University of Southern California Dissertations and Theses

Pediatric magnetic resonance image processing: applications to posterior fossa cancer and normal development

(USC Thesis Other)

Pediatric magnetic resonance image processing: applications to posterior fossa cancer and normal development

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Pediatric Magnetic Resonance Image Processing:
Applications to Posterior Fossa Cancer and Normal Development
by
Jeffrey Tañedo

A Dissertation Presented to the
FACULTY OF THE USC VITERBI SCHOOL OF ENGINEERING
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)

May 2022

Copyright 2022 Jeffrey Tañedo
ii

Epigraph
We all change, when you think about it, we're all different people; all through our lives, and
that's okay, that's good, you've gotta keep moving, so long as you remember all the people that
you used to be. I will not forget one line of this, not one day, I swear.
-Steven Moffat

iii

Dedication
For my mother Flocerfida and her mother Gregoria

iv

Acknowledgements
I begin with the beginning of me and thank my mother, Flocerfida Ocampo. You have
fought an uphill battle your entire life so that I may have the opportunities you dreamed of when
you were a child. This work and these words that I dedicate to you only skim the surface of the
depths of gratitude, respect, admiration, and love that I have for you. You have been the
inspiration in my life to keep going when the going gets rough. I love you and thank you, Mom.
Thank you to my committee. First, I thank Dr. Natasha Lepore for giving me a chance
and welcoming me to the CIBORG lab as I was contemplating leaving the PhD program. The
warmth of your genuine empathy and readiness to help guide me have kept me afloat all these
years and it has been many. Thank you not only for your guidance in scientific endeavors but
also for reminding me to remember the people both in front of and behind the research. You
have always been able to pull me back to see the bigger picture when I’m most lost in the
weeds. You have been an excellent researcher, professor, mentor, and above all, an excellent
human being.
Thank you, Dr. Michael Khoo, for all your support. I still remember working in your lab in
Spring 2015 when I didn't have funding. That saved me that semester and I am still grateful for
that.
Thank you, Dr. John Wood. I still remember when Eamon introduced me to you at CHLA
back in 2014 when I was transitioning between labs. The kindness you showed me when we
spoke showed me how encouraging advisors can be. All of that factored into my eventual
decision to come and work at CHLA with Natasha and Mary.
Thank you Dr. Vidya Rajagopalan. You have been thoughtful and supportive in
challenging me to re-examine my assumptions when it comes to research. You were also
incredibly supportive during the many times I felt down.
v

Thank you, Dr. Darryl Hwang. You have been one of the three elder siblings I felt I have
found in this lab’s family. You have challenged and taught me not only in matters of biomedical
engineering, but also in dance, martial arts, education, and general awesomeness despite some
of your worst jokes. Thank you for all your advice throughout the roughest points of the PhD.
Thank you to all the people in CIBORG lab who have shaped not just the PhD but my life
as well.
Thank you Dr. Niharika Gajawelli. You have been the second of the three elder siblings.
Your invitation for me to visit CHLA began all of this. Throughout all these years, you have been
a grounding presence for me. Beyond that, you even gave me opportunities to collaborate and
work with you. I’ve always been empowered by the trust you have given me that I can contribute
to your work.
Thank you Dr. Sinchai Tsao. You have been the third of the three elder siblings. Just like
Darryl and Niharika, you’ve guided me through the PhD, but specifically you’ve taught me to
always keep an eye on the future. I can’t say I’ve totally gotten the hang of it, but I at least feel
comforted by your humility in admitting you’re still figuring things out too. We all are, and I think
acceptance of that gives rise to true compassion.
Thank you, Dr. Natacha Paquette. Thank you for being the friend and the light I needed
during my darkest times in the lab. Literally, we kept the lights half on because it was too bright.
Seriously though, our conversations helped balance the emotional turmoil that sometimes
comes with the PhD.
Thank you, Dr. Mary Nelson, for your mentorship and for helping me get started in the
lab. The work we did on the posterior fossa tumor pediatric studies formed a large part of my
understanding of MRI processing and a large part of my thesis. I am grateful for the
opportunities you have given me.
vi

Thank you, Dr. Julie Coloigner, Dr. Yaqiong Chai, Dr. Matt Borzage, Dr Iris Miaowei, Dr.
Yi Lao, Dr. Sharon O’Neil, Dr. Marvin D. Nelson, Dr. Roza Vlasova, Dr. Arthur Olch, Dr. Kenneth
Wong, Dr. Benita Tamrazi.
Thank you to all the collaborators outside of USC and CHLA. I believe that your
willingness to share your ideas and successes are the only true path for scientific advancement.
Thank you Dr. Yalin Wang, Dr Sean Deoni, Dr. Siri Khalsa, Dr. Cormac Maher, Dr. Qunxi Dong,
Dr. Pablo Hernáiz Driever, Dr. Daniela Saccheto and Dr. Adam Bush.
Thank you to all the research assistants in the lab. I hope you learned as much from
working with as we have learned from you. I have had the wonderful opportunity to become
friends with many of you have impacted my life beyond research. Please know that you have
also impacted all the lives that this research will reach as well. Thank you Athelia Pauli, Sharon
Guo, Malia Valder, Adam Walker, Jonathan Hong, Katrina Huft, Christine Obioha, Scarlet Cho,
Andrea Ezis, Yeun Kim, Brittany Randles, Nick Chapman, Aditi Tondulkar, Sofia Dhanani,
Katherine Maina, Hannah Telle, Ipek Narbay, Shringala Chelluri, Dixith Reddy Gomari,
Francesca Trane, Binh Nguyen.
Thank you also to all the staff at CHLA who have helped smooth out the protocols and
miles of paperwork that come with research. I have felt the compassion you have towards just
helping a fellow human. It has been a pleasure to work with all of you. Thank you, Erika Keka
Martinez, Sandra Esquer, Yuri Trujillo, Nathan Menard, Emy Arango, Kayla Marie Guzman,
Julia Castro, Elizabeth Kim.
Thank you to the faculty and staff at USC. Thank you Mischal Diasanta, you are the
reason I was able to come out to Los Angeles in the first place. You were patient with me and
worked with me as I tried to find funding for my first year and you were accommodating and
supportive all the way until you left USC. Thank you to William Yang. When you took over for
Mischal as the BME Graduate Advisor, you had big shoes to fill but you have filled them and
then some. I have felt supported by you all the same since you’ve started. Thank you, Dr. Jean-
vii

Michel Maarek, for being an amazing educator and an inspiration for me as I seek to pursue my
own teaching career.
Thank you to all my family and friends who have supported me along the way. The color
of your lives will forever be on my life’s canvas. Thank you Roel for being my best friend and
support throughout all these years since high school. Time to crack open that Patron when I get
home because this counts as a life event. Thank you Mey Mercado for being my little sister. As
much as we love saying we hate each other, I have always felt the care and love we have
shared. Thank you, Tito Roel Mercado and Tita Marilou Mercado, – you have been my second
family and my safe haven since I was in high school. Thank you for allowing me to stay at your
house for so many nights, for inviting me to share so many meals, and of course thank you for
raising two of the most important people in my life. Thank you, Dad, for supporting me and Mom
as much as you could. Thank you Quelmor Tañedo for reminding me what it’s like to be a
teenager. You are already worthy and brilliant, and I hope you can see that one day. Thank you
to my partner, Sarah Berry. Your love, support and acceptance of me throughout the end of this
PhD helped remind me to love, support and accept myself. I hope you have felt those same
things from me as I grow in love for me, for you, and for us. Thank you, Dr. Lindsay Kramer. I
know you’ll say I have done all the work, which is true. But you have calmly and patiently shown
me the path to that work. Thank you for you and for your work. Thank you, Austin Carter, for
being my brother these last few years. We have relied on each other and supported each other
through some of the toughest moments of both of our lives and I know we’ll continue that for the
rest of our lives. Thank you, Lorena Bravo and Jonathan Atkinson, for sharing the love and
discipline of dance with me. Everything I do now is some form of dance. Thank you to my DnD
group for sharing your stories, your enthusiasm and your creativity. You have kept me sane
these last few years despite all our insane adventures – Blusean Velasco, Shoaib Zaheeruddin,
Nilhson Zelaya, Bob Liu. Thank you, Dara Wedler, for being that glue that keeps our friend
group hanging out even as we get older and start to drift away from the things that once kept us
viii

in the same room. Thank you, Dr. Neil Agarwal, Dr. Max Pflueger, Chris Girard, Dr. Kaveh
Shahabi, Dr. Sadaf Soleimani, Dr. Samantha McBirney, Adam Mergenthal, Ben KangWoo Lee,
Dr. Kathleen Lo.
Finally, thank you D.U. You have stuck by me my entire life. We have laughed together,
cried together, failed, won, survived, and thrived together. No matter how many times you ask
me if it’ll be ok, I will always tell you, “Yes, you’ll be fine”, but now I want to add, don’t worry so
much. You got this.

ix

Table of Contents
Epigraph ..................................................................................................................................... ii
Dedication .................................................................................................................................. iii
Acknowledgements .................................................................................................................... iv
List of Tables .............................................................................................................................. x
List of Figures ............................................................................................................................ xi
Abstract.................................................................................................................................... xiii
Chapter 1. Introduction ............................................................................................................... 1
Chapter 2. Brain Biomarkers and Neuropsychological Outcomes of Pediatric Posterior Fossa
Brain Tumor Survivors Treated with Surgical Resection with or without Adjuvant Chemotherapy
.................................................................................................................................................27
Chapter 3. White Matter Tract Changes in Pediatric Posterior Fossa Brain Tumor Survivors
after Surgery and Chemotherapy ..............................................................................................49
Chapter 4. Normal Cerebral Ventricular Shape in Early Childhood............................................67
Chapter 5. Conclusion and Future Work ..................................................................................82
Publications ..............................................................................................................................86
References ...............................................................................................................................88

x

List of Tables
Table 1. Patient demographics by treatment group compared to controls .................................39
Table 2. Neuropsychological assessment scores and fractional anisotropy R values * .............42
Table 3. Patient demographics by treatment group compared to controls .................................55
Table 4. Distribution of subjects in different age groups ............................................................70

xi

List of Figures
Figure 1. Example of Aliasing in T1 MR image of brain .............................................................13
Figure 2. T1 MR Image, Sagittal slice of human brain before bias field correction (left) and after
bias field correction (right) .........................................................................................................14
Figure 3. T1 MRI of human brain, before skullstripping (left), and after skullstripping (right) ......15
Figure 4. Example T1 MR Image Brain Template ......................................................................17
Figure 5. Top Left Histogram of Patient Ages, Top Right Histogram of Controls Ages, Bottom
Histogram of All subject ages ....................................................................................................19
Figure 6. Example of Diffusion Weighted Image sagittal slice with interslice signal intensity
drops .........................................................................................................................................22
Figure 7. Skull stripping of DWI in DSI Studio ...........................................................................23
Figure 8. DWI, Sagittal Slice before EPI correction (top left), after EPI correction (top right), with
example of T1 registration target (bottom) .................................................................................24
Figure 9. DTI Eigenvalues on brain slices in ITK-SNAP; Lambda 1 (top left), Lambda 2 (top
right), Lambda 3 (bottom) ..........................................................................................................25
Figure 10. Image Processing details of single subject and group-level image processing. Abbr.:
AFNI, analysis of functional neuroimages; ANTS, advanced neuroimaging tools; DWI, diffusion-
weighted imaging; FSL, FMRIB software library; NIHPD, NIH pediatric data .............................33
Figure 11. Areas of fractional anisotropy (FA) differences between patient groups in composite
axial and sagittal diffusion tensor images. .................................................................................42
Figure 12. WM Tract Results in the Surgery vs Healthy Control comparison. All results displayed
on a transparent glass brain. The background tract is in blue while the significant clusters after
multiple comparisons correction are displayed in red. The three images, left to right, display
results in a leftward facing sagittal view, inferior facing axial view, and rightward facing sagittal
xii

view in the (a) Corpus callosum (CC); (b) Bilateral corticospinal tract (CST); and (c) Right
inferior frontal occipital fasciculus (right IFOF)...........................................................................59
Figure 13. WM Tract Results in the Surgery and Chemotherapy vs Healthy Controls
comparison. All results displayed on a transparent glass brain. The background tract is in blue
while the significant clusters after multiple comparisons correction are displayed in red. The
three columns, left to right, display results in a leftward facing sagittal view, inferior facing axial
view, and rightward facing sagittal view in the (a) Corpus callosum (CC); (b) Bilateral
corticospinal tract (CST); (c) Right inferior frontal occipital fasciculus (right IFOF); and (d) left
uncinate fasciculus (left UF) ......................................................................................................61
Figure 15. Global significance maps displaying MADMTBM between the 1st and 2nd year old
groups (upper left), 2nd and 3rd year old groups (upper right), 3rd and 4th year old groups (left
second row), 4th and 5th year old groups (right second row) and 1st and 5th y year old groups
(bottom row). * Indicates that the whole map for that ventricle and age comparison is significant.
.................................................................................................................................................74
Figure 16. MAD Mean Ratio. Yellow and red colors indicate areas of greater ventricle thickness
in the older age group. Light blue to blue indicates areas of less thickness in the older age
group. .......................................................................................................................................75
Figure 17. Bilateral ventricular ratio. Gray dots represent individual data points. Curved lines
represent from top to bottom 95th, 90th, 75th, 50th, 25th, 10th, and 5th percentile curves ........76
Figure 18. Standard deviation of medial axial distance values across each individual subject
group in mm ..............................................................................................................................77

xiii

Abstract
This thesis work is composed of Magnetic Resonance Imaging (MRI) studies on the
pediatric brain in posterior fossa tumor survivors in response to treatment such as surgery,
chemotherapy, and radiation as well as MRI studies on the normal developing pediatric brain.
We have analyzed both the structural T1-weighted MR and Diffusion Weighted Images
(DWI) of the brains of pediatric survivors of posterior fossa brain tumors. The main goal of this
first project was to understand the differential effects of treatment on the neurobiology of the
pediatric brain. Many studies have shown that surgical tumor resection, chemotherapy and
radiation have continued detrimental effects on the developing pediatric brain beyond the tumor
site. While these treatments are necessary for survival, the neuroanatomical biomarkers of
neurotoxicity for each of these treatments are not well understood, especially in distinction from
one another. Thus, the first part of this project has been to compare the neural structure
between groups who had received increasing levels of treatment to parse out the difference in
treatment effect. After an initial voxel-wise supratentorial brain FA analysis, we found several
sites of FA differences in comparisons between the treatment groups and healthy controls.
Some of these sites were correlated with known white matter tracts.
These results informed the next study on the same cohort wherein we focused the
analysis to the specific white matter tracts found in the first study. We found clusters of
significantly different FA between the patient groups and the healthy controls, but none were
found between the patient groups who had received differential treatment. This supports prior
literature which theorizes damage to the supratentorial parts of the brain far away from the
tumor site due to some effect of the tumor or treatment. This also highlighted how difficult it is to
study pediatric brain tumor populations. In the process of analyzing the data, we had to exclude
data due to inadequate normalization. We found much of this was due to abnormally large
cerebral ventricles.
xiv

The large deformations necessary to transform brain images with large ventricles create
inadequate registration results. We realized that an understanding of how the presentation of
ventricles develops in healthy children was needed first. In fact, after conversations with
collaborating neurosurgeons, we found that there was a need for a more refined understanding
of developmental variability to better understand pathological variability. Most analyses on
healthy development of the brain involved either linear metrics on a single axial slice of an MR
image or volumetric analyses – both of which summarize whole brain structures and tissues as
a scalar value. However, there may be scenarios where indicators of pathology may lie not in
volume but in finer morphometric changes that would otherwise be averaged out in gross
analyses.
Towards that end, we worked to characterize the development of healthy pediatric lateral
ventricles by applying a multivariate tensor-based morphometry method to localize changes in
shape of the ventricular system. We worked with our collaborator, Dr. Sean Deoni, and his T1
MRI database of healthy children. Thus, we were able to identify several areas of significant
shape difference between each consecutive year between ages 1 through 5. In addition, we
were able to characterize the directionality of such changes by calculating the normalized
medial axial distance ratio between ages and displaying those ratios on models of the
ventricular surface. Finally, in an attempt to create a normative map of expected thickness
variability of the ventricle, we also displayed the standard deviation of the medial axial distance
values on the ventricular surface. These characterizations provide a baseline to be able to
separate typical developmental variability from pathological variability. The knowledge discerned
also provides neurosurgeons with another metric to evaluate dysmorphic ventricles with
otherwise normal volumes.
1

Chapter 1. Introduction
Central nervous system (CNS) tumors are second in frequency only to leukemia among
cancers affecting children but are still the most common cause of cancer death [Udaka, Yoko T.,
et al., 2018] in children ages 0-14 years in the United States with an incidence rate of
approximately 5.83 per 100,000 person-years [Ostrom, Quinn T., et al., 2017]. Among patients
in their first decade of life, brain tumors are most commonly located in the posterior fossa.
The current standard of care for most brain tumors involves a combination of surgical
resection, chemotherapy, and cranial irradiation. Developments in these therapies as well as
earlier detection and improved post-treatment monitoring have increased the average 5-year
survival rate for pediatric brain tumors for ages 0-14 years from 59.1% in 1977 to 76.5% in 2016
[SEER].

Unfortunately, both the lesion and treatments may induce potentially debilitating long-
term neurodevelopmental sequelae as children are most susceptible to the adverse effects of
treatment during this period of significant development in their lives. Hence, with an increasing
population of survivors, the effort to study late-onset adverse effects is also becoming
increasingly important. These long-term effects include but are not limited to stroke, seizure,
hearing loss, and learning disabilities [King, A., et al., 2016]

While some chemotherapeutic
agents are known to be neuro-toxic, [Anderson, F.S., et al., 2009]

cranial radiation therapy is the
most significant contributor towards increased risk of neurocognitive decline. The physiological
cause of this decline may be due to necrosis of white matter tissue caused by the death of
oligodendrocytes [Siu A., et al., 2012].
Numerous studies [Jariyakosol S., et al., 2015; Gupta P., et al., 2017; Rey-Casserly C.,
et al., 2019; Baron Nelson, M., et al. 2021] have shown how both the tumor and the treatments
have led to deficiencies in physical, social, emotional, and cognitive domains, which result in an
overall decrease in quality of life for survivors. Surgical resection of posterior fossa tumors can
2

result in deficits to motor speech production [Huber, J. et al., 2006], visual function [Peeler, C.,
et al., 2017], and cognition [Mulhern RK, et al., 2004]. Depending on the region being irradiated,
radiation dose has been associated with impairments in memory, physical performance, and
social functioning [Armstrong, G, et al., 2010]. Current research on the late effects of
chemotherapy in CNS patients is sparse. However, a meta-analysis of these late effects in
childhood leukemia survivors shows that chemotherapy contributes cognitive deficits [Iyer NS,
et al., 2015].
While the exact neurobiological mechanisms leading to these adverse outcomes are not
yet understood, surgery, chemotherapy [Ikonomidou, C. 2018] and radiotherapy [Kim, J. et al.,
2008; Jacob, J., et al., 2018] have been shown to induce additional neurological injury which
can be monitored and analyzed through magnetic resonance imaging (MRI). Using voxel-based
morphometry (VBM) methods to quantify brain tissue volume, global reductions in both grey and
white matter (WM) volumes have been seen and correlated with neurocognitive decline after
treatment [Ailion, A, et al., 2017]. WM is important in mediating the functional connectivity for
many neurobehavioural operations [Filley, C., et al., 2016] and is susceptible to damage from
radiation and chemotherapy. Thus, particular attention has been given to WM tract alterations
through the use of diffusion-weighted imaging (DWI).
DWI is an MRI method which captures the diffusion of water molecules through
structures in the brain, especially to WM structures. Healthy WM typically consists of bundles of
myelinated axons organized into tracts which connect different parts of the brain. The restricted
flow of water molecules through these bundles can be characterized through a metric called
fractional anisotropy (FA), which is obtained through a processed form of DWI called Diffusion
Tensor Imaging (DTI) [Bammer, R. 2003]. Lower FA values in a particular region may indicate
more unrestricted flow of water molecules and thus can be indicative of a loss of microstructural
integrity, reduced bundle organization or axonal damage [Rueckriegal, SM., et al., 2010].
3

Several studies have demonstrated lower FA in patient groups after receiving treatment,
thus possibly indicating WM damage due to treatment. The combination of chemotherapy and
radiation has been shown to be detrimental with many studies documenting reduced FA in
several brain structures globally, including the corpus callosum and frontal WM [Fouladi, M., et
al., 2004]. Although literature on solely chemotherapy’s effects on pediatric CNS brain tumor
patients is sparse, we can extrapolate from a recent study of its effects on pediatric leukemia
survivors. This study indicated a reduction in grey and WM volume in addition to a global
reduction in FA values [Deprez, Sabine, et al., 2013]. However, before either adjuvant therapies,
surgical resection of posterior fossa tumors alone impacts the supratentorial brain, as evidenced
by lower FA and decreases to WM volume in structures such as the corpus callosum and
corona radiata [Glass, J., et al., 2017]. Thus, examining differences in FA values is important in
determining WM damage in relation to different treatments.
Group Image Techniques and Analysis
For medical applications, individual MR and DW images are critical for the diagnosis of
some diseases and injuries. If we wish to begin studying groups of patients for commonalities
across diseases, treatments, or normal development, we need tools to help distinguish
interesting patterns from several images.
Image Registration
Image registration in general is simply the process of transforming an image from one
coordinate space to another. This can be across modalities, for example, registering a DWI or
CT image to T1 or T2 MRI space. Registration can also be across dimensions, for example, in
tensor-based morphometry- registering a three-dimensional mesh to a two-dimensional
conformal map.
4

When registering two 3-D images, we can define the nature of the deformation as such-
Rigid registration has six degrees of freedom across three dimensions - translation and rotation
(2) in the x, y, z plane. Affine registration expands to having twelve degrees of freedom, adding
the ability to shear and scale in the three planes. You can also think of this as transforming
parallel lines to parallel lines. Finally, mapping lines to curves is considered nonlinear or elastic
registration.
The motivation for registration is different at each stage of preprocessing and will be
discussed in the workflow section.
Templates
One routine step in MR image processing is the spatial normalization of subject scans to
enable comparison between subjects and between groups of subjects. This is accomplished
through a process of registration towards a common template. Templates provide a priori
information about the expected voxel intensities of the expected image of the population brain.
In the voxel-based methods which will be described, statistics are calculated voxel by voxel
across patients. Thus, nonlinear registration is important to ensure that values from the same
brain structure are sampled across the patients in the group. However, it is important to note
that every registration degrades the data and introduces interpolation error with larger
deformations. In pediatric imaging, registration to an adult template may cause significant
interpolation error due to the large geometric transformations required to match a pediatric brain
to an adult brain. One study on pediatric neuroimaging has shown that usage of an adult
template vs a pediatric template can influence measures of surface and volumetric morphology
of the registered subject brain scans [Yoon, Uicheul, et al., 2009]. These differences would
introduce bias to the group analysis statistics leading to spurious results or obscuring true
results. Therefore, the choice of template must be deliberate in its attempt to represent the
characteristics of the group being studied.
5

Tract Specific Analysis
Many approaches have been developed to identify the FA values in WM tracts. Manually
drawn regions of interest (ROIs) have been used to ensure anatomical accuracy. However,
voxel-based methods (VBM) like this have been replaced by the development of automated
methods of WM identification as the number of subjects in studies has increased. Tract-based
spatial statistics (TBSS) [Smith, SM., et al., 2006] became the standard automated method in
the study of WM tracts as it allowed researchers to process more data without sacrificing
anatomical accuracy. TBSS accomplishes this by projecting volumetric data, such as FA values,
onto a WM skeleton. With the method’s rise in popularity, researchers have become
increasingly aware of its limitations. For example, TBSS’s projection onto an entire WM skeleton
does not allow a researcher to distinguish between distinct but adjacent WM tracts [Bach,
Michael, et al, 2014].
A newer method, tract specific analysis (TSA), addresses this pitfall by segmenting
individual WM tracts onto population specific templates [Pecheva, Diliana, et al., 2017]. Tensors
are then projected onto a medial sheet which both defines the skeleton and informs the
boundary of a WM tract. The maximum or mean tensor values can be calculated along a spoke
extending perpendicularly from a point on the medial sheet to the tract boundary. From these
values, DTI metrics such as FA can be calculated. Thus, TSA can provide FA values for
specific WM tracts without noise from adjacent tracts. A previous study from this lab comparing
TBSS and TSA has shown that in a comparison between a congenital blind group and healthy
sighted controls, TSA shows higher sensitivity in detecting subtle differences in WM [Lao, Yi, et
al., 2015].

6

Statistics
Multiple Comparisons Correction
With multiple simultaneous statistical tests, there is a possibility that some tests will
falsely reject the null hypothesis by pure chance. This is also called a Type I error. When
applying statistical tests to whole brain analyses of MR images, this problem is exacerbated by
the sheer number of voxels being tested. Human brain volume is on the order of 1000 cm
3
. If
the spatial resolution of the MRI is 1mm
3
and all voxels in the brain were tested, this would
amount to 1 million statistical tests.
The Bonferroni correction method was one of the first methods to correct for multiple
comparisons. If the significance level was set to 𝛼 = 0.05 and n = 100 statistical tests are
performed, then the null hypothesis would only be rejected if the p-value of a test were 𝛼 /𝑛 =
0.05 ∗ 10
−2
. This is much too conservative for statistical tests involving MR images.
Instead of rejecting all Type I errors, the proportion of acceptable Type I errors can be managed
by controlling the False Discovery Rate (FDR). If the FDR were set to 𝛼 = 0.05 , that means
accepting that 5% of significant results will be false positives.
One method to address FDR is to use the Benjamini-Hochberg procedure. Of the parametric
methods to correct for multiple comparisons, this is one of the most popular.
The intuitive explanation for the procedure is as follows:
Find p-values for all statistical tests planned
Sort and assign a rank for all p-values such that the smallest p-value has a rank of 1 and
the next p-value has a rank of 2.
Calculate the Benjamini Hochberg critical value,
𝑖 𝑚 𝑄 for each p-value, such that i = rank
of p value, m = total number of tests, and Q = FDR 𝛼 value.
7

Looking at the other end of the sorted p-value list. Find the largest p-value, pL <
𝑖 𝑚 𝑄 and
every p-value < pL is significant.
Another popular practice is a non-parametric statistical algorithm involving permutations of data
labels to approximate the null distribution.
Cluster-based Permutation Analysis
Clusters are defined by areas of contiguous significant voxels in MR statistic tests. The
intuition is that significant differences that can be seen at the resolution of MR images are most
likely going to be groups of voxels rather than singular lone voxels. Cluster-based permutation
analysis between two groups of data begins by assuming that no difference exists between the
groups. Another way of saying this is that the null hypothesis is true for all data. Thus, the
group labels of data can be randomly assigned without consequence. Within this new
assignment of groups, statistical tests can be performed. The number of voxels (i.e., the area) of
the largest cluster can then be recorded. Once again, the labels of the data are randomly
assigned or rather, permuted, and the process begins again. In this way, a histogram of largest
clusters can be obtained. This histogram represents the null distribution. A statistical test on the
original labeling of the data will then give clusters of different sizes, which can be compared to
the null distribution and thus given a cluster-wide significance value.

8

Study Responsibilities
Population Criteria
The main dataset I have studied consists of patient participants who met the following
criteria: (1) age at the time of study was between 6-17 years old, inclusive ; (2) at least one year
had elapsed since the participant’s last therapy; (3) tumors were located in the posterior fossa,
either in the cerebellum or fourth ventricle; and (4) all tumors were completely resected with no
evidence of tumor or metastasis at least 1 year after treatment with either surgery or surgery
chemotherapy or surgery, chemotherapy and radiation.
Control participants in the study all met the following criteria: (1) age at the time of study was
between 6-17 years old, inclusive (2) participants had no prior history of traumatic brain injury or
mental disease.
All participants were screened for exclusion by the following criteria: (1) metal in the
body which prevented them from taking an MRI, (2) preterm birth, (3) developmental disabilities,
(4) traumatic brain injury or (5) turning 18 through the duration of their participation in the study.
Controls were excluded if they were not able to sit still in an MRI for thirty minutes without
sedation. Patients were excluded who (1) had a history of recurrent tumor or residual disease
outside of the posterior fossa, (2) had a history of posterior fossa syndrome, (3) had been
treated with radiation, (4) who were no longer followed at the hospital.
All participants were required to undergo a three-hour neuropsychological assessment
as well as a minimum 20–30-minute MR scan to obtain a T1 and DTI.

9

Recruitment
The original study design was to collect data from 10 healthy sibling controls, 20 children
treated with surgery only, 20 treated with surgery/chemotherapy and 10 children treated with
surgery/chemotherapy and radiation. The study was extended to include more patients and
different analyses. In the end, we obtained data from 73 patients in total - 17 children treated
with surgery only, 11 children treated with surgery and chemotherapy, 19 children treated with
surgery, chemotherapy, and radiation, 5 children treated with surgery and radiation, and finally
21 healthy controls.
Throughout that timeframe, I would regularly search through patient databases and
neuro-oncology lists to find potentially eligible subjects for the study. Upon confirming eligibility
with Dr. Mary Baron Nelson, or if alerted to a potentially eligible subject, I would initiate contact
with the primary guardian, parent, or household and thus begin the recruitment process in one
of four ways 1) phone call, 2) stamped mail, 3) email, or 4) greeting them during a regularly
scheduled clinical visit. Phone calls and greeting patients in person proved to be the most
successful in at least beginning a conversation.
Research Coordination
During that conversation, in phone or in person, I would give a brief outline of the study’s
purpose, design and requirements and ask if both the family and the patient or healthy control
would be interested in participating. An affirmative response would allow me to move on to the
next phase of coordinating the schedules of the family, the radiology scheduling team, the
radiology technicians, the neuropsychology team and - for some Spanish-speaking families - a
hospital translator. In the days and weeks leading up to their appointments, I would also check
in with all aforementioned groups via phone, email or in person to confirm scheduling, technical
aspects of the scan, or logistics such MR machine availability or quiet neuropsychological
assessment rooms.
10

Data Acquisition and Quality Control
On scheduled days or nights, I would meet the patients at the hospital, and guide them
through the consent process. Upon confirmation of their signatures, I would produce the
necessary copies for our records, the family’s records, and the hospital’s records.
During their MR scans, I reminded the patient to stay still. I sat with the MR technician to assure
the same scan parameters were used as well as to confirm scan quality. Before subjects left the
hospital (preferably before they left the scanner), I would also process the images from raw
DICOM data type to the more research friendly NIFTI data type to quickly scan for abnormalities
or critical noise artifacts.
Upon completion of their scan, I would thank the patient and their family for participating
and escort them to their next appointment, whether it be a regular clinical visit or the
neuropsychological examination.
Database Management
Before any of the data was obtained, my first contribution was building the computer
station, which would eventually house the study data and act as the main processing unit for the
image-heavy calculations.
After obtaining the DICOM data from the scanners, I would then enter the workflow
described in the next section. Data from each stage would be saved separately and into a file
structure I devised so that we could locate and edit data. Considering that there are no pediatric
MRI-specific research programs, I needed to experiment with different tools at each step of
processing. Often, tools would not work or would produce nonoptimal results. Thus, the data
structure I created allowed me to revisit every step of the trial-and-error process. I was also in
charge of uploading and auditing neuropsychological and demographic data into the Research
Electronic Data CAPture (REDCAP) resource at USC.

11

Challenges
There are many difficulties along each stage of developing and processing such a
dataset of pediatric MR images. Initially, the search to find eligible patients turned up many
possible results, but further inspection was required to confirm eligibility. I would go through lists
of hundreds of patients, searching through their records to confirm that the patient had no
history of recurrent tumor, excessive hydrocephalus, or other diseases.
Many times, patient families would not respond back either by phone or by mail.
Sometimes, the family had moved, or otherwise had different contact information. There were
also, of course, patients’ families who declined to be in the study. Of the patients who did
respond and agreed to be in the study, there were some who did not show up to one or both the
MR scan and the neuropsychological study.
Since we are scanning children, the likelihood of their motion creating noisy and
unusable data was high. Thus, I was present during most scans to check that the data was still
usable. This was especially critical during the 13-minute DTI sequence. Often, children would
move during this period. As a result, I had to manually extract Diffusion Weighted Imaging
volumes within the sequence to continue using the rest of the data. Otherwise, I was able to
scan patients again. A few times, we were able to do this on the same day while they were still
in the scanner. Other times, we had to scan them on the day of their neuropsychological
examination. These difficulties were just with collecting the dataset. New challenges arise with
the processing of the MR data.
The current research tools to process MR data are primarily geared towards adults
and the aging population. The difficulty of even obtaining MR datasets of children, for the
reasons above, has led to the scarcity of pediatric specific tools. Compared to adult MR images,
pediatric MR images suffer from high noise, low contrast, and anatomical variability. For this
reason, every step of the preprocessing pipeline had to be tested with different software to
confirm that the tool or software we were using was correct for the population. For example,
12

there are very few templates for children for registration. Throughout the course of the project,
we’ve switched the target template to account for the evolving characteristics of the dataset as it
grew. The details of this and the development of the final template used are in a later chapter.
The scans were restricted in quality to reduce time spent in the MRI. Otherwise, our protocol
would impact MR scheduling, would incur fees associated with using the scanner, and would
exhaust the patient’s ability to lie still during an extended MR session. In short, obtaining and
processing a pediatric MR dataset is difficult.

13

Pediatric Specific MR Processing Workflow
T1 Preprocessing
Visual Assessment

Figure 1. Example of Aliasing in T1 MR image of brain

First, images were assessed during the scan and immediately after for artifacts such as
zipper artifact, aliasing, etc. If not caught during the clinical scan session, patients would need to
be rescanned at the end of the clinical sequence, rescheduled for another scan, or excluded
from the study. Above is an example of Ringing or Gibbs artifact in Figure 1. If a patient moves
during scanning, there’s a chance that the scanner is unable to sample higher frequencies at
boundaries such as that between the head and surrounding air.

14

Bias Field Correction

Figure 2. T1 MR Image, Sagittal slice of human brain before bias field correction (left) and after bias field correction
(right)
In every MRI acquisition, on top of the true image, there is an additional low frequency
layer of noise that is commonly called the bias field [Vovk, Uro, et al., 2007] as shown in Figure
2. The field is noise introduced by several sources: the properties of the MRI device itself, eddy
currents driven by the field gradients, and dielectric properties of the object or person being
scanned. The bias field has no significance in clinical studies and if left untouched, could
obfuscate any process or analysis over the MR image which relies on voxel intensity. Thus, I
begin with Bias Field Correction (BFC). The established method for that is to use Nicholas
Tustison’s N4ITK algorithm [Tustison, Nicholas J., et al., 2010] which is built into most MR
research computer programs. After reviewing the bias field images extracted by BFC, I found
that applying the correction 2-3 times is sufficient in removing the bias field from the image.
Thus, the pipeline requires BFC three times before the next step.

15

Skullstripping or Brain Extraction

Figure 3. T1 MRI of human brain, before skullstripping (left), and after skullstripping (right)
Studies of the brain typically focus on the brain matter itself. Thus, all non-brain material
such as the skull or dura are removed as shown in Figure 3. The standard of MR research has
been FMRIB Software Library’s (FSL’s) Brain Extraction Tool (BET) [Smith, Stephen, et al.,
2001] which is included in many other programs. Recently, Brainsuite released an open-source
brain extraction algorithm, Brain Surface Extraction (BSE), which is quickly replacing FSL’s BET
[Rajagopal, Gautham., et al., 2017]. With any automated program, there is the possibility of
failing to remove certain elements which are not brain matter. Thus, time intensive manual
editing of the skull stripping is needed. However, BSE improves on BET by adapting itself to the
image and calculating the best parameters for extraction, reducing the number of hours needed
for manual editing. Despite that, the need for manual editing still exists. In this study, the manual
editing of these brain extractions took anywhere between 10-20 manhours per brain.

16

Image Registration
Image registration is critical to generalize MR studies across groups of patients.
Comparisons of the brain should be independent of any potential differences in spatial
resolution, field of view, or orientation.
In MR image processing, the target of this image registration is called a template and the
coordinate space associated with it is the template space. With brain images, the standard has
been Montreal Neurological Institute’s or MNI template space. Below is an example of one such
template for 7.5- to 13.5-year-old subjects.
I found FSL’s FMRIB Linear Image Registration Tool (FLIRT) and FMRIB Nonlinear
Image Registration Tool (FNIRT) to be the best at automatic registration for T1 weighted
images. Insight Toolkit’s ITK-SNAP has also been useful in manual registration. I found it
especially useful in multi-modality studies as seen in my conference submissions on CT
dosages and MR images.
17

Figure 4. Example T1 MR Image Brain Template

A linear transformation is first applied to address artifacts caused by orientation or pose
of the subject in the scanner. Nonlinear transformations are then only applied if the desired goal
is to analyze differences in qualities of the tissue such as the DTI metrics in a tissue. If the goal
were to research the morphometry or size of structures in the brain, then registration would not
stretch beyond simple translation and rotation to preserve discerning information. The object of
these transformations is called a template as shown in Figure 4.

18

Template
In a review by Dickie et al. 2017 of all whole brain MR templates created up to 2017, 4
groups have created templates of healthy and demographically representative children covering
the age range of our study (6-year-old. - 17-year-old)
Overall, MNI pediatric atlases from Fonov et al 2011 were chosen for the following reasons:
- The use of their templates is well documented across many journal articles
- Their templates are divided into ranges based off pediatric developmental phases
- Avants et al. only has 1 template for the whole age range
- Sanchez et al. has templates in increments of 0.5 years, which complicates the
decision process in covering our 11-year age range.
- Wu et al. breaks their group into 4-8 y.o., 9-15y.o. And 16-26 y.o. Their extreme
ages overlap and go much farther beyond the edges of our study’s age range.

Fonov et al. 2011 divides their templates into the following age ranges in years:
4.5 - 18.5 (the whole age range)
4.5 - 8.5 (prepuberty)
7 - 11 (pre to early puberty)
7.5 - 13.5 (pre to mid puberty)
10 - 14 (early to advanced puberty)
13 - 18.5 (post puberty)

Considering that our study spans 11 years across a child’s development, I divided the
study into smaller age-specific groups. This was to reduce the geometric distance to the
corresponding MNI pediatric atlas. Using a histogram, I created bin boundaries to keep the
19

number of subjects in each age group as equal as possible. I adjusted the boundaries to reflect
the ranges of the templates in bold above.
The histogram below indicates how the patient ages were balanced. Unfortunately, there
was only 1 control in the upper age group as can be seen below. Even if I had increased the
range of the upper group to include 14-year-old subjects, the third group would only gain one
extra control while shifting 4 patients from age group 2 to age group 3.

Figure 5. Top Left Histogram of Patient Ages, Top Right Histogram of Controls Ages, Bottom Histogram of All subject
ages

Using the histogram data in Figure 5., I decided on the age ranges below
● [6 ,10] using 7- 11 y.o. Pre to early puberty template (4-year range)
20

● (10 ,14] using 10 -14 Early to advanced puberty template (4-year range)
● (14,17] using 13 - 18.5 Post Puberty Template (3-year range)

The proposed scheme is to register each of the three templates to a fourth overall template
that exists geometrically in between all three age-specific templates. Then each of the subjects
will be registered to their group’s age-specific template. The transformations from subject to
age-specific template and from age-specific template to overall template will be combined and
applied to the original subject. Registration here is defined as applying FLIRT with mutual
information cost function with 12 pt DOF followed by FNIRT with default parameters.

To create a fourth overall template, I present several different options aligning across three
categories of groupwise registration summarized from Liu, Q. et al. 2014.

1. Group-mean
a. Doing pairwise registration and averaging the results to obtain the group mean Seghers
et al. 2004
i. The main drawback cited for this method is that the method treats all subjects
equally, which can create fuzzy images if the beginning subjects are not aligned.
b. Sharp-Mean - only subjects similar to group-mean image can participate in the updating
of the group mean. Guorong W. et al. 2011
i. Available on the UNC website called GLIRT (Groupwise and Longitudinal Image
Registration Toolbox)

Pairwise registration from the group-mean methods introduces bias to the template
when created with individual subjects. To escape this, methods were developed to register all
images while maintaining their relative distribution in geometric space.

2. Graph-theory based
21

a. ABSORB - Atlas Building by Self Organized Registration and Bundling- Jia H. et al.
2010.
i. Each subject image is constrained to deform locally with respect to its closet
geometric neighbor. The main benefit of this is maintaining the relative
distribution of subject images.
ii. Available on UNC website as ABSORB
b. HUGS - Hierarchical Unbiased Graph Shrinkage - Ying, S. et al 2014
i. In keeping with maintaining the data distribution during registration, HUGS builds
off ABSORB and introduces graph theory with each node representing an image
and each edge representing the geodesic pathway between two images.
Shrinking the nodes to be closer together while constraining them to maintain
topology maintains the distribution of images.
ii. Recently improved upon in Guorong, W. et al 2016
1. Doesn’t seem to provide significant improvement over HUGS in small
datasets (16 - 30)

Wang, Q. et al. 2010 suggested that “simplistic utilization of voxel-wise image intensity is
not sufficient to establish reliable correspondences, since it lacks important contextual
information.” And thus Wang, Q. propose to use attribute vectors instead of image intensity to
guide correspondence detection in groupwise registration.

3. Feature Based
a. Guorong W. et al. 2012 improved upon, implemented, and released this in their GLIRT
package at UNC linked above.

22

The main drawback to the group-mean method is introduced by applying the idea to
individual subjects. However, in creating this template, I am not using individual subjects, but
rather other templates which are spatial, and intensity normalized. The other methods will not
improve the creation of the fourth template significantly.
DWI Preprocessing
Visual Assessment

Figure 6. Example of Diffusion Weighted Image sagittal slice with interslice signal intensity drops
Motion during the scan can cause artifacts such as drops in interslice signal intensity, as
seen here in Figure 6. If I couldn’t reschedule the patient for a scan, whole volumes in the DW
collection would need to be thrown out before processing. This is an example of one such
volume which was removed. I would use ITK-SNAP to assess DW image quality and DSI Studio
to remove noisy volumes.
23

Eddy Current Correction and Subject Motion Correction
Eddy currents caused by the powerful diffusion gradients cause distortions in the image.
I used FSL’s eddy to register the diffusion weighted images to their b0 image both to address
eddy current artifacts as well as patient motion before skull stripping all volumes in the DWI.
Skull Stripping

Figure 7. Skull stripping of DWI in DSI Studio
I used the tool DSI Studio as shown in Figure 7 to accomplish brain surface extraction of
the DW image. The mask overlaying the axial view of the B0 image is seen as a red tint over
the image.

24

Registration to T1w & EPI Correction

Figure 8. DWI, Sagittal Slice before EPI correction (top left), after EPI correction (top right), with example of T1
registration target (bottom)
The first image acquired in a DWI sequence is the B0 image, which is usually an Echo
Planar Image (EPI) acquisition. EPI is sensitive to B0 field inhomogeneities which appear as
nonlinear distortions which is demonstrated in Figure 8 above.
25

In the example labeled “before,” there is some compression of the frontal and occipital
lobes due to the artifact. This is addressed by registering the DW image to the T1 image which
does not suffer similar distortions. This is all done with Brainsuite’s Brainsuite Diffusion Pipeline
(bdp) tool.
Diffusion Tensor Estimation

Figure 9. DTI Eigenvalues on brain slices in ITK-SNAP; Lambda 1 (top left), Lambda 2 (top right), Lambda 3 (bottom)
After the processing above to ensure that the diffusion weighted images are aligned, I
also process the estimation of the diffusion tensor. In Figure 9, examples of the three
26

eigenvalues, Lambda 1, 2 and 3 labeled as L1, L2 and L3 are shown. At this point it is easy to
calculate to DTI metrics- FA, MD, AD, and RD. In this example, I used Brainsuite’s BDP pipeline
to estimate the tensors. However, FSL’s dtifit is also applicable if we’re planning to use Diffusion
Tensor Imaging ToolKit (DTI-TK), for example, in tract specific analysis.
Apply Combined Registration to DTI
Finally, using the deformation information from registering the T1 to the template space,
I applied that same deformation map to the DTI metrics to obtain FA, MD, AD, and RD maps in
the template space using FSL’s applywarp.

27

Chapter 2. Brain Biomarkers and Neuropsychological Outcomes of
Pediatric Posterior Fossa Brain Tumor Survivors Treated with Surgical
Resection with or without Adjuvant Chemotherapy

Mary C. Nelson
1,2
; Sharon H. O’Neil
2,3,4
;Jeffrey Tanedo
2,5
;Sofia Dhanani
3,6
; Jemily Malvar
7
;
Christopher Nuñez
8
; Marvin D. Nelson Jr.
9,10
; Benita Tamrazi
9,10
; Jonathan L. Finlay
11,12
; Vidya
Rajagopalan
2,10
; Natasha Lepore
2,5,10

1
Departments of Medical Education and Pediatrics,
Keck School of Medicine of USC, Los Angeles, California

2
Radiology Department, CIBORG Laboratory,
Children’s Hospital Los Angeles, Los Angeles, California

3
The Saban Research Institute, Children’s Hospital Los Angeles, Los Angeles, California

4
Division of Neurology, Children’s Hospital Los Angeles, Los Angeles, California

5
USC Viterbi School of Engineering, Los Angeles, California

6
Keck School of Medicine of USC, Los Angeles, California

7
Division of Hematology, Oncology and Blood and Marrow Transplantation,
Children’s Hospital Los Angeles, Los Angeles, California

8
Neuropsychologist in private practice, Los Angeles, California

9
Department of Radiology, Keck School of Medicine of USC, Los Angeles, California

10
Department of Radiology, Children’s Hospital Los Angeles, Los Angeles, California

11
The Ohio State University College of Medicine, Columbus, Ohio

12
Nationwide Children’s Hospital, Columbus, Ohio

28

Abstract
Purpose: Children with brain tumors experience cognitive late effects, often related to cranial
radiation therapy (CRT). We sought to determine differential effects of surgery and
chemotherapy on brain structure and neuropsychological outcomes in children who did not
receive CRT.
Methods: Twenty-eight children with a history of posterior fossa tumor (17 treated with surgery,
11 treated with surgery and chemotherapy) underwent neuroimaging and neuropsychological
assessment a mean of 4.5 years to 9 years (respectively) posttreatment, along with 18 healthy
sibling controls. Psychometric measures assessed IQ, language, executive functions,
processing speed, memory, and social-emotional functioning. Group differences and
correlations between diffusion tensor imaging findings and psychometric scores were examined.
Results: The z-score mapping demonstrated fractional anisotropy (FA) values were ≥2
standard deviations lower in white matter tracts, prefrontal cortex gray matter, hippocampus,
thalamus, basal ganglia, and pons between patient groups, indicating microstructural damage
associated with chemotherapy. Patients scored lower than controls on visuoconstructional
reasoning and memory (P ≤ .02). Lower FA in the uncinate fasciculus (R =−0.82 to −0.91) and
higher FA in the thalamus (R = 0.73-0.91) associated with higher IQ scores, and higher FA in
the thalamus associated with higher scores on spatial working memory (R = 0.82).
Conclusions: Posterior fossa brain tumor treatment with surgery and chemotherapy affects
brain microstructure and neuropsychological functioning years into survivorship, with spatial
processes the most vulnerable. Biomarkers indicating cellular changes in the thalamus,
hippocampus, pons, prefrontal cortex, and white matter tracts associate with lower psychometric
scores.
29

1. Introduction
Central nervous system (CNS) tumors are the most common cause of cancer death in
children of age 0-14 years in the United States [Ostrom, QT., et al, 2015]. Treatment regimens
of surgery, chemotherapy, and irradiation are based upon risk including tumor molecular
genetics, location, pathology, and age at diagnosis [Pollack, IF., et al., 2011]. Cranial radiation
therapy (CRT) is a factor in neuropsychological deficits in pediatric cancer survivors ranging
from a global loss of IQ points to memory and attention deficits [Kim, JH., et al., 2008; Turner,
CD., et al., 2009]. Evidence from large cohorts, such as the Childhood Cancer Survivor Study
[Anderson, FS., et al., 2009] and Children’s Oncology Group, supports decreasing CRT dose or
elimination when feasible [Packer RJ., et al., 1999]. Children under 6 years may be treated with
irradiation-sparing intensive chemotherapy regimens when possible without compromising
survival [Dunkel IJ., et al., 2002; Marachiellian A., et al., 2008; Grill, J., et al., 2005; Rutkowski
S., et al., 2005], as cognitive and social-emotional effects appear mitigated with this approach
[Sands S., et al., 1998; Sands SA., et al., 2010; Sands, SA., et al., 2011; Fay-McClymont TB., et
al., 2017; O’Neil SH., et al., 2020]. Late effects of systemic and CNS-directed chemotherapy on
the developing brain in the absence of CRT are less well defined, but in children with leukemia,
include deficits in attention, executive functions, visual processing, and visual motor deficits
[Anderson FS., et al., 2009; Beaulieu C., et al., 2013; Kesler SR., et al., 2010; Reddick WE., et
al., 2006; Robinson EK., et al., 2009].
Treatment induces neurotoxicity through oxidative stress, and inflammation triggered by
irradiation and/or chemotherapy [Kim, JH., et al., 2008; Myers JS; et al., 2008]. Cell membrane
breakdown and death, demyelination, and loss of blood-brain barrier integrity contribute to
edema and a cycle of continuing injury [Kim JH., et al., 2008]. Magnetic resonance imaging
(MRI) illustrates extent of damage and injury subtypes in vivo. Diffusion tensor imaging (DTI),
30

characterizing three-dimensional water diffusion as a function of spatial location, provides
indices of mean diffusivity (MD) and fractional anisotropy (FA). MD reflects cell size, shape,
integrity, and molecular motion across tissues [Cercignani M., et al., 2001; Kumar R., et al.,
2006]. High MD may indicate edema or loss of axons and demyelination [Kumar R., et al., 2006;
Iannucci G., et al., 2001]. Comparatively low FA represents loss of microstructural integrity,
particularly in myelinated regions, reduced tissue organization [Cercignani M., et al., 2001;
Mukherjee P., et al., 2002], and axonal damage [Mac Donald CL., et al., 2007].
Rueckriegel et al (2010) demonstrated that children with low-grade brain tumors treated
with surgery showed decreased FA in the white matter (WM) tract skeleton, though to a lesser
extent than those treated with surgery, irradiation, and chemotherapy [Rueckriegel SM., et al.,
2010]. Similar populations treated with surgery and chemotherapy showed deficits in processing
speed, visual sustained attention [Aarsen FK., et al., 2009], and visual working memory
[Peterson RK., et al., 2009], indicating that brain tumor presence, surgical resection, and
chemotherapy results in WM disruption. However, few such neuroimaging studies exist, and
neuropsychological outcomes remain unclear.
In a pilot study, we investigated brain injury in childhood brain tumor survivors treated
with surgery and chemotherapy (n = 7) compared to healthy controls (n = 9) using DTI [Baron
Nelson M., et al., 2014; Baron Nelson M., et al., 2016]. Higher MD values indicated significant
changes in the thalamus, pons, basal ganglia, and mammillary bodies in patients compared to
controls [Baron Nelson M., et al., 2016]. Due to sample size, it could not be determined if this
pattern of injury was due to tumor, surgery, chemotherapy, or a combination of the three. The
purpose of the current study was to differentiate effects of surgery alone, and surgery followed
by chemotherapy on brain microstructure and neuropsychological function in pediatric posterior
fossa tumor survivors. Our a priori hypotheses were: (a) neuroimaging would indicate a pattern
of injury to subcortical and brainstem structures in children treated with surgery and
31

chemotherapy (S + C) compared to children treated with surgery (S) and healthy controls (HC),
(b) children in the S + C group would score lower than those in the S and HC groups on
neuropsychological outcomes, and (c) injury to pons, hippocampus, basal ganglia, and
thalamus would correlate with poorer neuropsychological outcomes.

32

2. Methods
A cross-sectional comparative design was used with participants completing MRI with
DTI and neuropsychological assessment. Patient groups included children at least 12 months
from last treatment for a posterior fossa brain tumor, currently 6-17 years old, treated with either
surgery (S group) or surgery and chemotherapy (S + C group). The study was approved by the
Institutional Review Board. Potential subjects were identified from the neuro-oncology database
and clinic lists. Parents were approached in clinic or contacted by mail or phone. At the time of
patient enrollment, parents were invited to enroll healthy 6-17-year-old siblings as controls.
Parents and participants were required to speak and read either English or Spanish.
Exclusion criteria for all participants included metal in the body precluding MRI, preterm
birth, neurodevelopmental disability, and traumatic brain injury. Controls had to be able to
undergo MRI without sedation. Patients with recurrent tumor or residual disease outside the
posterior fossa were excluded, as were those with a history of posterior fossa syndrome, since
related deficits, including lower IQ, working memory, and processing speed [Schreiber JE., et al,
2017] are significant confounding variables.
All data were stored in REDCap v6.14.2.32,33

33

Figure 10. Image Processing details of single subject and group-level image processing. Abbr.: AFNI, analysis of
functional neuroimages; ANTS, advanced neuroimaging tools; DWI, diffusion-weighted imaging; FSL, FMRIB
software library; NIHPD, NIH pediatric data
2.1 Imaging data and preprocessing
Three-dimensional T1-weighted images were obtained on a 3.0 T Philips Achieva
scanner with voxel size of 1.0 × 1.0 × 1.0 mm3 with parameters: TR 9.9 ms; TE 4.6 ms; 240 ×
231 matrix; FOV 24 cm. DWI imaging sequence was acquired with parameters: 70 axial slices
(2-mm thick), FOV = 256 × 256 × 140 mm, TR/TE 8657/86 ms, no gap, with a 128 × 126
acquisition matrix, 28 gradient directions collected with b-value = 1500.
T1 images were bias field corrected using ANTs [Mac Donald CL., et al., 2007], N4
BFC26 tool, and manually skull-stripped in Brainsuite16 [Aarsen FK., et al., 2009]. Twelve-point
linear registration was done using FSL’s [Peterson RK., et al., 2019; Baron Nelson M., et al.,
2014; Baron Nelson M., et al., 2016] FLIRT [Schreiber JE., et al., 2017; Harris PA., et al., 2009]
tool and nonlinear registration to the McConnell Brain Imaging Center’s (MBIC) natural pediatric
34

template for children [Harris PA., et al., 2019; Cox RW., et al., 1996] with FSL’s FNIRT tool. The
overall template space is an average of three age appropriate MBIC templates. DW images
were corrected with FSL19 for eddy current and subject motion. Resulting DWIs were skull-
stripped in DSI Studio. Brainsuite’s BDP Pipeline co-registered DWI images to preprocessed
T1w images, then registered to the MBIC mean template space by applying combined
transformations from the T1w image’s registration pipeline (Figure 10).
2.2 Neuropsychological assessment
Participants completed a battery of well-validated measures (i.e., Wechsler scales,
NEPSY II Memory for Designs, California Verbal Learning Test, Children’s Version,
Receptive/Expressive One-Word Picture Vocabulary Tests, NIH Toolbox, and Achenbach Child
Behavior Checklist) commonly used to assess domains potentially affected by posterior fossa
brain tumors, including intelligence, processing speed, memory, executive functions, language,
and psychosocial functioning. Assessments were performed by a board-certified pediatric
neuropsychologist or by doctoral trainees under her supervision. For full details of these
assessments please see Baron Nelson, Mary C., et al 2021.
2.3 Statistical methods
Whole-brain voxel-wise analysis was performed in analysis of functional neuroimages
(AFNI)34 using 3dttest++ to generate pairwise z-score maps. We masked the cerebellum from
image analysis due to our focus on structures outside the surgical resection. Age at time of scan
was a regression covariate. False discovery rate (FDR) was applied at P ≤ .05 to voxel-wise
analyses. Within AFNI’s graphical user interface output, z-maps were thresholded to show only
voxels where FA had z scores ≥ 2 SD from mean. Cluster maps were exported to ITK-SNAP
35

[Yushkevich PA., et al., 2006] where mean and SD of FA values in each cluster were extracted.
Anatomical labels are in concordance with MRI atlas [Oishi K., et al., 2011].
2.3.1 Post hoc analyses
Using AFNI’s 3dttest++, whole-brain voxel t-statistic maps were created for each DTI
metric using group-level analysis in three pairwise analyses between S, S + C, and HC groups.
AFNI’s 3dcalc was then used to convert each t-statistic map into a z-score map. The z-maps
were filtered to display regions of the brain where FA was |z| > 2, indicating areas with ≥2 SD
difference between treatment groups and then between each treatment group and HCs, using a
Q-value of 0.05 (hypothesis 1). WM clusters with 100 or more surviving contiguous voxels are
reported. In the thalamus, pons, and basal ganglia, we decreased the number of contiguous
voxels required to 30 or more to allow analysis of areas specified in our hypotheses, since those
structures are generally smaller in size than WM regions and allowed further analysis of FA
findings in regions identified in a priori hypothesis (thalamus, basal ganglia, hippocampus).
The χ2 test and t-test examined differences in patient demographics and medical history among
groups, and neuropsychological scores were compared using linear regression. ANOVA
compared scores in patients to HCs with age at diagnosis as a covariate (hypothesis 2).
Analyses were performed in Stata [Statistical Software., et al., 2015]] using two-sided tests.
FDR was controlled by applying the Benjamini Hochberg procedure [Benjamini Y., et al., 1995]
(Q-value 0.012) with significance set at P ≤ .02.
Pearson correlation analysis using SPSS v.25.0 [IBM SPSS Statistics for Windows., et
al., 2017] assessed the relationship of FA in the eight clusters that significantly differed between
groups to neuropsychological scores, controlling for age and time off treatment in patient groups
(hypothesis 3).

36

3 Results
3.1 Demographics
One hundred thirty-one children were potentially eligible. After excluding those treated
with CRT, turning 18 during recruitment, with PFS or severe developmental delay, or no longer
followed at the institution, there were 48 patients. Of those, five families declined participation
and 15 families did not respond to mail/telephone invitations, for a participation rate of 58%. The
20 who did not participate compared to the 28 who did were not significantly different in current
age, age at diagnosis or years since treatment, but insurance status suggested a larger
percentage were of higher socioeconomic status than in the final sample. However, the sample
was representative of that the institution serves.
Seventeen children in the S group, 11 in the S + C group, and 18 HCs participated in the
study. Demographic data are presented in Table 1. Although age at study was not different
between groups, age at diagnosis was, thus the distribution of current ages between patient
groups varied with the S group skewed toward younger ages, and the S + C group skewed
toward older ages and longer time from treatment. There was also a difference between
diagnoses, as most patients in the S group had pilocytic astrocytoma and most in the S + C
group had medulloblastoma (P = .0001), which was expected related to assigned treatment. For
all 46 participants, there were no group differences in preschool attendance (P = .62), receiving
special education (P = .43), family history of developmental delays, learning disabilities,
psychiatric, or medical illnesses (P = .06, .31, .18, .08, respectively). The S + C group was
treated with carboplatin, etoposide, and thiotepa (with 82% receiving each of these drugs),
vincristine and cyclophosphamide (73%), cisplatin (64%), methotrexate (36%), temozolomide
(27%), and dasatinib, lenalidomide, or irinotecan (9%).
37

3.2 Neuroimaging
There were no significant group differences in regional DTI after correction for multiple
comparisons with the false discovery rate (FDR) procedure, permutation analysis, and cluster
analysis. Despite comparable mean ages of subjects in each group, age-related effects
dominated the analysis. Therefore, we partialed out participant age as a regression covariate.
Areas of significantly different FA clusters by z masks between patient groups (SvsS +
C) are shown in Figure 11. FA was higher in the S group in several regions, including the
superior longitudinal fasciculus (SLF) and uncinate fasciculi (UF), posterior thalamic radiation
(PTR)/splenium of corpus callosum (CC), and posterior limb of internal capsule (PLIC), as well
as in the right pons, gray and WM of the prefrontal cortex (PFC), bilateral hippocampi, right
globus pallidus and putamen, and right thalamus. FA was lower in the S group in the left
thalamus.
Areas of significantly different FA clusters between each patient group and HCs are
shown in Figures S1 and S2 supporting findings between patient groups, demonstrating higher
FA in the pons, hippocampus, and thalamus in controls than in S patients (Figure S1). FA in
controls was higher than S + C patients in the hippocampus, putamen, PLIC, and inferior
frontooccipital fasciculus (Figure S2). FA was lower in controls in the thalamus and SLF than in
both patient groups. All supporting figures and tables are only available online.
3.3 Neuropsychological assessment
No monolingual Spanish-speaking children participated. Children completed the
Children’s Oncology Group Language Preference Survey. The single bilingual participant who
38

was not English dominant for testing was assessed by a bilingual examiner providing
instructions in both English and Spanish and completed bilingual versions of Receptive and
Expressive One-Word Picture Vocabulary tests.
There were no significant differences between patient groups in neuropsychological
performance. When factoring in HCs, there was no difference in FSIQ (96.50, 96.80, 102.81),
though HCs performed higher on Block Design, a subtest within FSIQ (P = .01) (Table S1).
Controls also performed higher than patients on immediate design recall and delayed recall of
designs and their spatial locations on Memory for Designs (P = .02, .005, respectively).
Distribution of scores is shown in Figure S3. Significant group differences on other tasks with a
motor component (i.e., spatial span, coding, symbol search, dimensional change card sort,
flanker inhibitory control, and pattern comparison processing speed) were not found. Children in
the S group and HCs scored lower (P = .02) than those in the S + C group on self-reported
NeuroQOL Pediatric Cognitive function, though both groups placed within the norm. Scores on
depression, fatigue, and interacting with peers were not different between groups; however,
parent questionnaires indicated clinical concerns regarding executive functions and social-
emotional functioning more frequently for patients (Table S2). The primary difference between
groups was in the Externalizing Problems Index of the CBCL, where 36% of children in S, 20%
of children in S + C, and no controls had parent-proxy scores in the clinical range.

39

Table 1. Patient demographics by treatment group compared to controls
Variable Treatment
Group
S (n=17) S+C (n=11) HC (n=18) Total (n=46) P-value
Age – mean (SD) 10.76(4.02) 12.18(3.52) 10.56(2.23) 11.02(3.25) .41
Age at diagnosis –
mean (SD)
6.02(3.57) 3.52(3.29) NA 5.04(3.63) .07
Sex Male 4 (24%) 6 (55%) 10(56%) 20(43%) .11
Female 13 (76%) 5 (45%) 8 (38%) 26 (53%)
Handedness Right 14 (82%) 10 (91%) 17 (94%) 41 (89%) .78
Left 3 (18%) 1 (9%) 1 (5%) 5 (11%)
Patient
race/ethnicity
White 15 (88%) 10 (91%) 16 (89%) 41 (89%) Race:.
84
Ethnicity
Non-Hispanic 7 (41%) 3 (27%) 10 (63%) 20 (43%)
Hispanic/Latino 9 (53%) 8 (73%) 6 (33%) 23 (50%)
Black/African
American
1 (6%) 0 (0%) 0 (0%) 1 (2%)
Asian 1 (6%) 1 (9%) 2 (11%) 4 (9%)
Diagnosis Medulloblastoma 0 (0%) 7 (64%) NA 7 (25%) .0001
Ependymoma 0 (0%) 1 (9%) 1 (4%)
Astrocytoma 2 (12%) 0 (0%) 2 (7%)
Pilocytic
Astrocytoma
14 (82%) 2 (18%) 16 (57%)
Other 1 (6%) 1 (9%) 2 (7%)
Time off treatment,
years – mean (SD)
4.47 (3.20) 8.09 (4.71) NA 5.89 (4.19) .02
Region of tumor
location
Midline / vermis 6 (35%) 7 (64%) NA 13 (46%) .18
Left cerebellar 4 (24%) 3 (27%) 7 (25%)
Right cerebellar 7 (41%) 1 (9%) 8(29%)
Hydrocephalus at
diagnosis
Yes 10 (59%) 7 (64%) NA 17 (61%) .75
No 5 (29%) 2 (18%) 7 (25%)
Unknown 2 (12%) 2 (18%) 4 (14%)
Hydrocephalus
Severity
Mild 3 (30%) 2 (29%) NA 5(29%) .87
Moderate 6 (60%) 3 (43%) 9 (53%)
Severe 0 (0%) 1 (14%) 2 (12%)
VP Shunt Yes 1 (6%) 6 (55%) NA 7 (25%) .002
No 16 (94%) 5 (45%) 21 (75%)
Bilingual No 10 (59%) 4 (36%) 10 (56%) 24 (52%) .82
Yes 5 (29%) 5 (45%) 6 (33%) 16 (35%)
Missing 2 (12%) 2 (18%) 2 (11%) 4 (9%)
Annual Income ≤$ 39 999 6 (35%) 4 (36%) 7 (39%) 17 (37%) .80
$40 000–79 999 1 (6%) 1 (9%) 2 (11%) 4 (9%)
≥$80 000 7 (41%) 3 (27%) 8 (44%) 18 (39%)
Missing 3 (18%) 3 (27%) 1 (6%) 7 (15%)
Did mother attend
college?
No 7 (41%) 5 (45%) 8 (44%) 20 (43%) .12
Yes – BS 1 (6%) 4 (36%) 2 (11%) 7 (15%)
Yes – Grad 6 (35%) 1 (9%) 8 (44%) 15 (33%)
Missing 3 (18%) 1 (9%) 0 (0%) 4 (9%)
Did father attend
college?
No 6 (35%) 5 (45%) 3 (17%) 14 (30%) .11
Yes - BS 4 (24%) 4 (36%) 9 (50%) 17 (37%)
Yes - Grad 3 (18%) 1 (9%) 6 (33%) 10 (22%)
Missing 4 (24%) 1 (9%) 0 (0%) 5 (11%)

40

3.4 Neuroimaging and neuropsychological assessment
Pearson correlations determined relationships between eight clusters of FA significance
in Figure 11 and psychometric scores in children with brain tumors, controlling for age and time
off treatment (Table 2). Higher FA in the right thalamus correlated with higher scores on spatial
working memory, perceptual reasoning, block design, FSIQ, and similarities in the S + C group.
Higher FA in the right pons correlated with lower scores on memory in the S + C group. Children
in the S + C group demonstrated inverse correlations between FA in the UF and FSIQ,
vocabulary, and similarities. In the S group, higher FA in the left hippocampus and right PLIC
correlated with improved scores in spatial working memory, while lower FA in prefrontal gray
matter (GM) correlated with higher scores in spatial working memory.
The sole significant correlation in HCs was between spatial working memory and left
thalamus FA (R = 0.63, P = .01).
Data are available on request from the authors.
41

42

Figure 11. Areas of fractional anisotropy (FA) differences between patient groups in composite axial and sagittal
diffusion tensor images.
Red clusters indicate areas where FA z scores were ≥2 SD lower in the S + C group than the S group; blue clusters
indicate where FA z scores were ≥2 SD lower in the S group than the S + C group. R and L indicate right and left on
axial images; A and P indicate anterior and posterior. Cluster size is noted for each structure. A, FA in bilateral SLF
higher in S than S + C. B, FA in prefrontal WM lower in S than S + C; FA in prefrontal GM higher in S than S + C. C,
FA in R thalamus higher in S than S + C; FA in L thalamus lower in S than S + C; FA in L PTR/CC higher in S than S
+ C. D, FA in R putamen and GP higher in S than S + C. E, FA in R UF higher in S than S + C. F, FA higher in R
hippocampus in S than S + C. G, FA higher in L hippocampus in S than S + C. H, FA higher in R pons in S than S +
C. I, FA higher in R PLIC and prefrontal WM in S than S + C. Abbreviations: CC, corpus callosum; GM, gray matter;
GP, globus pallidus, MFG, middle frontal gyrus; PLIC, posterior limb of internal capsule; PTR, posterior thalamic
radiation; SLF, superior longitudinal fasciculus; UF, uncinate fasciculus; WM, white matter

Table 2. Neuropsychological assessment scores and fractional anisotropy R values *
Domain/test L MFG GM R UF R PLIC R thalamus L
Hippocampus
R Pons
General Intelligence
Verbal comprehension
-0.84 S + C
Perceptual Reasoning 0.83 S + C
Block Design 0.91 S + C
FSIQ -0.91 S + C 0.88 S + C
Vocabulary -0.82 S + C
Similarities -0.84 S + C
Working memory
Digit span total
0.98 S
Spatial span total 0.82 S + C
Spatial span forward -0.98 S 0.98 S
Memory
Memory for designs
delayed total
-0.87 S + C
Language
Expressive one-word
vocabular test
-0.83 S + C
Abbreviations: GM, gray matter; MFG, middle frontal gyrus; PLIC, posterior limb of internal capsule; UF, uncinate
fasciculus.
*P<0.01

43

4 DISCUSSION
This study demonstrated FA differences indicating microstructural injury in brain areas,
many distant from the tumor site, in children treated with surgery and chemotherapy compared
to those treated with surgery only, accompanied by minor differences in cognitive performance.
FA differences were found in WM association pathways connecting temporal lobes with
orbitofrontal and prefrontal cortices (UF) [Von Der Heide RJ., et al., 2013; Schmahmann D., et
al., 2007], and those connecting parietal lobes to prefrontal cortices (SLF) [Schmanmann JD., et
al., 2007]; in the PLIC, PTR/CC; pons and hippocampi; basal ganglia and thalami; and the PFC.
These findings both confirm and expand upon results of our pilot study, where MD differences
indicated injury to subcortical GM and pons in children treated with S + C for brain tumors
[Baron Nelson M., et al., 2014; Baron Nelson M., et al., 2016], also in the absence of major
cognitive deficits [Baron Nelson M., et al., 2016].
Lower FA in certain brain structures of children in the S + C group may indicate
increased vulnerability of healthy tissue to chemotherapy apart from tumor or surgery effects.
WM in the PFC and association pathways continues to develop past adolescence [Sowell ER,.
et al., 2004] and structures such as the UF develop later in life [Riggs L., et al., 2014], possibly
rendering these tracts more vulnerable to neurotoxicity when injured earlier in development.
Children in the S + C group were younger at diagnosis than children in the S group.
Lower mean FA in WM compared to HCs is common to many neuropathological
conditions, including multiple sclerosis [Bammer R., et al., 2000], stroke [Werring DJ., et al.,
2000], epilepsy [Dumas de la Roque A., et al., 2005], Alzheimer’s disease [Fellgiebel A., et al.,
2004], and brain tumors including meningiomas, low-grade gliomas, and glioblastoma
multiforme [Lu S., at al., 2004], indicating demyelination, edema, and/or inflammation [Assaf Y.,
et al., 2008]. It is a marker of WM damage, but nonspecific as to exact underlying pathology.
44

Other pediatric brain tumor studies found decreased mean FA in WM areas identified in
our study, including the UF [Riggs L., et al., 2014; Aleksonias HA., et al., 2019], internal
capsule, PTR/splenium of CC, SLF [Palmer SL., et al., 2012], and frontal WM [Mabbott DJ., et
al., 2006]. However, unlike this study, most did not control for tumor location and included
children treated heterogeneously with combinations of surgery, CRT, and/or chemotherapy.
The additional neuronal damage in patients who received chemotherapy was expected,
as all patients in the S + C group received chemotherapeutic agents known to cross the BBB
(methotrexate, cisplatin, thiotepa, and temozolomide) [Myers JS., et al., 2008] and cause
neurotoxicity [Ikonomidou C., et al., 2018]. While direct effects of chemotherapy in the
developing brain are not fully understood, hypothesized mechanisms include neurotoxic injury to
cerebral parenchyma, induction of a secondary inflammatory response, microvascular injury,
indirect chemical toxicity, increased oxidative stress, altered neurotransmitter levels [Myers JS.,
et al., 2008], DNA damage, decreased neurogenesis, and shortening of telomeres [Ikonomidou
C., et al., 2018]. These disruptions interfere with normal myelination, synaptogenesis, and
pruning, all of which could contribute to our findings.
The S + C group demonstrated inverse correlations between FA in the UF and general
intelligence. One would not expect that lower FA in WM, indicative of myelin disruption, would
correlate with higher intelligence scores. One explanation is that FA in the UF is lower than
expected where the UF crosses the SLF [Schmahmann JK., et al., 2006], as crossing fibers
cause the diffusion tensor to become more spherical or oblate [Tournier J., et al., 2010],
reducing FA even in the presence of intact myelinated fiber tracts. Our methods did not allow
determination of whether the cluster of significance was within these crossing fibers.
While FA is most commonly measured in WM, we found decreased FA in the right
thalamus and increased FA in the left thalamus in the S + C group compared to the S group.
45

This contrary finding may be due to location of the FA clusters in different thalamic nuclei (right
ventrolateral nucleus and left medial thalamic nucleus), although imaging resolution did not
allow us to confirm location in each subject. The medial thalamic nucleus is located much closer
to the lateral ventricle than the ventrolateral nucleus, and as such may have been injured from
hydrocephalus, which was more prevalent in the S + C group. High FA in GM may be a
biomarker of neuronal injury in patients with chemotherapy-induced toxicity. The organization of
GM, primarily made up of cell bodies, unmyelinated axons, dendrites, and synapses, is less
directionally organized than WM, making interpretation of DTI findings less clear [Baron Nelson
M., et al., 2019] Increased FA in GM was associated with gliosis in an animal model of traumatic
brain injury [Budde MD., et al., 2011] Similar patterns of increased FA in GM are noted in
chronic subdural hematoma [Osuka., et al., 2012], where a compression dependent increase in
FA in the caudate and putamen decreased following evacuation of hematomas, suggesting
increased FA in GM could indicate damage caused by tissue compression. These studies
provide evidence that increased FA in the left thalamus could be a result of gliosis, damage to
surrounding WM, and/or compression of tissues from hydrocephalus.
Decreased FA in the basal ganglia, thalamus, and pons supports our earlier findings of
elevated MD in these areas in children with brain tumors treated with surgery and chemotherapy
[Baron Nelson M., et al., 2016], as both directional indices represent microstructural injury. In
five cognitive domains assessed (IQ, language, executive functions, processing speed, and
memory), group differences were visual-spatial, suggesting greater sensitivity of these
measures to broad CNS dysfunction [Lezak MD., et al., 2004], commonly reported in neurologic
conditions such as congenital hydrocephalus [Bigler ED., et al., 1988].
Block design, a visuoconstructional task requiring one to analyze part-whole
relationships to copy two-dimensional patterns, is sensitive to CNS dysfunction [Lezak MD., et
al.]. Patient groups scored significantly lower than the HC group, though all performed within
46

normal limits. This encouraging finding suggests that although children in the S + C group were
younger at diagnosis, often a risk factor for poorer outcomes, and had intensive treatment, there
were more years from diagnosis in which to, at least theoretically, develop neural plasticity. On
a spatial memory task, patients performed lower than controls on learning designs and on
delayed recall of both designs and locations, with those who received chemotherapy showing
an additional, though nonsignificant, reduction in performance. This suggests patients had
difficulty encoding, a process involving the hippocampus and WM connections between frontal
and parietal regions via the thalamus. It could also reflect breakdown in connections involved in
consolidating memory into long-term storage, a process thought to be mediated by medial
temporal and diencephalic structures, as well as specific unimodal and heteromodal cortices or
in retrieving it through activity in regions of neocortex without need for medial temporal or medial
diencephalic involvement [Blumenfeld H., et al., 2002].
Finally, it is notable that the S + C group performed like the S group on
neuropsychological assessment, but scored higher on QOL, even compared to controls. This
further supports the theory that these children are relatively functionally intact due to
neuroplasticity related to treatment at a younger age.

4.1 Limitations
As is common in pediatric brain tumor research, our sample size is relatively small,
making definitive conclusions difficult, and participants had a range of cognitive abilities from
impaired to superior, making aggregate data less representative. While only diagnosis and age
at diagnosis differed between patient groups, patients were not specifically matched by sex,
age, language, handedness, or tumor location/laterality.
47

5 CONSIDERATIONS/CONCLUSIONS
Few prior neuroimaging studies included children with brain tumors treated with
chemotherapy without CRT. We found clear patterns of brain injury in children with posterior
fossa tumors treated with surgery with or without chemotherapy.
While most patients performed within normal limits on neuropsychological assessment,
they performed lower than healthy controls on visual-constructional reasoning and spatial
memory. Earlier neurological injury in children in the S + C group could allow for
reorganization/compensation with other parts of the attention network with frontal or parietal
aspects of attention playing a more important role in performance. Although not tested in this
study, this is an important focus for future studies.
Factors underlying vulnerability or resilience in the face of neuro-toxic therapies merit
further evaluation to determine effects of genetics and environment. Future studies involving
multiple institutions enabling larger sample sizes over multiple time points are necessary to
pinpoint periods of neuronal vulnerability amenable to intensive interventions.
Our findings illustrate that treatment of pediatric posterior fossa brain tumors results in
long-term alterations to gray and WM microstructure, and that more pronounced differences are
seen when chemotherapy is used in addition to surgical intervention.

48

ACKNOWLEDGMENTS
This study was funded by K23NR014902 and Children’s Hospital Los Angeles Hemonc/BMT
Division Research Funds and supported by The Saban Research Institute (TSRI) of Children’s
Hospital Los Angeles, and USC CTSI REDCap UL1TR001855. This study was also funded in
part by the SC CTSI, which is part of the Clinical and Translational Science Awards (CTSA), a
national network funded through the National Center for Advancing Translational Sciences
(NCATS) at the NIH (Grant Number UL1TR000130). Under the mandate of “Translating Science
into Solutions for Better Health,” SC CTSI provides a wide range of ser-vices, funding, and
education for researchers, and promotes online collaboration tools such as USC Health
Sciences Profiles. Natasha Lepore is funded by R01EB025031 and TSRI 000013228. Vidya
Rajagopalan is funded by K01 HL153942 and TSRI 00011096. The content is solely the
responsibility of the authors and does not necessarily represent the official views of the National
Institutes of Health. The authors would like to thank Drs Ki Moore and Bradley Peterson, who
were mentors for this study; and Christine Obioha and Francesca Trane, who assisted with data
entry.

49

Chapter 3. White Matter Tract Changes in Pediatric Posterior Fossa
Brain Tumor Survivors after Surgery and Chemotherapy

Jeffrey Tanedo
1,2
, Niharika Gajawelli
1
, Sharon Guo
1
, Mary Baron Nelson
1,3,†
, Natasha Lepore
1, 2,3,†

1
CIBORG Laboratory, Children’s Hospital Los Angeles,
Department of Radiology, Los Angeles, CA, USA

2
Department of Biomedical Engineering,
University of Southern California, Los Angeles, CA, USA

3
Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
†
Denotes equal senior authorship

50

Abstract
Survivors of pediatric posterior fossa brain tumors are susceptible to the adverse effects
of treatment as they grow into adulthood. While the exact neurobiological mechanisms of these
outcomes are not yet understood, the effects of treatment on white matter (WM) tracts in the
brain can be visualized using diffusion tensor (DT) imaging. We investigated these WM
microstructural differences using the statistical method tract-specific analysis (TSA). We applied
TSA to the DT images of 25 children with a history of posterior fossa tumor (15 treated with
surgery, 10 treated with surgery and chemotherapy) along with 21 healthy controls. Between
these 3 groups, we examined differences in the most used DTI metric, fractional anisotropy
(FA), in 11 major brain WM tracts. Lower FA was found in the splenium of the corpus callosum
(CC), the bilateral corticospinal tract (CST), the right inferior frontal occipital fasciculus (IFOF)
and the left uncinate fasciculus (UF) in children with brain tumors as compared to healthy
controls. Lower FA, an indicator of microstructural damage to WM, was observed in 4 of the 11
WM tracts examined in both groups of children with a history of posterior fossa tumor, with an
additional tract unique to children who received surgery and chemotherapy (UF). Our findings
indicate that a history of tumor in the posterior fossa and surgical resection may have effects on
the WM in other parts of the brain.

51

Introduction
Central nervous system (CNS) tumors are second in frequency only to leukemia among
cancers affecting children but are still the most common cause of cancer death (Udaka et al.,
2018) in children ages 0-14 years in the United States, with an incidence rate of approximately
5.83 per 100,000 person-years (Ostrom et al., 2017). Of these tumors in children, over half are
in the posterior fossa, making it the most common location for CNS tumors. Treatment for these
tumors can include a combination of surgical resection, chemotherapy, and cranial or
craniospinal irradiation. New developments in these therapies, earlier detection, and improved
post-treatment monitoring have increased the survival rates for pediatric patients. However,
children are most susceptible to the adverse effects of treatment during this period of significant
brain development (Macartney et al., 2014). Thus, the urgency to examine the long-term
adverse outcomes of these treatments has also increased.
The exact neurobiological mechanisms leading to adverse outcomes from these
therapies are not yet understood. However, the effects on brain anatomy may be visualized,
quantified, and analyzed through magnetic resonance imaging (MRI) (Kim et al., 2008;
Ikonomidou 2018; Jacob et al., 2018). For example, global reductions in both grey and white
matter (WM) volumes have been seen and correlated with neurocognitive decline after
treatment (Ailion et al., 2017).
WM is important in mediating the functional connectivity for many neurobehavioral
operations (Filley et al., 2016) and is susceptible to damage from radiation and chemotherapy.
Thus, particular attention has been given to WM tract alterations with diffusion-weighted imaging
(DWI). DWI is an MRI method which captures the diffusion of water molecules through brain
tissues and is particularly useful in analyzing WM tracts. Healthy WM typically consists of
bundles of myelinated axons organized into tracts which connect different parts of the brain. The
restricted flow of water molecules through these bundles can be characterized through different
52

metrics, the most popular of which is called fractional anisotropy (FA). FA describes the degree
of deviation from purely isotropic Brownian motion of water molecules in the brain, where higher
FA values indicate highly anisotropic random motion (Basser et al., 1994).
Lower FA values in a particular region indicate less restricted flow of water molecules
and thus has been theorized to represent a loss of microstructural integrity, reduced bundle
organization or axonal damage (Scholz et al., 2014). Several studies have demonstrated lower
FA in the WM tracts of children with brain tumors after receiving variable combinations of the
three common treatments, surgery, chemotherapy, and radiation, thus possibly indicating WM
damage due to treatment. Before either chemotherapy or radiation, surgical resection of
posterior fossa tumors alone impacts the supratentorial brain, as evidenced by lower FA and
decreases to WM volume in structures such as the corpus callosum and corona radiata (Glass
et al., 2017; Rueckriegal et al., 2010; Reddick et al., 2005). The combination of surgery,
chemotherapy and radiation has been shown to be detrimental with many studies documenting
reduced FA in several brain structures, including the corpus callosum and frontal WM (Jacola et
al., 2014; Monje et al., 2007; Reddick et al., 2005; Fouladi et al., 2004). Although literature on
solely chemotherapy’s effects on children with brain tumors is sparse, our recent paper in Baron
Nelson et al., 2021 found patterns of FA differences in gray and white matter structures
associated with the effects of surgery and chemotherapy compared to surgery alone. Thus,
examining differences in FA values is important in determining WM damage in relation to
different treatments.
Many approaches have been developed to identify the FA values in WM tracts. Manually
drawn regions of interest (ROIs) ensure anatomical accuracy. However, voxel-based methods
like this have been replaced by the development of automated methods of WM identification as
the number of subjects in studies has increased. To address the increasing workload and time
required to analyze larger datasets, tract-based spatial statistics (TBSS) (Smith et al., 2006)
became the standard automated method in the study of WM tracts. TBSS accomplishes this by
53

first nonlinearly registering the FA maps from individual scans. The normalized FA maps are
averaged and eroded to produce a WM skeleton which represents the core of all the WM tracts
common to the initial scans. With the method’s rise in popularity, researchers have become
increasingly aware of its limitations. For example, TBSS’s projection onto an entire WM skeleton
does not allow a researcher to distinguish between distinct but adjacent WM tracts (Bach et
al.,2014).
Tract specific analysis (TSA) was designed to remedy these problems by segmenting
individual WM tracts onto population specific templates (Yushkevich et al., 2008; Zhang et al.,
2010). Tensors are then projected onto a medial sheet which both defines the skeleton and
informs the boundary of a WM tract. The maximum or mean tensor values can be calculated
along a spoke extending perpendicularly from a point on the medial sheet to the tract boundary.
From these values, DTI metrics such as FA can be calculated. Thus, TSA can provide FA
values for specific WM tracts without noise from adjacent tracts. A previous study from this lab
comparing TBSS and TSA has shown that in a comparison between a congenital blind group
and healthy sighted controls, TSA shows higher sensitivity in detecting subtle differences in WM
(Lao et al., 2015).
In this study, we investigate WM microstructure differences between three groups: (1)
pediatric brain tumor survivors who underwent surgery only, (2) those who underwent surgery
and received chemotherapy, and (3) healthy controls, by utilizing TSA to compare FA values
across 11 major WM tracts. We seek to further delineate neuroimaging findings from our
previous study on the same cohort which identified clusters of lower FA in children treated with
surgery and chemotherapy than in those treated with surgery in the superior longitudinal
fasciculus (SLF) bilaterally and in the left uncinate fasciculus (UF) using whole brain FA voxel-
based analysis (Baron Nelson et al., 2021). Based on our previous study and findings of others,
we hypothesize that TSA will indicate a pattern of injury to the corpus callosum, SLF, and UF in
children with brain tumors compared to healthy controls, and that those children treated with
54

chemotherapy in addition to surgery will have a pattern of injury that is more widespread than
children treated with surgery alone.
Materials and Methods
Participants
Participant demographics and recruitment are the same from a prior study (Baron
Nelson et al., 2021). All included participants were between 6 and 17 years old, inclusive and
were also required to speak and read either English or Spanish. All included patient participants
had 1) tumor location in the posterior fossa – cerebellum or fourth ventricle, 2) complete tumor
resection with no evidence of metastasis more than 1 year after treatment, and 3) at least one
year since the patient’s last treatment for brain tumor. All included control participants had no
prior history of traumatic brain injury or mental disease. Potential participants were excluded if
they met the following criteria: 1) metal in the body, 2) preterm birth, 3) neurodevelopmental
disability, 4) traumatic brain injury, or 5) turning 18 years old during study duration. Potential
patient participants were excluded if they had a recurrent tumor, residual disease outside of the
posterior fossa or a history of posterior fossa syndrome. Potential control participants were also
excluded if they needed sedation for an MRI scan. Patient demographics such as age, gender,
race/ethnicity, tumor diagnosis site and time off treatment can be found in Table 3.
The Institutional Review Board at Children’s Hospital Los Angeles approved this study.
The Neuro-oncology database and clinic lists were used to identify potential
subjects. Recruitment of subject families took place in clinic or by mail or phone.
Eight of 10 children (80%) in the chemotherapy group received intensive marrow-
ablative chemotherapy followed by autologous hematopoietic stem cell transplant (AuHSCT)
with some combination of thiotepa, etoposide or carboplatin as the conditioning regimen. Most
children in both groups received cisplatin, cyclophosphomide, and vincristine.
55

Two subject participants from the surgery group were excluded after pre-processing but before
analysis due to inadequate registration of the subject MRI data.

Table 3. Patient demographics by treatment group compared to controls
Variable Treatment group

Surgery
(n = 15)
Surgery + Chemo
(n = 10)
HC
(n = 21)
p-value
Age – mean (SD)
10.60 (4.12) 12.50 (3.54) 10.52 (2.27) 0.26
Age at Diagnosis – mean (SD)
5.55 (3.08) 3.77 (3.36) N/A 0.19
Sex Male 4 (27%) 6 (60%) 13 (62%)
0.09
Female 11 (73%) 4 (40%) 8 (38%)
Patient
Race/ethnicity
White

Non-Hispanic

Hispanic/Latino
13 (87%) 9 (90%) 18 (86%)
Race – 0.84

Ethnicity – 0.23
6 (46%) 3 (33%) 10 (56%)
7 (54%) 6 (67%) 8 (44%)
Black/African
American
1 (7%)

0 (0%) 0 (0%)
Asian 1 (7%) 1 (10%) 3 (14%)
Diagnosis Medulloblastoma 0 (0%) 6 (60%)
N/A 0.0005

Ependymoma 0 (0%) 1 (10%)
Astrocytoma 2 (14%) 0 (0%)
Pilocytic Astrocytoma 12 (80%) 2 (20%)
Other 1 (6%) 1 (10%)
Time Off Treatment (years) – mean (SD)
4.67 (3.29)

8.18 (4.97) N/A 0.04

Neuroimaging
T1-weighted images were obtained on a 3.0 T Philips Achieva scanner with voxel size 1.0 x 1.0
x 1.0 mm
3
with parameters: TR 9.9 ms; TE 4.6 ms; 240x 231 matrix; FOV 24 cm.
Diffusion Weighted Images (DWI) were acquired using a novel fast DWI sequence totaling 11
minutes, with parameters: 70 axial slices (2 mm thick), FOV = 256 mm x 256 mm x 140 mm,
TR/TE 8657/86 ms, no gap, with a 128x126 acquisition matrix, 28 gradient images collected
with b-value=1500.
56

Registration and Sampling
T1 and DWI data were visually inspected for major artifacts and signal drop off. T1
images were bias field corrected using ANTs (Avants et al., 2009), N4 BFC (Tustison et al.,
2009) tool, and manually skull-stripped in Brainsuite16 (Shattuck et al., 2001). DW images were
first visually inspected for motion artifact and noisy volumes were excluded. DW images were
processed through FSL’s eddy (Andersson et al., 2016) for eddy current correction and subject
motion correction. DW images were skull-stripped using DSI Studio followed by tensor
estimation using FSL’s DTIFIT. Tensors were then formatted for use with TSA with DTI-TK’s
toolbox. The generation of a dataset-specific template to create the medial representations of
white matter tracts created inconsistent registration results, thus we opted to use the adult
template publicly available on the DTI-TK site as the registration target. We sequentially
registered all subjects through a rigid, affine, then nonlinear registration process through the
DTI-TK toolbox to align subjects into the atlas space (Zhang et al., 2007). We used the same
toolbox to generate fractional anisotropy (FA) values from the registered subject DT images.
The FA values within the boundaries of a WM tract were then projected onto a thinner sheet-like
representation of the tract which snakes through the tract’s mid-plane. Each point on that
medially located surface held the average value of the FA values projected onto it.
Eleven major white matter tracts available in the standard release of the software were tested:
the corpus callosum (CC), and bilateral cortico-spinal tracts (CST), inferior fronto-occipital
fasciculi (IFOF), inferior longitudinal fasciculi (ILF), superior longitudinal fasciculi (SLF), and
uncinate fasciculi (UF).

57

Analysis
A supra-threshold statistical model was used to assess differences in FA between
patient groups and healthy controls (Yushkevich et al., 2008). At each point on the medial
surface of the tract, a two-sample t-test was computed. An arbitrary value t 0 was used to extract
clusters on the medial surface for which their t values are less than t 0. The size of the cluster (in
terms of number of points of the surface) was then collected into a histogram. This process was
repeated 10,000 times, but for each instance, the labels of the subjects were randomly
permuted. Thus, a non-parametric permutation-based cluster analysis method (Nichols &
Holmes, 2002) was used to correct for the family-wise error rate (FWER), considering the
number of WM tracts. The threshold p-value was set to 0.01 and the number of permutations to
10,000. We included age at study in the general linear model in the TSA pipeline to control for
relevant confounding factors. For quality assurance, all corrected clusters were overlaid across
each subject’s individual FA map. If the cluster aligned outside of or along the edge of the white
matter skeleton for most patients, then the cluster would be omitted as we could not determine if
the results were due to inadequate registration.
Visualizations of the statistically significant clusters as determined after multiple
comparisons correction were created in ParaView (Ahrens et al., 2005), an open-source data
analysis and visualization application.
Results
Demographics for the three comparison groups are shown in Table 3. Most children in
the surgery group had a diagnosis of pilocytic astrocytoma, while those in the surgery and
chemotherapy group were younger at diagnosis, most often with medulloblastoma, and had
been off treatment longer.
58

Tract Specific Analysis
Between the two treatment groups of children with brain tumors (surgery vs surgery and
chemotherapy), there were no significant FA differences in any of the 11 tracts after correcting
for multiple comparisons using the permutation-based cluster analysis. However, we found
statistically significant group differences in four tracts (the CC, left and right CST, and right
IFOF) between the surgery group and healthy controls and in five tracts (the CC, left and right
CST, right IFOF, and left UF) between the surgery and chemotherapy group and controls (Fig.
12). In each group comparison, the red clusters indicate areas of lower FA in the patient
population in comparison to healthy controls. There were no significant clusters which indicated
higher FA in the patient population compared to healthy controls.
59

Surgery vs Healthy Controls

Figure 12. WM Tract Results in the Surgery vs Healthy Control comparison. All results displayed on a transparent
glass brain. The background tract is in blue while the significant clusters after multiple comparisons correction are
displayed in red. The three images, left to right, display results in a leftward facing sagittal view, inferior facing axial
view, and rightward facing sagittal view in the (a) Corpus callosum (CC); (b) Bilateral corticospinal tract (CST); and (c)
Right inferior frontal occipital fasciculus (right IFOF)

In the CC, Fig. 12a, we see a large significant cluster in the splenium of the CC
indicating lower FA in the surgery group compared to healthy controls. This region contains
fibers from the superior temporal, inferior temporal, and occipital areas of the brain
(Schmahmann et al., 2007). In the bilateral CST, Fig. 12b, there are small clusters of significant
60

difference toward the posterior part of the tract. The corticospinal tract connects the motor
cortex to the spinal cord through the brainstem and thus is responsible for voluntary movements
of the limbs and trunk (Davidoff 1990). In the right IFOF, Fig. 12c, we see clusters of
significantly reduced FA in the surgery group in the posterior part of the tract. The IFOF
connects all lobes of the brain and is thought to play a key role in nonverbal semantic cognition
(Herbet et al., 2017), language and attention (Altieri et al., 2019).
There were also small clusters of significance found in the right inferior longitudinal fasciculus
(ILF), left IFOF, and bilateral superior longitudinal fasciculus (SLF). However, these were
omitted from visualization and the discussion as the cluster aligned outside of or along the edge
of the white matter skeleton for most patients, thus we could not determine if the results were
due to misregistration. No clusters of significance were found in the left ILF or bilateral UNC.
61

Surgery and Chemotherapy vs Healthy Controls

Figure 13. WM Tract Results in the Surgery and Chemotherapy vs Healthy Controls comparison. All results displayed
on a transparent glass brain. The background tract is in blue while the significant clusters after multiple comparisons
correction are displayed in red. The three columns, left to right, display results in a leftward facing sagittal view,
inferior facing axial view, and rightward facing sagittal view in the (a) Corpus callosum (CC); (b) Bilateral corticospinal
tract (CST); (c) Right inferior frontal occipital fasciculus (right IFOF); and (d) left uncinate fasciculus (left UF)
62

With visual comparison between Figure 12 and 13, there are similar findings between
each of the two patient treatment groups and healthy controls in the CC, bilateral CST, and R
IFOF. Interestingly, there are results in the L UF that exist in the surgery and chemotherapy
comparison and not in the surgery comparison. The uncinate fasciculus is a limbic fiber tract
which connects the orbitofrontal cortex to the anterior temporal lobes and may affect memory
retrieval mechanisms (Olson et al., 2015).
As for the other four tracts, visual inspection between the two comparisons does provide
some observable differences in cluster appearance. In the CC, both comparisons have many
clusters close to the splenium, but the results in the surgery vs healthy controls comparison
contain a single prominent cluster. In the bilateral CST, both comparisons have clusters of
significant difference in the posterior parts of the bilateral CST. However, the clusters in the
surgery and chemotherapy comparison appear larger. In the R IFOF, both comparisons have
clusters closer to the posterior parts of the R IFOF, surgery and chemotherapy comparison
seems to have larger clusters of significance compared to the surgery comparison. Despite
these visual differences between the two patient group comparisons to healthy controls, the
direct statistical tests between the surgery and surgery and chemotherapy groups did not
contain any significant clusters.
There were also small clusters of significance found in the bilateral ILF, left IFOF, and
bilateral SLF. However, these were omitted from visualization and the discussion as the cluster
aligned outside of or along the edge of the white matter skeleton for most patients, thus we
could not determine if the results were due to misregistration. No clusters of significance were
found in the right UNC.
63

Discussion
We examined the differences in Fractional Anisotropy (FA) in 11 major WM tracts
between two groups of pediatric posterior fossa brain tumor survivors who had received
treatment with either surgery alone or surgery + chemotherapy as well as a third group of
healthy controls. FA provides some indication of axonal density, organization, and degree of
myelination, which in turn may provide information about damage to white matter structures
caused by treatment (Rueckriegel et al., 2010). Our results were calculated using a medial
representation of several tracts and a deformable shape analysis technique, Tract Specific
Analysis (TSA), which projects areas of significant WM differences between groups onto
surfaces.
There were several significant clusters found in the comparisons between the children
treated with surgery and healthy controls and between the children treated with surgery and
chemotherapy and healthy controls, even after regressing out the age covariate. Both sets of
comparisons demonstrate significantly lower FA in the patients in the corpus callosum, bilateral
CST, and right IFOF than in healthy controls. Comparing the surgery and chemotherapy group
to controls demonstrated an additional cluster of lower FA in the left UF.
Our findings of lower FA in children with posterior fossa brain tumors in the CC and CST
support those found in previous studies of pediatric brain tumor survivors. In a longitudinal
study of pediatric brain tumor patients from baseline after surgery through treatment to 36
months later, Glass et al. 2017 utilized TBSS and reported reduced FA values in the CC and
CST of patients. However, this investigation studied children treated with a combination of
surgery, chemotherapy, and radiation therapy. Our study shows similar findings of lower FA in
these tracts in both patient groups, indicating that WM damage to these tracts may not be
attributed to cranial irradiation. Since lower FA was present in both patient groups, we are
unable to ascribe such changes to either surgery or to chemotherapy, or even to the effects of
64

the tumor alone. Notably, a study of children with bone and soft tissue tumors outside the CNS
who received chemotherapy also reports lower FA in the CC and CST after treatment (Sleurs et
al., 2018).
Brain structural WM loss or damage can have lasting effects on learning and cognition.
The CC plays a critical role in processing motor, sensory, and cognitive signals from both
hemispheres. Palmer et al. (2012) and Aukema et al. (2009) found FA in the CC to be positively
associated with processing speed in survivors of pediatric brain tumors who had been treated
with surgery, chemotherapy, and radiation. Although research linking reduced FA in the CST to
motor deficits in pediatric brain tumor populations is sparse, studies have found an initial decline
in FA in the CST in 3 children after brain tumor treatment (Hua et al, 2012), and in children after
proton beam irradiation (Uh et al, 2015) that recovered over time. The IFOF is an association
fiber system which connects the occipital cortex, temporo-basal areas and superior parietal lobe
to the frontal lobe and may play a role in reading, attention, and visual processing (Wu et al.,
2016). In studies of pediatric brain tumor survivors who received varying levels of treatment with
surgery, chemotherapy, and radiation, Aleksonis et al. (2019) and Aukema et al. (2009) reported
lower mean FA in the right IFOF with Aukema et al. finding a positive correlation between mean
FA and processing speed. In another study of pediatric survivors of brain tumors treated with all
three treatment modalities, lower mean FA was reported in the UF as well (Riggs, et al., 2014).
The critical difference between most studies reporting on WM microstructure in this population
and our study is that we stratified our population according to treatment type to understand the
effect that different treatments have on the recovering and developing pediatric brain. While the
studies cited above examined the effects of all treatments and found lower FA in the CC, CST,
and IFOF, our study examined children treated with surgery only and found similar results in all
three WM structures. These results may indicate lasting effects on the supratentorial brain by
infratentorial tumors or resection that may be exacerbated by or at least persist through
adjuvant treatment as evidenced by our similar findings in this study’s surgery and
65

chemotherapy group and from other studies on patient groups who had also been treated with
cranial irradiation. Supratentorial WM structures with projections to the cerebellum, such as the
CST, may suffer axonal degeneration if the tumor or surgical resection damages these
cerebellar extensions. Our lack of findings in the direct comparison between the two patient
groups makes it difficult to parse out the additive effect of chemotherapy to surgical
resection. However, our findings in the UF in the surgery and chemotherapy vs healthy control
comparison may indicate a relationship between chemotherapeutic agents and damage to the
UF because there are no similar significant findings in the comparison of children treated with
surgery to controls.
Limitations
To reduce the regions of interest to 11 specific white matter tracts, the continuous medial
representations of the WM tracts do not capture more distal neuronal extensions. Thus, TSA is
unable to reveal full WM connections - especially in deeper WM structures. Additionally, the
diffusion tensor model is inherently unable to accurately describe voxels which contain WM
fibers that may cross, fan, bend, or branch and may generate underestimated FA values in
those regions (Seunarine et al., 2014). Indeed, crossing fibers are a common limitation of DTI
studies. Thus, our future research will utilize methods which can distinguish multiple fibers
(Garyfallidis et al., 2014) These limitations as well as the small sample size of this study make
conclusive interpretations of the results difficult to produce. Finally, patients were not matched to
controls by gender, age, language, or handedness, although 48% of controls were siblings of
the participants.
66

Conclusion
In conclusion, our study is the first to investigate the impact of chemotherapy and/or
surgery separately on microstructural changes in the 11 major WM tracts. Statistically significant
clusters of decreased FA in the splenium of the CC, bilateral posterior CST, and posterior right
IFOF were observed in both groups of patients compared to their healthy age-matched sibling
controls. Decreased FA was also observed in the more superior part of the UF when comparing
the surgery and chemotherapy group to healthy controls. No significant clusters of FA
differences were found between the two patient groups. Findings in the surgery vs control group
indicate that surgical resection in the cerebellar region, while necessary, may have effects on
the relatively distant supratentorial white matter. Our study findings support others of children
with brain tumors that also report decreased mean FA in WM tracts. However, unlike many
previous studies, we controlled for tumor location and stratified our dataset by treatment type
(Baron Nelson et al., 2021). Further work on this study population will include a third patient
group who received surgery, chemotherapy, and radiation to further understand WM damage
and neurocognitive functioning in pediatric survivors of posterior fossa brain tumors.

67

Chapter 4. Normal Cerebral Ventricular Shape in Early Childhood
Jeffrey Tanedo
1,2
, Niharika Gajawelli
1
, Miao Wei
1
, Siri Sahib S. Khalsa
3
, Cormac O. Maher
3
, Qunxi Dong
4
,
Darryl Hwang
6
, Sean Deoni
5
, Yalin Wang
4
, Natasha Lepore
1,2,6

1
CIBORG Laboratory, Children’s Hospital Los Angeles,
Department of Radiology, Los Angeles, CA, USA

2
Department of Biomedical Engineering,
University of Southern California, Los Angeles, CA, USA

3
Department of Neurosurgery, University of Michigan, Ann Arbor, Michigan

4
Department of Computer Science, Arizona State University, AZ, USA

5
Department of Pediatrics, Warren Alpert Medical School at Brown University, RI, USA

6
Keck School of Medicine, University of Southern California, Los Angeles, CA, USA

68

Introduction
The lateral ventricles of the brain are centrally located cavities of cerebrospinal fluid that
support and cushion the surrounding tissues of the brain. The spectrum of ventricular
abnormality can extend from slit or collapsed ventricles to ventriculomegaly, otherwise known as
enlarged or expanded ventricles. Depending on the pathology, enlarged ventricles can either
cause or be the effect of damage to structures across the entire cortex. These conditions can
occur across the entire lifespan in clinical populations.
Utilizing MRI volumetric analyses, the enlargement of the lateral ventricles was
demonstrated in various psychological, neurological, and even endocrinological disorders that
arise during childhood and adolescence. Psychological disorders related to larger lateral
ventricles included but not limited to autism spectrum disorder (Turner, Greenspan, & van Erp,
2016), childhood-onset schizophrenia (Kumra et al., 2000; Rapoport et al., 1997), bipolar
disorder (Edmiston et al., 2011), maltreatment-related posttraumatic stress disorder (De Bellis et
al., 2002), anorexia nervosa (Katzman, Zipursky, Lambe, & Mikulis, 1997). It was also
associated with neurological disorders, such as Down syndrome (Pearlson et al., 1998),
Klinefelter Syndrome (J.N. et al., 2007), maternal opioid use (Q. et al., 2014), and cerebral palsy
(Kulak, Maciorkowska, & Goscik, 2016). Left lateral ventricle was found to be significantly larger
in the chorioamnionitis exposed children as compared to control group (Hatfield et al., 2011).
Preterm baby with Intraventricular hemorrhage is at risk for ventricular dilation (Qiu, Yuan,
Kishimoto, Ukwatta, & Fenster, 2013). Endocrinologically, children with Cushing syndrome had
significantly larger lateral ventricles when comparing to controls. However, because the
naturally occurring volume increase of lateral ventricles, scientists and clinicians must interpret
these results with caution as dysmorphic ventricular systems may happen to share the same
numerical volume as healthy ventricular systems.
69

The lateral ventricle also tends to have the highest variability of morphometric measures
due to its border with multiple structures (Lenroot & Giedd, 2006). Some disorders would involve
broader ventricle enlargement, for example, Fragile X syndrome was reported to cause the
enlargement of lateral ventricle (Lee et al., 2007), 3rd ventricle (Murphy et al., 1999) and 4th
ventricle (Reiss, Aylward, Freund, Joshi, & Bryan, 1991). Supratentorial ventricular system
volume was increased in asymptomatic children with achondroplasia (Calandrelli et al., 2017).
Increased total ventricular volume was associated with Apert syndrome and Chiari I
malformations (Jong et al., 2011). Children with new-onset idiopathic generalized epilepsy had
significantly larger lateral and 3rd ventricle volumes relative to the control group (Jackson et al.,
2011).
With a better understanding of normal ventricle development in early childhood, it may
be possible to identify children who are at high risk to develop these disorders earlier. The most
current analysis on normal ventricle development by Cutler et al. 2020 presents normal
percentile growth charts based on ventricular volume calculated from 3D MR images. To our
knowledge, no studies have published the precise variations in shape of the ventricular system.
In our previous works, we applied a multivariate tensor-based morphometry (mTBM) method to
preterm children to measure the morphometric deformations in preterm born neonates and full-
term infants in the cerebral ventricles (Paquette et al., 2017), the anterior and inferior portion of
the putamen, and the ventral portion of the thalamic nuclei (Lao et al., 2014; Shi et al., 2013).
We expand on this application by implementing the mTBM analysis in addition to an analysis of
the change of ventricle thickness to determine cerebral ventricular shape changes across the
first five years of normal child development.

70

Methods
Healthy pediatric sample
Our analyzed sample included brain MR data from 104 healthy pediatric subjects in ages
ranging from 347 - 1991 days old from the Advanced Baby Imaging Lab
(www.babyimaginglab.com). Only 12 of the 104 subjects had several time points. 1 female
subject had three time points. 6 male and 5 female subjects had two time points. This resulted
in a set of 117 MR images.
The data used consisted of T1 MP-RAGE MRI scans (1.4–1.8 mm
3
) of healthy typically
developing children with the following inclusion criteria: singleton birth between 37- and 42-
weeks gestation with no abnormalities on fetal ultrasound and no reported history of
neurological events or disorders during infancy. Detailed data acquisition parameters can be
found in (Deoni et al. 2003, 2006). The study was approved by the Institutional Review Board of
Brown University and informed consent was obtained from the guardians of all participants. All
data was deidentified before pre-processing. Subject handedness is not assessed in this study
due to lack of handedness information at early ages.
The dataset of 117 MR images was binned into 5 age groups at each consecutive age
from 1 to 5 years ± 6 months (e.g., 1 year old group spanned 6 months to 18 months, 2-year-old
group spanned 18 months to 30, etc.). Distribution details can be seen below in Table 4.
Table 4. Distribution of subjects in different age groups
Age Mean age / std. dev (days) Male Female Total
1-year-old 382.6 / 23.9 10 10 20
2-year-old 728.5 / 52.1 15 8 23
3-year-old 1067.7 / 88.1 18 6 24
4-year-old 1395.9 / 97.4 16 10 26
5-year-old 1854.9 / 105.6 11 13 24

71

Processing
MR images of brains were first skull-stripped using BrainSuite (Shattuck and Leahy
2002). The images were linearly registered with 6 degrees of freedom to an age matched
template with FSL FLIRT (M Jenkinson and Smith 2001; Jenkinson et al. 2002). Bias field
correction was then applied to the images using a geodesic intensity correction algorithm
(Gaonkar et al. 2015). Finally, each image was resampled to a 1×1×1 mm
3
space.
Comprehensive details on the method of extracting multivariate tensor based
morphometric (mTBM) features from surface maps can be found in Paquette et al., 2017. The
outline of the mTBM process will be described in this paper. First, bilateral ventricular structures
were delineated by an expert in the Insight Toolkit's SNAP program (Yushkevich et al., 2006)
and reviewed by an experienced pediatric neuro-radiologist. By applying a topology-preserving
level set method (Han et al., 2003), we can build surface models from the binary segmentations
of the ventricles. Finally, the marching cube algorithm (Lorensen and Cline, 1987) is applied to
the surface model to obtain the ventricular model represented as tetrahedral meshes, which are
discretized into surfaces described by vertices in ℝ
3
, edges and faces. To register the mesh so
that we can compare ventricular surfaces among subjects, we then conformally map the
ventricular surfaces to a two-dimensional hyperbolic plane which preserves the angles between
vertices. In this plane, we can register subject ventricular surfaces to the template ventricular
surface.
Multivariate tensor-based morphometry
The surface deformations that occur in the surface registration process are essential to
the multivariate tensor-based morphometry (mTBM) process (Lepore et al., 2008; Wang et al.,
2008). First, we compute the Jacobian matrix (J) at each vertex from the transformation
between the common template and the subject’s shape by multiplying [w3 – w1, w2 –w1] [v3 –
72

v1, v2 – v1]
−1
, where [v1,v2,v3] is a triangle on the subject mesh being mapped on a triangle on
a template mesh [w1, w2, w3] to obtain a 2x2 matrix. The deformation tensor is formed by
defining log(S) = log(sqrt(J
T
J)) which gives us three unique elements from the upper triangle of
the 2x2 matrix. This 3x1 feature vector is thus used as a multivariate measure (Lepore et al.,
2008). We also obtain the determinant of the Jacobian (det J) as a scalar measure of the
magnitude and directionality of local area changes on the surface of the ventricular mesh.
Medial Axial Distance
While the 3x1 mTBM feature vector and the scalar det J are sensitive to shear changes
along the tangent direction of the ventricular surface, we can further investigate extensional
changes along the normal direction of the ventricular surface by analyzing the Medial Axial
Distance (MAD). We calculate MAD at each vertex by calculating the radial distance from the
medial axis of the ventricles to that vertex. This medial axis is obtained by using the mid-point
of the iso-parametric curves on the ventricular surface. By combining the 3x1 mTBM feature
vector with the scalar MAD, we can construct the 4x1 multivariate feature vector named
MADMTBM. To intuitively display the direction of changes between groups, we mapped the
ratio of the mean MAD values of an older group divided by the mean MAD values of a younger
group. The result is such that values greater than 1 indicate expansion from younger to older at
that particular vertex whereas values lesser than 1 indicate contraction of the thickness at that
particular vertex. To characterize the normative thickness at each age, we also calculated and
displayed the standard deviation of MAD values at each vertex for the five age groups.
Statistical Analysis
We applied the multivariate Hotelling's T2 test, a multivariate generalization of the t-test
(Hotelling, 1931), with the four measures mentioned above (detJ, MAD, MTBM and MADMTBM)
to calculate vertex-wise group differences among each consecutive age group pairing (1v2, 2v3,
73

3v4, and 4v5) in addition to comparing the largest age group pairing between the 1- and 5-year-
old group. Each test was performed on both left and right ventricles separately.
To correct for multiple comparisons (Nichols and Holmes, 2002), we applied permutation tests
at N = 10,000. P-values are calculated by comparing the original data to the permutation
distribution. The resulting group comparison p-map was used to visualize the different
morphometry patterns between subsequent age groups. The permutation corrected p-value
allows us to measure the overall significance of the whole structure’s p-map at the 0.05 alpha
level.
Statistical analysis for growth curves were developed by applying the Generalized
Additive Models for Location Shape and Scale (GAMLSS) R package utilizing the lambda-mu-
sigma smoothing method (LMS). The LMS uses Box-Cox power transformation and maximum
penalized likelihood to obtain optimal smoothed coefficients for median (mu), variation (sigma)
and skewness (lambda) across the age range (Kuczmarski RJ et al., 2002).
Results
Global maps of significance for all measures at each comparison are displayed in Figure
15. Each block of images depicts 2 views of each ventricle for a total of 4 maps. The center two
images display the superior surface of both ventricles with the most anterior point facing
downwards. The “wings” of each block display the medial side of the right ventricle and the
medial side of the left ventricle. A * indicates that the whole global map for that particular
ventricle, measure, and age group comparison remained significant after permutation analysis.

74

Figure 14. Global significance maps displaying MADMTBM between the 1st and 2nd year old groups (upper left), 2nd
and 3rd year old groups (upper right), 3rd and 4th year old groups (left second row), 4th and 5th year old groups
(right second row) and 1st and 5th y year old groups (bottom row). * Indicates that the whole map for that ventricle
and age comparison is significant.

Large areas of significance between groups can be seen on the superior section of the
central part of the lateral ventricle. In the 2v3 and 3v4 comparisons, similar areas can be seen
extending anteriorly along the superior surface of the ventricle towards the frontal horn.
75

Figure 15. MAD Mean Ratio. Yellow and red colors indicate areas of greater ventricle thickness in the older age
group. Light blue to blue indicates areas of less thickness in the older age group.

In Figure 16, we see what appears to be large areas of expansion from ages 1 to 2 and
3 to 4 with areas of contraction or shrinking of the thickness measure in the comparison
between ages 2 and 3 then 4 to 5. The 1v5 comparison seems generally have areas of
shrinking along the lateral edges along with a few areas which seem to increase the MAD along
the superior surface.
76

Figure 16. Bilateral ventricular ratio. Gray dots represent individual data points. Curved lines represent from top to
bottom 95th, 90th, 75th, 50th, 25th, 10th, and 5th percentile curves
In Figure 17, we see the percentile curves fit to the bilateral ventricle volume ratio.
There is an overall decrease in ventricular volume from 1 year of age to 5 years of age.
77

Figure 17. Standard deviation of medial axial distance values across each individual subject group in mm
In Figure 18, we see the expected variability of thickness at each mesh in mm. Blue
denotes areas of low standard deviation of ventricle thickness whereas red indicates areas of
increased variability in mesh thickness for a given age group.

78

Discussion
In this study, we examined the morphological changes in the lateral ventricles in a
healthy population of children between 1 to 5 years of age. We precisely localized areas of
significant shape change between the ventricles of these groups via four measures of
morphological change - the determinant of J, MAD, MTBM and MADMTBM.
In reference to the significance maps, the analyses using MADMTBM were most
sensitive to areas of significant change along the ventricular surface. The significance maps of
det J, MAD and MTBM will be made available with the supplementary material. In the 1v2
comparison, we found areas of change on the central body, especially on the lateral sides of the
tip of the frontal horns. The most widespread areas of significance are seen in the comparisons
around the 3-year-old group with both the 2v3 and 3v4 comparisons displaying large areas of
significance along the superior surface of the ventricles, especially on the left side. The 3v4
comparison map also displays these significant areas spread further to the posterior bilaterally.
The 4v5 comparison map retains the display of these posterior areas, but notably lacking any
widespread areas closer to the frontal horn. Interestingly, the 1v5 comparison shows areas of
change along the superior surface similar to the areas seen in the 2v4 and 3v4 comparisons
even though the comparison did not include data from those groups. This result seems to
reinforce the notion that some significant shape change occurs in those areas at some point
during development around three years old.
As the lateral ventricles are pockets of cerebrospinal fluid in the brain, their change in
shape and size in development is reflected by the developmental changes happening to their
surrounding structures such as the corpus callosum. A similar analysis on those surrounding
structures would theoretically complement the analysis of the ventricles, especially with
MADMTBM’s inclusion of the medial axial distance. Indeed, our group recently submitted an
analysis of the corpus callosum in young children of ages 1-5 (Gajawelli et al., unpublished)
79

using the same measures. Knowing that the CC lies directly above the lateral ventricles, we
visually compared the significance maps and found that in 1v2, and 2v3 comparisons, the
medial section of the frontal horn of the ventricle seemed to match with areas of significance on
the inferior side of the CC. In our 4v5 comparisons, the superior surface of the central part of the
lateral ventricle seems to match with the inferior surface of the CC body. Finally, the widespread
areas of difference along the superior surface of the ventricles seen in the 1v5 comparison are
similarly reflected in areas of significance along the inferior surface of the CC extending from the
genu to the body. A formal correlation analysis between findings of both manuscripts will be
forthcoming.
To understand the results of this analysis in a more intuitive sense, we also examined
the directionality of MAD changes. Our results display an expansion of the ventricles along the
lateral and medial edges from years 1 to 2 and 3 to 4 with a subsequent contraction of the
ventricles along similar edges from years 2 to 3 and 4 to 5. As this is the first time a localized
analysis such as this has been performed on the lateral ventricles, there is no documentation on
the expected deformations that occur along the ventricle surface as the brain continues to
develop.
However, as some studies have demonstrated (Coupe et al., 2017, Remer et al., 2017,
Martini et al., 2018, Reynolds et al., 2019, Deoni et al., 2012), each structure of the brain does
not grow linearly or at the same rate as other structures of the brain at a given age. Richards et
al. 2015 found rapid increases in GM volume through early childhood compared to a more
gradual increase in WM volume in early childhood. They also found that occipital WM and GM
volume appeared to have little change from birth till 4 years old whereas frontal, parietal and
temporal WM and GM volumes increased in the same age range. Deoni et al., 2012 found that
WM myelination development can be different within the same structure such as the CC. Among
the functions of the lateral ventricles, one may infer that these fluid-filled pockets of the brain
can act as a buffer for the variations in pressure that could occur due to these complex growth
80

patterns across the brain. We hypothesize that in addition to the pressures of volumetric change
there are also changes to the density and thus physical compliance of white matter tissue as it
myelinates.
Our analysis on the volumetric ratio of the ventricular volumes compared to the total
intracranial volume may indicate that the ventricles account for a decreasing percentage of
space in the developing brain compared to developing gray and white matter. Indeed, this
analysis on volumetric ratio agrees with previous findings in Xenos et al., 2002 in which a
general decrease of volumetric was found in early childhood.
To demonstrate the utility of this level of detailed morphometric analysis, we also display
the map of standard deviations of thickness at each vertex location along the ventricular surface
for each age group. The map of thickness standard deviations provides a visualization of the
expected variability of thickness in exact locations on the ventricular surface. In future studies,
we may be able to identify dysmorphic ventricular anatomy as objectively outside the normal by
comparing pathologic ventricular surfaces to these maps.
Conclusion
In this study, we demonstrated the use of multivariate TBM to identify shape and size related
differences in the lateral ventricular surfaces of typically developing children between the ages
of 1 and 5 years. These changes were most widespread in the comparisons between 2- and 3-
year-old groups and 3- and 4-year-old groups. This extent of localized differences was also
reflected in the direct comparison between 1- and 5-year-old groups. We also demonstrated the
relative change in lateral ventricle thickness from year to year revealing a pattern of expansion
and contraction which may be related to competing intracranial pressure due to the brain’s
continued development. Finally, we displayed the localized variability of lateral ventricle
thickness to intuitively characterize the shape of typically developing ventricles.
81

Methodological Limitations and future directions
Longitudinal data could help strengthen our inferences, but such data does not exist in
sufficient quantities with this dataset. In the absence of longitudinal data (which will be
forthcoming in the following years), one promising future direction would be to extend the
morphometric analysis to subcortical structures which border the ventricles. In addition to the
ventricles and the CC, the mTBM methods described may be applied to the thalamus, and
caudate nucleus to understand not only the typical development of those structures in
childhood, but also their relationships to each other in the complex balancing act that is the
maturation of the brain.
Acknowledgements
The authors would like to thank Nicholas Chapman, Danielle Darakjian, Nathan Menard,
Brittany Randles, Shringala Chelluri, Hannah Telle, Ipek Narbay, Michelle Liang, Diandra
Decampos-Kundahl and Malia Valder for their support in segmenting the ventricular surface.

82

Chapter 5. Conclusion and Future Work
My thesis research began with an investigation into the differences in neuroanatomy
between survivors of posterior fossa brain tumors who received different treatments for their
tumor. These survivors often suffered from neurocognitive disorders which we believe to be
more related to structures in the supratentorial brain. We know that treatments of surgery,
chemotherapy and radiation generally incur further damage to neural structures, but the exact
mechanism and extent of that damage is understudied. Thus, we built an MR dataset of
pediatric survivors of posterior fossa tumors who had received combinations of surgery,
chemotherapy, and radiation in addition to age-matched healthy controls.
We saw and overcame many challenges in processing the MR dataset. Among them
was the lack of a pediatric specific pipeline and thus we became acquainted with all the MRI
research programs that were initially designed for adults. In addition to building a pipeline
piecemeal from different software, I developed a registration protocol to reduce the bias that
could be introduced when performing the normalization.
Normalization is a critical aspect of MR imaging to ensure that comparisons between
images of the brain are in the same space. However, that normalization can introduce bias if the
distance between the subject and the template is too large. Considering that the dataset ranged
from 6-17 years of age, there was a large amount of variation due to the typical variation
introduced by normal human development. Thus, I reduced that distance by dividing the group
into three and first registering each group to a template which was closer to their age group.
Then, I performed another registration to a template which existed in the space between the
initial three templates, thus bringing all patients into the same space.
With this, we were able to perform our first set of analyses into the DW images of three
groups: posterior fossa tumor survivors who received surgery, those who received surgery and
chemotherapy and healthy controls. Using whole brain voxel-wise FA as the primary
83

investigative measure, we discovered evidence of FA changes in several WM and GM
structures in the supratentorial brain. These changes were evident in both comparisons against
healthy controls, but also in comparisons between surgery and surgery and chemotherapy.
First, the comparisons to healthy controls were already enough to add to the literature
concerning possible sites of neural damage that occur due to surgery and the surgery and
chemotherapy combination. Secondly, the findings between surgery and chemotherapy
provided published literature on chemotherapy specific sites of possible neural damage –
important findings in what is a sparse set of literature on chemotherapy’s effects on pediatric
survivors of posterior fossa brain tumors.
Bolstered by this finding, we sought to investigate further into the differences by focusing
the analysis on WM tracts. This would help to improve statistical power over doing a whole brain
FA analysis by focusing the analysis on a smaller part of the whole brain. This way, there was
less of a chance of multiple comparisons correction nulling out any potential findings. Indeed,
the analysis displayed differences between posterior fossa brain tumor patients who had
surgery and surgery/chemotherapy combo versus healthy controls. However, we did not find
any differences between surgery and the surgery/chemotherapy group. Initially, I found this
disheartening so I went back to the proverbial drawing board to wonder about what next steps
could be taken.
One of the potential confounds of research across large age ranges in pediatric brain
research is the possibility of capturing the typical variability due to normal development. Of
course, we include age as a covariate when doing statistical analyses. The understanding of
that variability is necessarily set aside to look at the pathological variability of interest. I became
interested more in that typical variability and characterizing the normal development of
children’s brains as I believe that such a contribution would benefit not only pediatric brain
cancer research, but also any pediatric research.
84

During quality assessment of the pediatric posterior fossa tumor cohort, we frequently
noted abnormal ventricle size. During the initial processing of the dataset, this was a problem to
be overcome through registration or pruning of the data if the ventricles were too abnormal. I
wondered if the presentation of the ventricles could possibly be an avenue of research to
determine differences between treatment groups. Much of the literature on the healthy
development of ventricles primarily utilized linear measurements summarizing the ventricular
system such as ratios of distances found on a few slices of the MR image. This method was
initially developed to support neurosurgeons in making quick decisions for emergency MRIs.
There were also studies on the volume of the ventricles through development, but none were
able to locally characterize the shape and size changes that occurred to the ventricle through
development.
Based off the literature, we hypothesized that there would be some nonlinear pattern of
shape changes in the cerebral ventricles as children developed. Thus, with Natasha Lepore’s
lab history of exploring the shape change of subcortical structures in pediatric MR images, we
decided to apply the mTBM method to a large MR database of healthy children. With this, we
indeed found evidence of nonlinear changes across the ventricular surface. We were able to
corroborate two interesting findings from the latest ventricular volume study by Cutler et al.,
2020. We found several areas of significant difference indicating size and shape change
between the ventricles of 1 through 5 years old children. Specifically, there were widespread
areas of change from 2 through 4 years old that corresponded with a period around 3.5 years of
age from previous studies [Cutler et al., 2020] where ventricular volume switched from
decreasing with age to increasing with age. Another noteworthy finding from the significance
maps is that areas of significance in the comparison of 1- and 5-year-old children seemed to
reflect a combination of the areas of significance seen in the previous pairwise combinations
between age groups. We were also able to display an intuitive map of the ventricles indicating
the expected variability of ventricle thickness at each age from 1 through 5 years. This marks
85

the first step towards creating a more informed picture of what normative pediatric ventricular
morphology looks like.
The first arc of my thesis provides a pathway for others studying pediatric MRI to inform
them on how to design processing and analysis pipelines that address the challenges pediatric
MRI usually bring. The second arc of my thesis provides a step towards the understanding the
normal development of pediatric ventricles. I hope that both arcs support pediatric MR research
to come.

Future Work
I have already accepted a post-doctoral research fellowship with Dr. Natasha Lepore at
Children’s Hospital Los Angeles to continue studying normal child brain development.
Specifically, I will continue to characterize the morphological development of the hippocampus
in the same pediatric dataset and same age range. I am joining others such as Dr. Niharika
Gajawelli in mapping the normal developmental trajectories of the subcortical structures of the
brain. In addition to this continued research, I will pursue opportunities to teach classes at a
university level.

86

Publications

Tanedo, J., Gajawelli, N., Guo, S.
,
and Baron Nelson, M.
*
Lepore, N.
* ”
Tract specific analysis in
pediatric survivors of posterior fossa brain tumors with and without adjuvant chemotherapy.”
Submitted to Frontiers in Neuroscience (December 2021)

Baron Nelson, M.; O'Neil, S. H.; Tanedo, J.; Dhanani, S.; Malvar, J.; Nuñez, C.; ... Lepore, N.
“Brain biomarkers and neuropsychological outcomes of pediatric posterior fossa brain tumor
survivors treated with surgical resection with or without adjuvant chemotherapy.”
Pediatric Blood and Cancer (2020), e28817.

Tsao, S.; Gajawelli, N.; Olch, A.; Wong, K. ;, Chapman, N.; Tanedo, J. ; Nelson, M. ; Baron
Nelson, M; Lepore, N.. "Long-term neuroanatomical effects of germ cell tumors after cranial
radiation therapy." 15th International Symposium on Medical Information Processing and
Analysis. Vol. 11330. International Society for Optics and Photonics, 2020.

Dhanani, S., O, Neil, S.; Tanedo, J.; Malvar, J.; Tamrazi, B. ; Finlay, J. ; Trane, F. ;
Rajagopalan, V. ; Lepore, N. ; Baron Nelson, M. "NCOG-72. Differential Effects of Surgery and
Chemotherapy on Children with Posterior Fossa Brain Tumors." Neuro-Oncology 22.
Supplement (2020): ii145-ii145.

Tanedo, J.; Tsao, S.; Gajawelli, N.; Lepore, N.; and Baron Nelson, M. “Treatment related DTI
changes in the posterior thalamic radiation in survivors of childhood posterior fossa tumors.”
87

14th International Symposium on Medical Information Processing and Analysis. October 24-26,
2018

Tanedo, J.; O’Neil, S.; Huft, K.; Tondulkar, A.; Wong, K.; Olch, A.; Malvar, J.; Maina, K.; Tsao,
S.; Lepore, N.; Baron Nelson, M.. "QOL-40. Radiation Dose Effect on Neuropsychological
Outcome of Posterior Fossa Tumor Survivors." Neuro-oncology 20.Suppl 2 (2018): i165.

Tanedo, J.; O’Neil, SH.; Huft, K. ; Tondulkar, A. ; Wong, K. ; Olch, A.; Malvar, J.; Maina, K.;
Tsao, S.; Lepore, N.; Baron, Nelson M. “Structural radiation dose effect on neuropsychological
outcome of posterior fossa brain tumor survivors.” International Symposium on Pediatric Neuro-
Oncology. June 29 – July 3, 2018.

Baron Nelson M.; O’Neil SH.; Tanedo J.; Huft K.; Tondulkar A.; Wong K.; Olch A; Malvar J;
Maina K.; Tsao S.; Lepore N. “Cognitive and Psychosocial Outcomes of Posterior Fossa Brain
Tumor Survivors Treated With or Without Cranial Irradiation.” Eighth International Nursing
Conference on Child and Adolescent Cancer Survivorship. April 12 - 13, 2018.

J. Tanedo, D. Sacchetto, F. Yepes, J. Coloigner, M. Descoteaux, M.D. Nelson, Jr, N. Lepore, M.
Baron Nelson. “Brain structural changes after treatment of cerebellar tumors in children”.
Presentation No. 304.13 2016 Neuroscience Meeting Planner. San Diego, CA: Society for
Neuroscience, 2016. November 16, 2016

88

References
2017 Alzheimer's disease facts and figures Alzheimer's Association. Alzheimer’s & dementia: the journal
of the Alzheimer's Association. 2017 April; [journal]
Aarsen FK, Paquier PF, Arts WF, et al. Cognitive deficits and predictors 3 years after diagnosis of a
pilocytic astrocytoma in childhood. J Clin Oncol. 2009;27(21):3626-3632.
Ahrens, James, Geveci, Berk, Law, Charles, ParaView: An End-User Tool for Large Data Visualization,
Visualization Handbook, Elsevier, 2005, ISBN-13: 978-0123875822
Ailion, Alyssa S., Kyle Hortman, and Tricia Z. King. "Childhood brain tumors: a systematic review of the
structural neuroimaging literature." Neuropsychology review 27.3 (2017): 220-244.
Aleksonis, Holly A., et al. "Associations among diffusion tensor imaging and neurocognitive function in
survivors of pediatric brain tumor: a pilot study." Applied Neuropsychology: Child (2019): 1-12.
Altieri, Roberto, et al. "Inferior Fronto-Occipital fascicle anatomy in brain tumor surgeries: From anatomy
lab to surgical theater." Journal of Clinical Neuroscience 68 (2019): 290-294.
Anderson, F. S. and Kunin-Batson, A. S., “Neurocognitive late effects of chemotherapy in children: the
past 10 years of research on brain structure and function.,” Pediatric blood & cancer 52, 159–64 (Feb
2009).
Armstrong FD. Implications of health status, health-related quality of life, and socioeconomic outcomes
in childhood brain tumor survivors. Neuro Oncol. 2019 Aug 5;21(8):957-958. doi:
10.1093/neuonc/noz082. PMID: 31111904; PMCID: PMC6682210.
Armstrong, G. T., Liu, Q., Yasui, Y., Huang, S., Ness, K. K., Leisenring, W., Hudson, M. M., Donaldson,
S. S., King, A. A., Stovall, M., Krull, K. R., Robison, L. L., and Packer, R. J., “Long-Term Outcomes
Among Adult Survivors of Childhood Central Nervous System Malignancies in the Childhood Cancer
Survivor Study,” JNCI: Journal of the National Cancer Institute 101, 946–958 (Jul 2009).
Arsigny, V., Fillard, P., Pennec, X., Ayache, N., 2006. Log-Euclidean metrics for fast and simple calculus
on diffusion tensors. Magn. Reson. Med. 56 (2), 411–421. http://dx. doi.org/10.1002/mrm.20965.
89

Assaf Y, Pasternak O. Diffusion tensor imaging (DTI)-based white matter mapping in brain research: a
review. J Mol Neurosci. 2008;34(1): 51-61.
Aukema, Eline J., et al. "White matter fractional anisotropy correlates with speed of processing and
motor speed in young childhood cancer survivors." International Journal of Radiation Oncology* Biology*
Physics 74.3 (2009): 837-843.
Avants, Brian B., Nick Tustison, and Gang Song. "Advanced normalization tools (ANTS)." Insight j 2.365
(2009): 1-35.
Bach, Michael, et al. "Methodological considerations on tract-based spatial statistics (TBSS)."
Neuroimage 100 (2014): 358-369.
Bammer R, Augustin M, Strasser-Fuchs S, et al. Magnetic resonance diffusion tensor imaging for
characterizing diffuse and focal white matter abnormalities in multiple sclerosis. Magn Reson Med.
2000;44(4):583-591.
Bammer, Roland. "Basic principles of diffusion-weighted imaging." European journal of radiology 45.3
(2003): 169-184.
Baron Nelson M, Compton P, Macey PM, et al. Diffusion tensor imaging and neurobehavioral outcome in
children with brain tumors treated with chemotherapy. J Pediatr Oncol Nurs. 2016;33(2):119-128.
Baron Nelson M, Macey PM, Harper RM, et al. Structural brain alterations in children an average of 5
years after surgery and chemotherapy for brain tumors. J Neurooncol. 2014;119(2):317-326.
Baron Nelson M, O’Neil SH, Wisnowski JL, et al. Maturation of brain microstructure and metabolism
associates with increased capacity for self-regulation during the transition from childhood to
adolescence. J Neurosci. 2019;19(42):8362-8375.
Baron Nelson, Mary C., et al. "Brain biomarkers and neuropsychological outcomes of pediatric posterior
fossa brain tumor survivors treated with surgical resection with or without adjuvant chemotherapy."
Pediatric Blood & Cancer 68.2 (2021): e28817.
Beaulieu C, Genschaft M, Huebner T, et al. Impact of chemotherapy for childhood leukemia on brain
morphology and function. PLoS One. 2013;8(11): e78599.
90

Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to
multiple testing. J Royal Stat Soc. 1995;57(1):289-300.
Bigler ED. The neuropsychology of hydrocephalus. Arch Clin Neuropsychol. 1988;3(1):81-100.
Blumenfeld H. Neuroanatomy Through Clinical Cases. Sunderland, MA: Sinauer; 2002.
Budde MD, Janes L, Gold E, Turtzo LC, Frank JA. The contribution of gliosis to diffusion tensor
anisotropy and tractography following traumatic brain injury: validation in the rat using Fourier analysis of
stained tissue sections. Brain. 2011;134(Pt 8):2248-2260.
Calandrelli, R., Panfili, M., D’Apolito, G., Zampino, G., Pedicelli, A., Pilato, F., & Colosimo, C. (2017).
Quantitative approach to the posterior cranial fossa and craniocervical junction in asymptomatic children
with achondroplasia. Neuroradiology. https://doi.org/10.1007/s00234-017-1887-y
CBTRUS. Central Brain Tumor Registry of the United States Statistical Report: primary brain and central
nervous system tumors diagnosed in the United States in 2004-2008 Hinsdale, IL: CBTRUS; 2012 [cited
2012 June 23]. Available from: www.cbtrus.org.
Cercignani M, Inglese M, Pagani E, Comi G, Filippi M. Mean diffusivity and fractional anisotropy
histograms of patients with multiple sclerosis. Am J Neuroradiol. 2001;22:952-958.
Chollet, M.B., Aldridge, K., Pangborn, N., Weinberg, S.M., DeLeon, V.B., 2014. Landmarking the brain
for geometric morphometric analysis: an error study. PLoS One 9 (1).
http://dx.doi.org/10.1371/journal.pone.0086005.
Chung, M.K., Robbins, S.M., Dalton, K.M., Davidson, R.J., Alexander, A.L., Evans, A.C., 2005. Cortical
thickness analysis in autism with heat kernel smoothing. NeuroImage 25 (4), 1256–1265.
http://dx.doi.org/10.1016/j.neuroimage.2004.12.052.
Coupé, Pierrick, et al. "Towards a unified analysis of brain maturation and aging across the entire
lifespan: An MRI analysis." Human brain mapping 38.11 (2017): 5501-5518.
Cox RW. AFNI: software for analysis and visualization of functional magnetic resonance neuroimages.
Comput Biomed Res. 1996;29:162-173.
91

Cutler, N. S., Srinivasan, S., Aaron, B. L., Anand, S. K., Kang, M. S., Altshuler, D. B., ... & Khalsa, S. S.
S. (2020). Normal cerebral ventricular volume growth in childhood. Journal of Neurosurgery: Pediatrics,
26(5), 517-524.
Davidoff, Robert A. "The pyramidal tract." Neurology 40.2 (1990): 332-332.
De Bellis, M. D., Keshavan, M. S., Shifflett, H., Iyengar, S., Beers, S. R., Hall, J., & Moritz, G. (2002).
Brain structures in pediatric maltreatment-related posttraumatic stress disorder: a sociodemographically
matched study. Biological Psychiatry, 52(11), 1066–1078. https://doi.org/10.1016/s0006-3223(02)01459-
2
Deoni SCL, Rutt BK, Peters TM (2003) Rapid combined T1 and T2 mapping using gradient recalled
acquisition in the steady state. Magn Reson Med 49(3):515–526
Deoni SCL, Rutt BK, Peters TM (2006) Synthetic T1-weighted brain image generation with incorporated
coil intensity correction using DESPOT1. Magn Reson Imaging 24(9):1241–1248
Deoni, Sean CL, et al. "Investigating white matter development in infancy and early childhood using
myelin water faction and relaxation time mapping." Neuroimage 63.3 (2012): 1038-1053.
Deprez, Sabine, et al. "Diffusion tensor MRI of chemotherapy-induced cognitive impairment in non-CNS
cancer patients: a review." Brain imaging and behavior 7.4 (2013): 409-435.
Domi, Trish, et al. "Structural integrity of the corticospinal tract correlated with the degree of hand
recovery in pediatric patients following stroke." (2012): A4033-A4033.
Dumas de la Roque A, Oppenheim C, Chassoux F, et al. Diffusion tensor imaging of partial intractable
epilepsy. Eur Radiol. 2005;15:279-285.
Dunkel IJ, Finlay JL. High-dose chemotherapy with autologous stem cell rescue for brain tumors. Crit
Rev Oncol Hematol. 2002;41(2):197–204. https://doi.org/10.1016/s1040-8428(01)00156-1
Edmiston, E. E., Wang, F., Kalmar, J. H., Womer, F. Y., Chepenik, L. G., Pittman, B., … Blumberg, H. P.
(2011). Lateral ventricle volume and psychotic features in adolescents and adults with bipolar disorder.
Psychiatry Research, 194(3), 400–402. https://doi.org/10.1016/j.pscychresns.2011.07.005
92

Fan, Y., Wang, G., Dong, Q., Liu, Y., & Lepore, N. (n.d.). Tetrahedral Spectral Feature-Based Bayesian
Manifold Learning for Grey Matter Morphometry: Findings from the Alzheimer’ s Disease Neuroimaging
Initiative. To be published in Medical Image Analysis, 2021.
Fay-McClymont TB, Ploetz DM, Mabbott DJ, et al. Long-term neuropsychological follow-up of young
children with medulloblastoma treated with sequential high-dose chemotherapy and irradiation sparing
approach. J Neurooncol. 2017;133:119-128.
Fellgiebel A, Wille P, Muller MJ, et al. Ultrastructural hippocampal and white matter alterations in mild
cognitive impairment: a diffusion tensor imaging study. Dement Geriatr Cogn Disord. 2004;18(1):101-
108.
Filley, Christopher M., and R. Douglas Fields. "White matter and cognition: making the connection."
Journal of neurophysiology 116.5 (2016): 2093-2104.
Fonov, V., Evans, A. C., Botteron, K., Almli, C. R., McKinstry, R. C., Collins, D. L., & Brain Development
Cooperative Group. (2011). Unbiased average age-appropriate atlases for pediatric
studies. Neuroimage, 54(1), 313-327.
Fouladi, Maryam, et al. "White matter lesions detected by magnetic resonance imaging after
radiotherapy and high-dose chemotherapy in children with medulloblastoma or primitive
neuroectodermal tumor." Journal of Clinical Oncology 22.22 (2004): 4551-4560.
Gajawelli, N., Deoni, S., Paquette, N., Salazar, C., Paulli, A., Dean, D., O’Muircheartaigh, J., Nelson.,
M.D., Wang, Y., Lepore, N. (2022). Surface based morphometry of the corpus callosum in young
children of ages 1-5. Unpublished
Gao, Tingran, et al. "Gaussian process landmarking for three-dimensional geometric morphometrics."
SIAM Journal on Mathematics of Data Science 1.1 (2019): 237-267.
Gaonkar B, Macyszyn L, Bilello M, Sadaghiani MS, Akbari H, Attiah MA, Ali ZS et al (2015) Automated
tumor volumetry using computer-aided image segmentation. Acad Radiol 22(5):653–661
93

Giedd, J. N., Clasen, L. S., Wallace, G. L., Lenroot, R. K., Lerch, J. P., Wells, E. M., ... & Samango-
Sprouse, C. A. (2007). XXY (Klinefelter syndrome): a pediatric quantitative brain magnetic resonance
imaging case-control study. Pediatrics, 119(1), e232-e240.
Glass, J. O., Ogg, R. J., Hyun, J. W., Harreld, J. H., Schreiber, J. E., Palmer, S. L., ... & Reddick, W. E.
(2017). Disrupted development and integrity of frontal white matter in patients treated for pediatric
medulloblastoma. Neuro-oncology, 19(10), 1408-1418.
Glass, John O., et al. "Disrupted development and integrity of frontal white matter in patients treated for
pediatric medulloblastoma." Neuro-oncology 19.10 (2017): 1408-1418.
Global Health Observatory (GHO) data. WHO.2018. Available from: http://www.who.int/gho/hiv/en/.
Gregory T. Armstrong, Neelam Jain, Wei Liu, Thomas E. Merchant, Marilyn Stovall, Deo Kumar
Srivastava, James G. Gurney, Roger J. Packer, Leslie L. Robison, Kevin R. Krull, Region-specific
radiotherapy and neuropsychological outcomes in adult survivors of childhood CNS malignancies,
Neuro-Oncology, Volume 12, Issue 11, November 2010, Pages 1173–1186,
https://doi.org/10.1093/neuonc/noq104
Grill J, Sainte-Rose C, Jouvet A, et al. Treatment of medulloblastoma with postoperative chemotherapy
alone: an SFOP prospective trial in young children. Lancet Oncol. 2005;6:573-580.
Gupta P, Jalali R. Long-term Survivors of Childhood Brain Tumors: Impact on General Health and
Quality of Life. Curr Neurol Neurosci Rep. 2017 Nov 8;17(12):99. doi: 10.1007/s11910-017-0808-0.
PMID: 29119343.
Hamilton, Richard S. "The Ricci flow on surfaces." Mathematics and general relativity, Proceedings of
the AMS-IMS-SIAM Joint Summer Research Conference in the Mathematical Sciences on Mathematics
in General Relativity, Univ. of California, Santa Cruz, California, 1986. Amer. Math. Soc., 1988.
Han, X., Xu, C., Prince, J.L., 2003. A topology preserving level set method for geometric deformable
models. IEEE Trans. Pattern Anal. Mach. Intell. 25 (6), 755–768. http://
dx.doi.org/10.1109/TPAMI.2003.1201824.
94

Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an international community of
software partners. JBiomedInform. 2019;95:103208.
Hatfield, T., Wing, D. A., Buss, C., Head, K., Muftuler, L. T., & Davis, E. (2011). Magnetic resonance
imaging demonstrates long-term changes in brain structure in children born preterm and exposed to
chorioamnionitis. American Journal of Obstetrics and Gynecology.
https://doi.org/10.1016/j.ajog.2010.10.085
Herbet, Guillaume, Sylvie Moritz-Gasser, and Hugues Duffau. "Direct evidence for the contributive role
of the right inferior fronto-occipital fasciculus in non-verbal semantic cognition." Brain Structure and
Function 222.4 (2017): 1597-1610.
Hoppe, H., 1996. Progressive meshes. In: Proceedings of the 23rd Annual Conference on Computer
Graphics and Interactive Techniques - SIGGRAPH ‘96, pp. 99–108. http://
dx.doi.org/10.1145/237170.237216.
Hotelling, Harold. "The generalization of Student’s ratio." Breakthroughs in statistics. Springer, New
York, NY, 1992. 54-65.
Iannucci G, Rovaris M, Giacomotti L, Comi G, Filippi M. Correlation of multiple sclerosis measures
derived from T2-weighted, T1-weighted, magnetization transfer, and diffusion tensor MR imaging. Am J
Neuro-radiol. 2001;22:1462-1467.
IBM SPSS Statistics for Windows, Version 25.0 [computer program]. Armonk, NY: IBM Corp; 2017.
Ikonomidou, C. Chemotherapy and the pediatric brain. Mol Cell Pediatr 5, 8 (2018).
Jackson, D. C., Irwin, W., Dabbs, K., Lin, J. J., Jones, J. E., Hsu, D. A., … Hermann, B. P. (2011).
Ventricular enlargement in new-onset pediatric epilepsies. Epilepsia. https://doi.org/10.1111/j.1528-
1167.2011.03323.x
Jacob, Julian, et al. "Cognitive impairment and morphological changes after radiation therapy in brain
tumors: a review." Radiotherapy and Oncology 128.2 (2018): 221-228.
Jacola LM, Ashford JM, Reddick WE, Glass JO, Ogg RJ, Merchant TE, Conklin HM. The relationship
between working memory and cerebral white matter volume in survivors of childhood brain tumors
95

treated with conformal radiation therapy. J Neurooncol. 2014 Aug;119(1):197-205. doi: 10.1007/s11060-
014-1476-4. Epub 2014 May 22. PMID: 24847967; PMCID: PMC4133306.
Jariyakosol S, Peragallo JH. The Effects of Primary Brain Tumors on Vision and Quality of Life in
Pediatric Patients. Semin Neurol. 2015 Oct;35(5):587-98. doi: 10.1055/s-0035-1563571. Epub 2015 Oct
6. PMID: 26444404.
Jenkinson M, Bannister P, Brady M, Smith S (2002) Improved optimization for the robust and accurate
linear registration and motion correction of brain images. Neuroimage 17(2):825–841
Jenkinson M, Smith S (2001) A global optimization method for robust affine registration of brain images.
Med Image Anal 5(2):143–156
Jesper L. R. Andersson and Stamatios N. Sotiropoulos. An integrated approach to correction for off-
resonance effects and subject movement in diffusion MR imaging. NeuroImage, 125:1063-1078, 2016.
Jong, T., Rijken, B. F. M., Lequin, M. H., Veelen, M. L. C., Mathijssen, I. M. J., de Jong, T., …
Mathijssen, I. M. J. (2011). Brain and ventricular volume in patients with syndromic and complex
craniosynostosis. Child’s Nervous System.
Katzman, D. K., Zipursky, R. B., Lambe, E. K., & Mikulis, D. J. (1997). A longitudinal magnetic
resonance imaging study of brain changes in adolescents with anorexia nervosa. Archives of Pediatrics
& Adolescent Medicine, 151(8), 793–797. https://doi.org/10.1001/archpedi.1997.02170450043006
Kesler SR, Tanaka H, Koovakkattu D. Cognitive reserve and brain volumes in pediatric acute
lymphoblastic leukemia. Brain Imaging Behav. 2010;4(3-4):256-269.
Kim JH, Brown SL, Jenrow KA, Ryu S. Mechanisms of radiation-induced brain toxicity and implications
for future clinical trials. JNeu-rooncol. 2008;87:279-286.
King, A. A., Seidel, K., Di, C., Leisenring, W. M., Perkins, S. M., Krull, K. R., Sklar, C. A., Green, D. M.,
Armstrong, G. T., Zeltzer, L. K., Wells, E., Stovall, M., Ullrich, N. J., Oeffinger, K. C., Robison, L. L., and
Packer, R. J., “Long-term neurologic health and psychosocial function of adult survivors of childhood
medulloblastoma/PNET: a report from the Childhood Cancer Survivor Study,” Neuro-Oncology 19,
now242 (Dec 2016).
96

Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and
development. Vital Health Stat 11. 2002; (246): 1–190.
Kulak, P., Maciorkowska, E., & Goscik, E. (2016). Volumetric Magnetic Resonance Imaging Study of
Brain and Cerebellum in Children with Cerebral Palsy. BioMed Research International, 2016, 5961928.
https://doi.org/10.1155/2016/5961928
Kumar R, Macey PM, Woo MA, Alger JR, Harper RM. Elevated mean diffusivity in widespread brain
regions in congenital central hypoventilation syndrome. J Magn Reson Imaging. 2006;24(6):1252-1258.
Kumra, S., Giedd, J. N., Vaituzis, A. C., Jacobsen, L. K., McKenna, K., Bedwell, J., … Rapoport, J. L.
(2000). Childhood-onset psychotic disorders: magnetic resonance imaging of volumetric differences in
brain structure. The American Journal of Psychiatry, 157(9), 1467–1474.
https://doi.org/10.1176/appi.ajp.157.9.1467
Lao, Y., Wang, Y., Shi, J., Ceschin, R., Nelson, M.D., Panigrahy, A., Leporé, N., 2014b. Thalamic
alterations in preterm neonates and their relation to ventral striatum disturbances revealed by a
combined shape and pose analysis. Brain Struct. Funct. http://dx.doi.org/10.1007/s00429-014-0921-7.
Lao, Yi, et al. "A study of brain white matter plasticity in early blinds using tract-based spatial statistics
and tract statistical analysis." Neuroreport 26.18 (2015): 1151-1154.
Lee, A. D., Leow, A. D., Lu, A., Reiss, A. L., Hall, S., Chiang, M. C., … Thompson, P. M. (2007). 3D
pattern of brain abnormalities in Fragile X syndrome visualized using tensor-based morphometry.
NeuroImage. https://doi.org/10.1016/j.neuroimage.2006.09.043
Lenroot, R. K., & Giedd, J. N. (2006). Brain development in children and adolescents: Insights from
anatomical magnetic resonance imaging. Neuroscience and Biobehavioral Reviews, 30(6), 718–729.
https://doi.org/10.1016/j.neubiorev.2006.06.001
Lepore, N., Brun, C., Chou, Y.-Y., Chiang, M.-C., Dutton, R.A., Hayashi, K.M., ... Thompson, P.M., 2008.
Multivariate statistics on deformation tensors. IEEE Trans. Med. Imaging 27 (1), 129–141.
http://dx.doi.org/10.1109/TMI.2007.906091. Generalized.
97

Lezak MD, Howieson DB, Loring DW. Neuropsychological Assessment. 4th ed. Oxford: Oxford
University Press; 2004.
Loop, C., 1987. Smooth subdivision surfaces based on triangles. ACM Siggraph Retrieved from.
http://www.citeulike.org/group/5490/article/2864922.
Lorensen, W.E., Cline, H.E., 1987. Marching cubes: a high-resolution 3D surface construction algorithm.
Proceedings of the 14th Annual Conference on Computer Graphics and Interactive Techniques -
SIGGRAPH ‘87 21 (4), 163–169. http://dx.doi. org/10.1145/37402.37422.
Lu S, Ahn D, Johnson G, Law M, Zagzag D, Grossman RI. Diffusion-tensor MR imaging of intracranial
neoplasia and associated peritumoral edema: introduction of the tumor infiltration index. Radiology.
2004;232(1):221-228.
Mabbott DJ, Noseworthy MD, Bouffet E, Rockel C, Laughlin S. Diffusion tensor imaging of white matter
after cranial radiation in children for medulloblastoma: correlation with IQ. Neuro Oncol. 2006;8(3):244-
252.
Mac Donald CL, Dikranian K, Song SK, Bayly PV, Holtzman DM, Brody DL. Detection of traumatic
axonal injury with diffusion tensor imaging in a mouse model of traumatic brain injury. Exp Neurol.
2007;205(1):116-131.
Macartney G, Harrison MB, VanDenKerkhof E, Stacey D, McCarthy P. Quality of Life and Symptoms in
Pediatric Brain Tumor Survivors: A Systematic Review. Journal of Pediatric Oncology Nursing.
2014;31(2):65-77. doi:10.1177/1043454213520191
Marachelian A, Butturini A, Finlay J. Myeloablative chemotherapy with autologous hematopoietic
progenitor cell rescue for childhood central nervous system tumors. Bone Marrow Transplant.
2008;41:167-172.
Martini, Markus, et al. "Head circumference-a useful single parameter for skull volume development in
cranial growth analysis?." Head & face medicine 14.1 (2018): 1-8.
98

Monje, Michelle L., et al. "Impaired human hippocampal neurogenesis after treatment for central nervous
system malignancies." Annals of Neurology: Official Journal of the American Neurological Association
and the Child Neurology Society 62.5 (2007): 515-520.
Mukherjee P, Miller JH, Shimony JS, et al. Diffusion-tensor MR imaging of gray and white matter
development during normal human brain maturation. Am J Neuroradiol. 2002;23:1445-1456.
Murphy, D. G., Mentis, M. J., Pietrini, P., Grady, C. L., Moore, C. J., Horwitz, B., … Rapoport, S. I.
(1999). Premutation female carriers of fragile X syndrome: a pilot study on brain anatomy and
metabolism. Journal of the American Academy of Child and Adolescent Psychiatry, 38(10), 1294–1301.
https://doi.org/10.1097/00004583-199910000-00019
Myers JS, Pierce J, Pazdernik T. Neurotoxicology of chemotherapy in relation to cytokine release, the
blood-brain barrier, and cognitive impairment. Oncol Nurs Forum. 2008;35(6):916-920.
Neuropsychological Rehabilitation: The International Handbook. United Kingdom: Routledge; 2017.
Nichols, Thomas E., and Andrew P. Holmes. "Nonparametric permutation tests for functional
neuroimaging: a primer with examples." Human brain mapping 15.1 (2002): 1-25.
O'Shea, Keiron, and Ryan Nash. "An introduction to convolutional neural networks." arXiv preprint
arXiv:1511.08458 (2015).
O’Neil SH, Whitaker AM, Kayser K, et al. Neuropsychological out-comes on Head Start III: a prospective,
multi-institutional clinical trial for young children diagnosed with malignant brain tumors. Neurooncol
Pract. 2020;7(3):329-337.
Oishi K, Faria A, Van Zijl PCM, Mori S. MRI Atlas of Human White Matter. 2nd ed. Boston, MA: Elsevier;
2011.
Olson, Ingrid R., et al. "Development of the uncinate fasciculus: Implications for theory and
developmental disorders." Developmental cognitive neuroscience 14 (2015): 50-61.
Ostrom QT, de Blank PM, Kruchko C, et al. Alex’s Lemonade Stand Foundation infant and childhood
primary brain and central nervous system tumors diagnosed in the United States in 2007-2011. Neuro
Oncol. 2015;16:x1-x36.
99

Ostrom, Quinn T., et al. "CBTRUS statistical report: Primary brain and other central nervous system
tumors diagnosed in the United States in 2013–2017." Neuro-oncology 22.Supplement_1 (2020): iv1-
iv96.
Osuka S, Matsushita A, Ishikawa E, et al. Elevated diffusion anisotropy in gray matter and the degree of
brain compression. JNeurosurg. 2012;117:363-371.
Ouahba J, Luton D, Vuillard E, Garel C, Gressens P, Blanc N, Elmaleh M, Evrard P,Oury JF. Prenatal
isolated mild ventriculomegaly: outcome in 167 cases. BJOG.2006 Sep;113(9):1072-9.
Packer RJ, Goldwein JL, Nicholson HS, et al. Treatment of children with medulloblastoma with reduced-
dose cranio-spinal radiation and adjuvant chemotherapy: a Children’s Cancer Group study. J Clin Oncol.
1999;17:2127-2136.
Palmer SL, Glass JO, Li Y, et al. White matter integrity is associated with cognitive processing in
patients treated for a posterior fossa brain tumor. Neuro Oncol. 2012;14(9):1185-1193.
Paquette, N., et al. "Ventricular shape and relative position abnormalities in preterm neonates."
NeuroImage: Clinical 15 (2017): 483-493.
Pearlson, G. D., Breiter, S. N., Aylward, E. H., Warren, A. C., Grygorcewicz, M., Frangou, S., … Pulsifer,
M. B. (1998). MRI brain changes in subjects with Down syndrome with and without dementia.
Developmental Medicine and Child Neurology, 40(5), 326–334.
Pecheva, Diliana, et al. "A tract-specific approach to assessing white matter in preterm infants."
Neuroimage 157 (2017): 675-694.
Peeler, Crandall E. "A review of visual and oculomotor outcomes in children with posterior fossa tumors."
Seminars in pediatric neurology. Vol. 24. No. 2. WB Saunders, 2017.
Peterson RK, Tabori U, Bouffet E, et al. Predictors of neuropsychological late effects and white matter
correlates in children treated for a brain tumor without radiation therapy. Pediatr Blood Cancer.
2019;66(10):e27924.
Pollack IF. Multidisciplinary management of childhood brain tumors: a review of outcomes, recent
advances, and challenges. J Neurosurg Pedi-atr. 2011;8(2):135-148.
100

Qiu, W., Yuan, J., Kishimoto, J., Ukwatta, E., & Fenster, A. (2013). Lateral ventricle segmentation of 3D
pre-term neonates US using convex optimization. Lecture Notes in Computer Science (Including
Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8151 LNCS(PART
3), 559–566. https://doi.org/10.1007/978-3-642-40760-4_70
Rajagopal, Gautham, Anand A. Joshi, and Richard M. Leahy. "An algorithm for automatic parameter
adjustment for brain extraction in BrainSuite." Medical Imaging 2017: Image Processing. Vol. 10133.
International Society for Optics and Photonics, 2017.
Rapoport, J. L., Giedd, J., Kumra, S., Jacobsen, L., Smith, A., Lee, P., … Hamburger, S. (1997).
Childhood-onset schizophrenia. Progressive ventricular change during adolescence. Archives of General
Psychiatry, 54(10), 897–903. https://doi.org/10.1001/archpsyc.1997.01830220013002
Reddick WE, Shan ZY, Glass JO, et al. Smaller white-matter volumes are associated with larger deficits
in attention and learning among long-term survivors of acute lymphoblastic leukemia. Cancer.
2006;106(4):941-949.
Reddick, Wilburn E., et al. "Atypical white matter volume development in children following craniospinal
irradiation." Neuro-oncology 7.1 (2005): 12-19.
Reiss, A. L., Aylward, E., Freund, L. S., Joshi, P. K., & Bryan, R. N. (1991). Neuroanatomy of fragile X
syndrome: the posterior fossa. Annals of Neurology, 29(1), 26–32.
https://doi.org/10.1002/ana.410290107
Remer, Justin, et al. "Quantifying cortical development in typically developing toddlers and young
children, 1–6 years of age." Neuroimage 153 (2017): 246-261.
Rey-Casserly C, Diver T. Late effects of pediatric brain tumors. Curr Opin Pediatr. 2019 Dec;31(6):789-
796. doi: 10.1097/MOP.0000000000000837. PMID: 31693589.
Reynolds, Jess E., et al. "Global and regional white matter development in early
childhood." Neuroimage 196 (2019): 49-58.
Riggs L, Bouffet E, Laughlin S, et al. Changes to memory structures in children treated for posterior
fossa tumors. J Int Neuropsychol Soc. 2014;20(2):168-180.
101

Robinson KE, Livesay KL, Campbell LK, et al. Working memory in survivors of childhood acute
lymphocytic leukemia: functional neuroimaging analyses. Pediatr Blood Cancer. 2009;54(4):585-590.
Rueckriegel SM, Driever PH, Blankenburg F, Ludemann L, Henze G, Bruhn H. Differences in
supratentorial damage of white matter in pediatric survivors of posterior fossa tumors with and without
adjuvant treatment as detected by magnetic resonance diffusion tensor imaging. Int J Radiat Oncol Biol
Phys. 2010;76(3):859-866.
Rutkowski S, Bode U, Deinlein F, et al. Treatment of early childhood medulloblastoma by postoperative
chemotherapy alone. NEnglJMed. 2005;352(10):978-986.
Sands SA, Oberg JA, Gardner SL, Whiteley JA, Glade-Bender JL, Finlay JL. Neuropsychological
functioning of children treated with intensive chemotherapy followed by myeloablative consolidation
chemotherapy and autologous hematopoietic cell rescue for newly diagnosed CNS tumors: an analysis
of the Head Start II survivors. Pediatr Blood Cancer. 2010;54:429-436.
Sands SA, Pasichow KP, Weiss R, et al. Quality of life and behavioral follow-up study of Head Start I
pediatric brain tumor survivors. JNeu-rooncol. 2011;101:287-295.
Sands SA, van Gorp WG, Finlay JL. Pilot neuropsychological findings from a treatment regimen
consisting of intensive chemotherapy and bone marrow rescue for young children with newly diagnosed
malignant brain tumors. Childs Nerv Syst. 1998;14:587-589.
Schaeffer, David J., et al. "The relationship between uncinate fasciculus white matter integrity and verbal
memory proficiency in children." Neuroreport 25.12 (2014): 921.
Schmahmann JD, Pandya DN, Wang R, et al. Association fibre pathways of the brain: parallel
observations from diffusion spectrum imaging and autoradiography. Brain. 2007;130:630-653.
Schreiber JE, Palmer SL, Conklin HM, et al. Posterior fossa syndrome and long-term neuropsychological
outcomes among children treated for medulloblastoma on a multi-institutional prospective study. Neuro
Oncol. 2017;19(12):1673-1682.
102

SEER*Explorer: An interactive website for SEER cancer statistics [Internet]. Surveillance Research
Program, National Cancer Institute. [Cited 2020 Sep 14]. Available from
https://seer.cancer.gov/explorer/.
Seunarine, Kiran K., and Daniel C. Alexander. "Multiple fibers: beyond the diffusion tensor." Diffusion
Mri. Academic Press, 2014. 105-123.
Shattuck DW, Leahy RM (2002) BrainSuite: an automated cortical surface identification tool. Med Image
Anal 6(2):129–142
Shattuck DW, Sandor-Leahy SR, Schaper KA, Rottenberg DA, and Leahy RM (2001) Magnetic
Resonance Image Tissue Classification Using a Partial Volume Model NeuroImage 13(5):856-876.
Shi, J., Stonnington, C.M., Thompson, P.M., Chen, K., Gutman, B., Reschke, C., ... Wang, Y., 2015.
Studying ventricular abnormalities in mild cognitive impairment with hyperbolic Ricci flow and tensor-
based morphometry. NeuroImage 104, 1–20. http://dx.doi.org/10.1016/j.neuroimage.2014.09.062.
Shi, J., Thompson, P.M., Gutman, B., Wang, Y., 2013a. Surface fluid registration of conformal
representation: Application to detect disease burden and genetic influence on hippocampus.
NeuroImage 78, 111–134. http://dx.doi.org/10.1016/j. neuroimage.2013.04.018.
Shi, J., Wang, Y., Ceschin, R., An, X., Lao, Y., Vanderbilt, D., ... Lepore, N., 2013b. A multivariate
surface-based analysis of the putamen in premature newborns: regional differences within the ventral
striatum. PLoS One 8 (7), e66736. http://dx.doi.org/ 10.1371/journal.pone.0066736.
Siu, A., Wind, J. J., Iorgulescu, J. B., Chan, T. A., Yamada, Y., and Sherman, J. H., “Radiation necrosis
following treatment of high-grade glioma review of the literature and current understanding,” Acta
Neurochirurgica 154, 191–201 (Feb 2012).
Sleurs, Charlotte, et al. "Advanced MR diffusion imaging and chemotherapy‐related changes in cerebral
white matter microstructure of survivors of childhood bone and soft tissue sarcoma?." Human brain
mapping 39.8 (2018): 3375-3387.
103

Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E., Watkins,
K.E., Ciccarelli, O., Cader, M.Z., Matthews, P.M., Behrens, T.E., 2006. Tract-based spatial statistics:
voxel-wise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505.
Smith, Stephen, et al. "FSL: New tools for functional and structural brain image analysis." NeuroImage
13.6 (2001): 249.
Sowell ER, Thompson PM, Leonard CM, Welcome SE, Kan E, Toga AW. Longitudinal mapping of
cortical thickness and brain growth in normal children. J Neurosci. 2004;24(38):8223-8231.
Statistical Software: Release 14 [computer program]. College Station, TX: StataCorp LP; 2015.
The American Psychiatric Publishing Textbook of Alzheimer Disease and Other Dementias. Bigler
E.USA: The American Psychiatric Publishing; 2009. Traumatic Brain Injury [book]
Tournier J. The biophysics of crossing fibers. In: Jones DK, ed. Diffusion MRI. Oxford University Press;
2010:467.
Turner CD, Rey-Casserly C, Liptak CC, Chordas C. Late Effects of Therapy for Pediatric Brain Tumor
Survivors. Journal of Child Neurology. 2009;24(11):1455-1463. doi:10.1177/0883073809341709
Turner, A. H., Greenspan, K. S., & van Erp, T. G. M. (2016). Pallidum and lateral ventricle volume
enlargement in autism spectrum disorder. Psychiatry Research. Neuroimaging, 252, 40–45.
https://doi.org/10.1016/j.pscychresns.2016.04.003
Tustison, Nicholas J., et al. "N4ITK: improved N3 bias correction." IEEE transactions on medical imaging
29.6 (2010): 1310-1320.
Udaka, Yoko T., and Roger J. Packer. "Pediatric brain tumors." Neurologic clinics 36.3 (2018): 533-556
Von Der Heide RJ, Skipper LM, Klobusicky E, Olson IR. Dissecting the uncinate fasciculus: disorders,
controversies, and a hypothesis. Brain. 2013;136(6):1692-1707.
Vovk, Uro, Franjo Pernus, and Botjan Likar. "A review of methods for correction of intensity
inhomogeneity in MRI." IEEE transactions on medical imaging 26.3 (2007): 405-421.
Wang, Y., Song, Y., Rajagopalan, P., An, T., Liu, K., Chou, Y.Y., ... The Alzheimer's Disease
Neuroimaging Initiative, 2011c. Surface-based TBM boosts power to detect disease effects on the brain:
104

an N=804 ADNI study. NeuroImage 56 (4), 1993–2010. http://
dx.doi.org/10.1016/j.neuroimage.2011.03.040.
Wang, Y., Yin, X., Zhang, J., Gu, X., Chan, T., Thompson, P. M., & Yau, S. T. (2008, October). Brain
mapping with the Ricci flow conformal parameterization and multivariate statistics on deformation
tensors. In 2nd MICCAI Workshop on Mathematical Foundations of Computational Anatomy (pp. 36-47).
Wang, Y., Zhang, J., Gutman, B., Chan, T.F., Becker, J.T., Aizenstein, H.J., ... Thompson, P.M., 2010.
Multivariate tensor-based morphometry on surfaces: Application to mapping ventricular abnormalities in
HIV/AIDS. NeuroImage 49 (1), 2141–2157. http://dx.doi.org/10.1016/j.neuroimage.2009.10.086.
Werring DJ, Toosy AT, Clark CA, et al. Diffusion tensor imaging can detect and quantify corticospinal
tract degeneration after stroke. J Neurol Neurosurg Psychiatry. 2000;69(2):269-272.
Westerhausen, René, et al. "Selective increase in posterior corpus callosum thickness between the age
of 4 and 11 years." Neuroimage 139 (2016): 17-25.
Wu, Yupeng, et al. "Subcomponents and connectivity of the inferior fronto-occipital fasciculus revealed
by diffusion spectrum imaging fiber tracking." Frontiers in neuroanatomy 10 (2016): 88.
Xenos, Chris, Spyros Sgouros, and Kalyan Natarajan. "Ventricular volume change in childhood." Journal
of neurosurgery 97.3 (2002): 584-590.
Yoon, Uicheul, et al. "The effect of template choice on morphometric analysis of pediatric brain data."
Neuroimage 45.3 (2009): 769-777.
Yuan, Q., Rubic, M., Seah, J., Rae, C., Wright, I. M., Kaltenbach, K., ... & Oei, J. L. (2014). Do maternal
opioids reduce neonatal regional brain volumes? A pilot study. Journal of Perinatology, 34(12), 909-913.
Yushkevich PA, Piven J, Hazlett HC, et al. User-guided 3D active con-tour segmentation of anatomical
structures: significantly improved efficiency and reliability. Neuroimage. 2006;31(3):1116-1128.
Yushkevich, P., Piven, J., Cody, H., Ho, S., 2006. User-guided level set segmentation of anatomical
structures with ITK-SNAP. NeuroImage 31 (3), 1116–1128.
Yushkevich, P.A., Zhang, H., Simon, T.J., Gee, .J.C., Structure-specific statistical mapping of white
matter tracts, NeuroImage. 2008 Jun; 41(2):448-61
105

Zebrack, B., Gurney, J. G., Whitton, J., Zebrack, B. J., Oeffinger, K., Packer, R. J., Mertens, A., Turk, N.,
Castleberry, R., Dreyer, Z., Robison, L. L., and Zeltzer, L. K., “Psychological outcomes in long-term
survivors of childhood brain cancer: A report from the Childhood Cancer Survivor Study Birth Defects
Survival View project Psychosocial care in Multiple Myeloma View project Psychological Outcomes in
Long-Term Survivors,” Article in Journal of Clinical Oncology 22, 999–1006 (2004).
Zeng, W., Samaras, D., Gu, X.D., 2010. Ricci flow for 3D shape analysis. IEEE Trans. Pattern Anal.
Mach. Intell. 32 (4), 662–677. http://dx.doi.org/10.1109/TPAMI.2009. 201.
Zhang, Fan, et al. "Supra-threshold fiber cluster statistics for data-driven whole brain tractography
analysis." International Conference on Medical Image Computing and Computer-Assisted Intervention.
Springer, Cham, 2017
Zhang, H., Avants, B.B, Yushkevich, P.A., Woo, J.H., Wang, S., McCluskey, L.H., Elman, L.B., Melhem,
E.R., Gee, J.C., High-dimensional spatial normalization of diffusion tensor images improves the
detection of white matter differences in amyotrophic lateral sclerosis, IEEE Transactions on Medical
Imaging, 26(11):1585-1597, November 2007. PMID: 18041273

Abstract (if available)

Abstract This thesis work is composed of Magnetic Resonance Imaging (MRI) studies on the pediatric brain in posterior fossa tumor survivors in response to treatment such as surgery, chemotherapy, and radiation as well as MRI studies on the normal developing pediatric brain. ❧ We have analyzed both the structural T1-weighted MR and Diffusion Weighted Images (DWI) of the brains of pediatric survivors of posterior fossa brain tumors. The main goal of this first project was to understand the differential effects of treatment on the neurobiology of the pediatric brain. Many studies have shown that surgical tumor resection, chemotherapy and radiation have continued detrimental effects on the developing pediatric brain beyond the tumor site. While these treatments are necessary for survival, the neuroanatomical biomarkers of neurotoxicity for each of these treatments are not well understood, especially in distinction from one another. Thus, the first part of this project has been to compare the neural structure between groups who had received increasing levels of treatment to parse out the difference in treatment effect. After an initial voxel-wise supratentorial brain FA analysis, we found several sites of FA differences in comparisons between the treatment groups and healthy controls. Some of these sites were correlated with known white matter tracts. ❧ These results informed the next study on the same cohort wherein we focused the analysis to the specific white matter tracts found in the first study. We found clusters of significantly different FA between the patient groups and the healthy controls, but none were found between the patient groups who had received differential treatment. This supports prior literature which theorizes damage to the supratentorial parts of the brain far away from the tumor site due to some effect of the tumor or treatment. This also highlighted how difficult it is to study pediatric brain tumor populations. In the process of analyzing the data, we had to exclude data due to inadequate normalization. We found much of this was due to abnormally large cerebral ventricles. ❧ The large deformations necessary to transform brain images with large ventricles create inadequate registration results. We realized that an understanding of how the presentation of ventricles develops in healthy children was needed first. In fact, after conversations with collaborating neurosurgeons, we found that there was a need for a more refined understanding of developmental variability to better understand pathological variability. Most analyses on healthy development of the brain involved either linear metrics on a single axial slice of an MR image or volumetric analyses – both of which summarize whole brain structures and tissues as a scalar value. However, there may be scenarios where indicators of pathology may lie not in volume but in finer morphometric changes that would otherwise be averaged out in gross analyses. ❧ Towards that end, we worked to characterize the development of healthy pediatric lateral ventricles by applying a multivariate tensor-based morphometry method to localize changes in shape of the ventricular system. We worked with a T1 MRI database of healthy children. Thus, we were able to identify several areas of significant shape difference between each consecutive year between ages 1 through 5. In addition, we were able to characterize the directionality of such changes by calculating the normalized medial axial distance ratio between ages and displaying those ratios on models of the ventricular surface. Finally, in an attempt to create a normative map of expected thickness variability of the ventricle, we also displayed the standard deviation of the medial axial distance values on the ventricular surface. These characterizations provide a baseline to be able to separate typical developmental variability from pathological variability. The knowledge discerned also provides neurosurgeons with another metric to evaluate dysmorphic ventricles with otherwise normal volumes.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Unveiling the white matter microstructure in 22q11.2 deletion syndrome with diffusion magnetic resonance imaging

PDF

Diffusion MRI of the human brain: signal modeling and quantitative analysis

PDF

Characterization of the brain in early childhood

PDF

Pattern detection in medical imaging: pathology specific imaging contrast, features, and statistical models

PDF

Shape, pose, and connectivity in subcortical networks across the human lifespan

PDF

A virtual reality exergaming system to enhance brain health in older adults at risk for Alzheimer’s disease

PDF

Functional connectivity analysis and network identification in the human brain

PDF

Neuroimaging markers of risk & resilience to brain aging and dementia

PDF

Neuroimaging in complex polygenic disorders

PDF

Development of a toolbox for global functional brain imaging of wake and sleep states in zebrafish

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Applications of graph theory to brain connectivity analysis

PDF

Systems pharmacology for acute myeloid leukemia and feedback control pharmacodynamic models

PDF

Ancestral inference and cancer stem cell dynamics in colorectal tumors

PDF

Correction, coregistration and connectivity analysis of multi-contrast brain MRI

PDF

Development of optical instrumentation and signal analysis for biomedical applications

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

Estimation of cognitive brain activity in sickle cell disease using functional near-infrared spectroscopy and dynamic systems modeling

PDF

3D vessel mapping techniques for retina and brain as an early imaging biomarker for small vessel diseases

PDF

Cerebrovascular disease of white matter in patients with chronic anemia syndrome

Asset Metadata

Creator Tañedo, Jeffrey Ocampo (author)

Core Title Pediatric magnetic resonance image processing: applications to posterior fossa cancer and normal development

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Degree Conferral Date 2022-05

Publication Date 02/03/2022

Defense Date 01/20/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cerebral lateral ventricles,diffusion tensor imaging,growth curves,magnetic resonance imaging,morphometric analysis,multivariate tensor based morphometry,neuroimaging,normal childhood development,OAI-PMH Harvest,pediatric brain tumor,survivors of childhood cancer,tract specific analysis,white matter

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Lepore, Natasha (committee chair), Khoo, Michael (committee member), Wood, John (committee member)

Creator Email joctanedo@gmail.com,jtanedo@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC110618981

Unique identifier UC110618981

Legacy Identifier etd-TaedoJeffr-10374

Document Type Dissertation

Format application/pdf (imt)

Rights Tañedo, Jeffrey Ocampo

Type texts

Source 20220207-usctheses-batch-911 (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. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

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