in the non-SCD cohort.
Appendix A 110
Figure A.2: AD and AK: T-maps displaying voxels that were statistically significant (p<0.05)
when comparing SCD|non-SCD patients to CTL individuals (upper|lower rows). The left three
columns reveal a mixture of decreased and increased axial diffusivity (magma and blue color scales),
while the middle three columns depict an overall decrease in axial kurtosis when controlling for
logarithm of age and sex. Additionally, the right three columns highlight statistically significant
voxels when the hemoglobin effects are also regressed out from the AK maps. The only remaining
signal was found in the AK model of SCD vs CTL, whereas non-SCD patients showed no statistical
difference. Moreover, in both cases, the AD maps showed no statistical differences when also
removing the effects of hemoglobin. These findings strongly suggest that anemia drives the observed
AD’s decrements|increments and AK’s decrements in the non-SCD cohort.
Appendix A 111
Figure A.3: T-maps depict voxels exhibiting statistically significant differences (p<0.05) when
comparing MK maps between individuals with Sickle Cell Disease (SCD) and healthy control (CTL)
individuals, utilizing magma color scales. All analyses were adjusted for the logarithm of age,
sex, and hemoglobin (Hb). Additional blood markers were systematically integrated to account
for any residual signal from the Hb model. Results, incorporating Leukocytes (Leuk) and fetal
hemoglobin (Hb F), reveal clusters with decreased MK closely resembling the Hb model. Moreover,
additional blood markers, including lactose dehydrogenase (LDH) and absolute reticular count
(ARC), elucidate the remaining Hb signal, as demonstrated in the far-left and right two panels,
respectively.
Appendix A 112
Figure A.4: ODI and FISO: T-maps displaying statistically significant voxels (p<0.05) when
comparing SCD|non-SCD patients to CTL individuals (upper|lower rows) while adjusting for
the logarithm of age and sex. The three left columns of the T-maps show a mix of decreased
and increased orientation dispersion index (ODI) values, with predominantly increased values in
non-SCD patients (represented by the blue color scale) and decreased values in SCD patients
(represented by the magma color scale). The three right columns show mostly decrements in free
water fraction (FISO) values in both cohorts. However, the extent of the clusters is greater in SCD
patients. No significant results were found when hemoglobin was added to the statistical model.
Abstract (if available)
Abstract
Worldwide prevalence of chronic anemia (CA) affects roughly 802 million people and creates a more significant burden than asthma, diabetes, and cardiovascular disease combined. CA is a condition where the number of erythrocytes or hemoglobin concentration is low, causing insufficient oxygenation. The resulting hypoxia is particularly damaging to the brain due to its high metabolic demand. Under acute circumstances, the brain receives preferential blood flow; however, this becomes insufficient when anemia is chronic, leading to hypoxia, neuroinflammation, and remodeling of white matter (WM). CA is commonly seen in patients with Hb synthesis or Hb structure disorders, such as thalassemia syndromes or sickle cell disease. These conditions have been associated with brain abnormalities such as volume loss, silent infarctions, and abnormal diffusion MRI (dMRI) parameters in the WM. Thus, the primary objective of the present study was to quantify the extent of WM damage resulting from hemoglobinopathy-induced CA using dMRI. This imaging technique identifies contrasts based on the movement of water molecules, enabling the inference of microstructural properties of WM. Various models can be derived from the dMRI signal. Our analysis focused on three different diffusion models: diffusion tensor, diffusion kurtosis tensor, as well as neurite orientation dispersion and density imaging. Our results revealed that the WM damage—long thought limited to the watershed areas— extends far beyond them, suggesting that WM remodeling may have a different pathophysiology.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Neurological consequences of chronic anemia
PDF
Unveiling the white matter microstructure in 22q11.2 deletion syndrome with diffusion magnetic resonance imaging
PDF
Cerebrovascular disease of white matter in patients with chronic anemia syndrome
PDF
Susceptibility-weighted MRI for the evaluation of brain oxygenation and brain iron in sickle cell disease
PDF
Methods for improving reliability and consistency in diffusion MRI analysis
PDF
Novel theoretical characterization and optimization of experimental efficiency for diffusion MRI
PDF
Decoding the neurological and genetic underpinnings of chronic pain
PDF
Diffusion MRI of the human brain: signal modeling and quantitative analysis
PDF
Diffusion MRI white matter tractography: estimation of multiple fibers per voxel using independent component analysis
PDF
Reproducibility and management of big data in brain MRI studies
PDF
Correction, coregistration and connectivity analysis of multi-contrast brain MRI
PDF
Assessing cerebral oxygen supply and demand in sickle cell anemia using MRI
PDF
Quantitative MRI for the measurement of cerebral oxygen extraction fraction in sickle cell disease
PDF
Brain-based prediction of chronic pain progression: a longitudinal study of urologic chronic pelvic pain syndrome using baseline resting state connectivity from the periaqueductal gray
PDF
Vascular contributions to brain aging along the Alzheimer's disease continuum
PDF
Relationships between lifetime chronic stress exposure, vascular risk, cognition, and brain structure in HIV
PDF
Age differences in diffusivity in the locus coeruleus and its ascending noradrenergic tract
PDF
Learning from limited and imperfect data for brain image analysis and other biomedical applications
PDF
Pattern detection in medical imaging: pathology specific imaging contrast, features, and statistical models
PDF
Neuroimaging in complex polygenic disorders
Asset Metadata
Creator
González Zacarías, Clio
(author)
Core Title
Anemia and white matter: a diffusion MRI analysis of the human brain
School
College of Letters, Arts and Sciences
Degree
Doctor of Philosophy
Degree Program
Neuroscience
Degree Conferral Date
2024-12
Publication Date
12/17/2024
Defense Date
12/12/2023
Publisher
Los Angeles, California
(original),
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
anemia,chronic inflammation,demyelination,dMRI,DTI,kurtosis,MRI,neuroinflammation,NODDI,OAI-PMH Harvest,sickle cell disease,thalassemia,white matter
Format
theses
(aat)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Thompson, Paul (
committee chair
), Kutch, Jason (
committee member
), Leahy, Richard M. (
committee member
), Mack, William (
committee member
), Wood, John C. (
committee member
)
Creator Email
cliog@usc.edu,cliogzphd@gmail.com
Unique identifier
UC11399EVHY
Identifier
etd-GonzlezZac-13697.pdf (filename)
Legacy Identifier
etd-GonzlezZac-13697
Document Type
Dissertation
Format
theses (aat)
Rights
González Zacarías, Clio
Internet Media Type
application/pdf
Type
texts
Source
20241217-usctheses-batch-1229
(batch),
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Repository Email
cisadmin@lib.usc.edu
Tags
anemia
chronic inflammation
demyelination
dMRI
DTI
kurtosis
MRI
neuroinflammation
NODDI
sickle cell disease
thalassemia
white matter