University of Southern California Dissertations and Theses

Risk factors and survival outcome in childhood alveolar soft part sarcoma among patients in the Children’s Oncology Group (COG) Phase 3 study ARST0332

(USC Thesis Other)

Risk factors and survival outcome in childhood alveolar soft part sarcoma among patients in the Children’s Oncology Group (COG) Phase 3 study ARST0332

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

Risk factors and survival outcome in childhood alveolar soft part sarcoma among patients
in the Children’s Oncology Group (COG) Phase 3 study ARST0332
By
Jiaqi Hu

A thesis presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA

In Partial Fulfillment of the Requirements for the Degree
MASTER OF SCIENCE
(APPLIED BIOSTATISTICS AND EPIDEMIOLOGY)

Dec 2018
2

Acknowledgements
My appreciation goes to my thesis advisor, Dr. Donald Barkauskas for his patient tutoring and guidance
on the data preparation and analysis. I would also like to thank my thesis committee members
Dr.Meredith Franklin and Dr. Todd Alonzo for their valuable suggestions on editing the manuscript.

3

TABLE OF CONTENTS
ACKNOWLEDGEMENTS ---------------------------------------------------------------------------------------- 2
LIST OF TABLES ---------------------------------------------------------------------------------------------------- 4
LIST OF FIGURES -------------------------------------------------------------------------------------------------- 4
ABSTRACT ------------------------------------------------------------------------------------------------------------ 5
1.INTRODUCTION--------------------------------------------------------------------------------------------------- 6
2.METHODS ---------------------------------------------------------------------------------------------------------- 8
3.RESULTS -----------------------------------------------------------------------------------------------------------10
4.DISCUSSION ----------------------------------------------------------------------------------------------------- 11
5.FIGURES AND TABLES --------------------------------------------------------------------------------------- 12
6.REFERENCES ---------------------------------------------------------------------------------------------------- 21

4

LIST OF TABLES
Table 1. Overall survival time for patients alive at last contact
Table 2. Comparison of baseline characteristics between ASPS and non-ASPS patients
Table 3. Associations between each baseline characteristics among ASPS patients (n=24)
Table 4. Pairwise comparisons between associated variables among ASPS patients (n=24)
Table 5. Log-rank test stratified by risk group to compare survival outcomes between ASPS and non-
ASPS
Table 6.1 Median Survival time of ASPS and non-ASPS in EFS.
Table 6.2 Median Survival time of ASPS and non-ASPS in OS.
Table 7.1 Log-rank test stratified by risk group for each covariate in EFS for ASPS patients.
Table 7.2 Log-rank test stratified by risk group for each covariate in OS for ASPS patients.
Table 8. Estimated predictors among patients with ASPS

LIST OF FIGURES
Figure 1A. Risk group classification system.
Figure 1B. Treatment strategy.
Figure 2.1 Kaplan-Meier Plot of Event-Free Survival (n=529).
Figure 2.2 Kaplan-Meier Plot of Overall Survival (n=529)
Figure 3.1 Kaplan-Meier Plot of Event-Free Survival among patients with ASPS by risk group
Figure 3.2 Kaplan-Meier Plot of Overall Survival among patients with ASPS by risk group

5

Abstract
Data from ARST0332, a risk-based treatment study of non-rhabdomyosarcoma soft tissue sarcomas
(NRSTS) in pediatric patients from the Children’s Oncology Group (COG), was used to explore the risk
factors and survival outcome among patients diagnosed with childhood alveolar soft part sarcoma (ASPS).
There were a total 24 of ASPS patients and 505 non-ASPS patients eligible and evaluable for analysis.
The ASPS patients were not statistically significantly different in demographics, such as age at enrollment,
sex, race and ethnicity, compared with non-ASPS patients. On the other hand, ASPS patients were more
likely to be metastatic, and classified as high-risk group consequently. Log-rank tests stratified by risk
group were used to compare the event-free survival (EFS) and overall survival (OS) between ASPS and
non-ASPS patients. Statistically significant differences between ASPS and non-ASPS were seen in EFS
but not for OS. The small number of death events among ASPS patients in certain risk stratum might
explain the unusual pattern of OS. Univariate log-rank test stratified by risk group was used to screen for
covariates to be considered for multivariate Cox-proportional hazard model. In both EFS and OS, risk
group was the only predictor in the Cox-proportional model. In EFS, high risk group compared to
intermediate risk group had a larger hazard rate. On the other hand, no statistical meaningful conclusion
was found in OS, likely due to too few events.

6

1. Introduction
1.1 Disease background
Non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) represent a heterogeneous group of neoplasms
with different biological mechanisms and clinical features, many of which are tumors typical of adulthood
[1]
. Pediatric NRSTS account for 7% of all childhood malignancies and approximately 50% of all
sarcomas in children
[2]
. Unlike the adult counterparts, pediatric NRSTS seem to be insensitive to
chemotherapy, with a complete/partial remission rate of less than 10% with conventional chemotherapy
[3]
.
On the other hand, chemotherapy and radiotherapy might also pose a greater threat to young children
compared to adults. These aspects greatly challenge the current treatment strategy for childhood NRSTS.
Alveolar soft part sarcoma (ASPS), as one of the histological subgroups of NRSTS, is a rare soft tissue
tumor that is highly vascularized with small vascular spaces separating nests of cells, and primarily
affects children. ASPS is also distinguished by its unusual patterns of metastatic spread, since the brain
has been reported to be the most common metastatic site
[4]
. Due to the rarity of childhood ASPS cases,
there is only one publication describing the disease outcomes in pediatric patients treated prospectively,
while the majority of reports concerning ASPS are case reports
[5]
. Evaluating prospective data from
ARST0332 treated with a risk-based treatment strategy will provide insights into prognostic factors of
ASPS, and further potentially improve the clinical outcomes.
1.2 Study design and treatment plan
Since childhood ASPS is rare, no standardized treatment guidelines have yet been defined. Surgery
remains as the main treatment for localized ASPS. However, surgery alone may not be sufficient, since
local recurrence or metastases could still occur many years after diagnosis, even after adequate resection
[6]
. Recent study shows that combined chemotherapy and radiotherapy after surgery might have more
potential in improving the clinical outcome. Therefore, we adopted the COG protocol “ARST0332: Risk-
based Treatment for Non-Rhabdomyosarcoma Soft Tissue Sarcomas in Patients Under 30 Years of Age.”
By comparing the characteristics between ASPS and other NRSTS, and evaluating the prognostic factors
and survival outcome, we hope to have a better understanding of the treatment effect for this fatal disease.
The study design and treatment plan are shown below
[7]
:

7

Figure 1A. Risk group classification system. Patients with non-metastatic grossly resected tumor, and
either low POG grade or high POG grade with maximal tumor diameter less than or equal to 5cm were
classified as low risk. Patients with non-metastatic and un-resected tumor, or high POG grade with
maximal diameter more than 5cm tumor were classified as intermediate risk. Patients with tumor
metastases either to lymph node or other distant sites were classified as high risk.
8

Figure 1B. Treatment strategy. Patients on treatment arms A, B or C must have had surgery done prior to
study enrollment. Low-risk patients were assigned to either treatment arm A or B, depending on the
microscopic margins. Intermediate-risk patients were assigned to arm C or D, depending on whether or
not the tumor was resected. High-risk patients were assigned to arm A, C, or D, depending on the POG
grade, resection status, and microscopic margins.
2. Methods
2.1 Data
2.1.1 Demographic information
Patient characteristics are composed of age at enrollment, gender, race and ethnicity. Age at enrollment
was categorized into three groups (0–9 years, 10–17 years, 18–30 years). Gender was treated as binary
(male, female). Race was divided into four groups (White, Black or African American, Other, and
Unknown). Ethnicity was categorized into three groups (Hispanic or Latino, non-Hispanic or non-Latino,
and unknown ethnicity).
2.1.2 Prognostic factors
Clinical prognostic factors include metastatic status (metastatic and non-metastatic), primary tumor size
(≤5cm and >5cm), primary tumor depth (superficial and deep), primary tumor invasiveness (invasive and
non-invasive), microscopic margins for patients whose tumor were resected before enrollment (negative,
9

positive and indeterminate), un-resectable tumor for patients assigned on arm D (yes and no), and
histologic tumor grade. The histologic tumor grade was classified according to Pediatric Oncology Group
(POG) Non-Rhabdomyosarcoma Soft Tissue Sarcoma Grading System, as well as grading systems of
French Federation of Cancer Centers (FNCLCC) Sarcoma Group
[8]
, during central pathology review.
Both systems were classified into two categories (low and high grade). Three records including central
imaging review, central surgery review and on study form were used to prepare the clinical data for
analysis. If the record in central imaging review was missing, then central surgery review would be used.
The On Study case report form was only used when both the central imaging review and surgery review
were missing, since the On Study case report form was reported for all patients at the individual study
sites.
2.1.3 Outcome
Event-free survival (EFS) was defined as the length of time from enrollment to an event (recurrence,
secondary malignancy, or death) occurrence or date last seen (censored). Recurrence and secondary
malignancy time were compared to determine which event occurred first. Overall survival (OS) was
defined as the length of time from enrollment to date of death (event), or date last seen (censored).
2.2 Statistical methods
Fisher’s exact tests were conducted to compare the demographic and clinical features between ASPS and
non-ASPS patients. Variables tested included age at enrollment, gender, ethnicity, tumor size, tumor
grade, tumor depth, invasiveness, resection of primary tumor, metastases, microscopic margins, assigned
treatment arms, and risk groups. Fisher’s exact tests were further employed to examine the association
between each pair of covariates, just among the ASPS patients.
Log-rank test stratified by risk group was used to compare the differences between ASPS and non-ASPS
patients, for both EFS and OS. Kaplan-Meier survival curves were constructed by each risk group (low,
intermediate or high), to compare the differences between ASPS and non-ASPS outcomes, in the event-
free and overall survival. Kaplan-Meier survival curves were further plotted by risk group among ASPS
patients, to compare the differences between each risk group, in the event-free and overall survival.
Probabilities of EFS and OS, as well as the median survival time were estimated by Kaplan-Meier
product limit method
[9]
. The 95% confidence interval for the median survival time was estimated
according to the log-log transformation.
Cox proportional hazards models were used to explore significant prognostic factors that related to the
event-free survival and overall survival. Covariates with p-value less than 0.2 in the univariate analysis
among ASPS patients based on the stratified log-rank test were taken into considerations for stepwise
selection to construct the Cox proportional hazard regression models for EFS and OS. Breslow’s method
was utilized to handle tied event times. A significance level of 0.05 to enter and to stay in the model was
used. Risk group was forced to enter the Cox proportional hazard model, while microscopic margins were
not taken into variable selection since approximately 40% of the data was indeterminate. Relative risks
were presented as hazard ratios (HR), using the intermediate risk group as reference.
10

Unless otherwise stated, a two-sided Type I error rate of 5% was employed throughout the study. SAS
software, version 9.4 (SAS institute Inc., Cary, NC, USA) was used. Procedures including PROC FREQ,
PROC LIFETEST, PROC PHREG were used in the analysis.
3. Results
3.1 Exploratory analysis of ARST0332 patients’ characteristics
3.1.1 Comparison between ASPS and non-ASPS patients
A total of 529 eligible and evaluable patients (24 ASPS and 505 non-ASPS) were analyzed in the COG
ARST0332 study. There were 18 ASPS patients and 393 non-ASPS patients alive at last contact (Table 1).
ASPS patients had shorter median follow-up time (5.75 years) compared to non-ASPS patients (6.51
years).
There were statistically significant differences in tumor POG grade, tumor metastases and risk group
between ASPS and non-ASPS patients (Table 2), while other demographics such as age at enrollment, sex,
race and ethnicity were not statistically significantly different between ASPS and non-ASPS. Specifically,
all ASPS patients were high POG grade, and had a significant higher proportion of metastatic tumors
(45.8% in ASPS compared to 13.7% in non-ASPS). Not surprisingly, ASPS patients were more likely to
be classified in high risk group compared to non-ASPS (p-value<0.0001), since tumor POG grade and
metastases were two important factors for risk group classification. For microscopic margins, around 40%
of the data were indeterminate, thus it was not included in further analysis.
3.1.2 Comparison within ASPS patients
Pairwise comparison of demographic and clinical characteristics within ASPS patients and Fisher’s exact
p-values are shown in Table 3. Due to risk group classification and treatment arm assignment, some of
the pairs such as risk group with tumor size, metastases and un-resectability, as well as treatment group
with tumor size, metastases and un-resectability, were omitted, even though the p-values were less than
0.05. Age group with tumor size (p=0.0515), race with FNCLCC grade (p=0.0568) were marginally
statistically significantly associated. There were total 10 pairs of variables with significant associations
between each other (Table 4). Older age group were more likely to have un-resectable tumors (p=0.0363).
Higher proportion of non-Hispanic had non-metastatic tumors (p=0.0411) and were more likely to be
classified in low risk group (p=0.0078). Not surprisingly, bigger tumor sizes were more likely to be
invasive, metastatic and un-resectable. Invasive tumor had greater possibility to be metastatic, and
therefore to be classified in higher risk group and assigned to treatment arm C or D. However, all of these
associations might be spurious due to the small sample size of ASPS patients.
3.2 Log-rank test stratified by risk group to compare survival outcome
3.2.1 EFS and OS between ASPS and non-ASPS
For all 529 ARST0332 patients, log-rank test stratified by risk group showed that there were statistically
significant differences in the event-free survival (p=0.0020), but not overall survival (p=0.3243) between
ASPS and non-ASPS patients (Table 5).
3.2.2 EFS and OS among ASPS patients
11

In the EFS high risk stratum, ASPS had shorter median survival time 0.49 years (95% CI 0.15–1.14),
compared to non-ASPS patients (1.03 years, 95% CI 0.85–1.63). This was also shown in Figure 2.1C, the
ASPS patients had a much lower survival than non-ASPS patients in high risk stratum (log-rank
p=0.0010). For intermediate and low risk stratum, ASPS patients had lower survival compared to non-
ASPS patients (Figures 2.1A and 2.1B). EFS comparison was significant (log-rank p=0.0058) in the
intermediate risk stratum, but not significant in the low risk stratum (log-rank p=0.5717). The median
event-free survival time for non-ASPS patients was not reachable in these two strata due to the heavy
censoring of EFS events (Table 6.1).
In the OS and high-risk stratum, interestingly, ASPS patients had longer median survival time 5.59 years
(95% CI 1.64–∞), compared to non-ASPS patients (2.37 years, 95% CI 1.48–3.36) (Table 6.2). This was
also shown in Figure 2.2C, the ASPS patients had a longer survival time than non-ASPS patients in high
risk stratum, although the comparison was not significant (log rank p=0.2656). On the other hand, for
intermediate and low risk stratum, the survival curve between ASPS and non-ASPS patients do not differ
a lot (Figure 2.2A and 2.2B). In both strata, the log-rank p-values to compare the overall survival
differences between ASPS and non-ASPS patients were less than 0.05.
3.3 Cox proportional hazard regression results
Univariate stratified log-rank tests were shown in Table 7. Other than the risk group, covariates with p-
values less than 0.2 were selected to be considered in the multivariate Cox proportional hazard regression
analysis. For EFS, age at enrollment (p=0.0733) and tumor invasiveness (p=0.1845) were included. For
OS, FNCLCC grade (p=0.1762) and treatment group (p=0.1181) were included.
The multivariate Cox-proportional hazard model for EFS showed that there were no other significant
prognostic factors associated with event-free survival, once risk group was presented in the model, but
risk group was significant (p-value=0.0025). Using the intermediate-risk group as the reference group,
low risk group had a hazard ratio of 0.06 (95% CI 0.01, 0.64), while high risk group had a hazard ratio of
3.02 (95% CI 0.53, 17.20) (Table 8).
For OS, risk group was also the only predictor (p-value=0.9229) in the multivariate Cox-proportional
hazard model. Using intermediate risk group as the reference group, the low risk group had hazard ratio
of 0.00 compared to intermediate risk group. And the 95% CI was not applicable since the death events
were extremely rare in the low risk group. Surprisingly, high risk group had a hazard ratio of 0.64 (95%
CI 0.07, 5.77) compared to intermediate risk group.
4. Discussion
In this ARST0332 study examining the risk factors and survival outcome among patients diagnosed with
childhood alveolar soft part sarcoma, 411 of a total 529 evaluable patients were alive at last contact. The
median survival time for non-ASPS patients was longer than ASPS patients. According to the exploratory
analysis of baseline characteristics, ASPS patients had higher proportion of POG grade and metastatic
tumor, thus higher proportion to be classified into higher risk group, compared to non-ASPS patients. On
the other hand, demographics such as age at enrollment, sex, race and ethnicity do not differ between
ASPS and non-ASPS patients. For both ASPS and non-ASPS, the highest proportion fell into the 10–17
years age group, which further indicated that teenagers were more likely to get non-rhabdomyosarcoma
12

soft tissue sarcomas (NRSTS). Specifically, White and non-Hispanic were more likely to experience
ASPS or non-ASPS, based on the race and ethnicity distribution. Generally speaking, ASPS patients
were more likely to have high POG grade, deep, invasive, un-resectable and metastatic tumor compared
to non-ASPS patients. Therefore, higher proportion of ASPS patients were classified into high risk group,
and further assigned combined treatment or neoadjuvant radio-chemotherapy.
Comparison within the 24 ASPS patients showed that age at enrollment might be associated with tumor
size (p=0.0515) and tumor un-resectability (p=0.0363), since older patients might have bigger tumor size
and thus had greater possibility to be un-resectable. However, there was no significant trend of age group
distribution with tumor un-resectability, probably due to the small sample size. On the other hand, non-
Hispanics were more likely to have non-metastatic tumor (p=0.0411), and all of them were classified as
low risk group. Furthermore, tumor size was associated with invasiveness, metastases and un-resectability,
since bigger tumor had a higher possibility to become metastatic, invasive and un-resectable. However,
all of these associations within ASPS patients might be ambiguous due to the small sample size.
Log-rank test stratified by risk group showed that there was significant difference of event-free survival
between ASPS and non-ASPS (p=0.0020), while the overall survival did not differ a lot (p=0.3243). For
EFS, in the high-risk group, ASPS patients had shorter median survival time compared to non-ASPS. In
the intermediate-risk group, the median survival was still comparable, and ASPS patients experienced a
shorter median survival time. In the low risk group, median survival time were not reachable for both
ASPS and non-ASPS, due to heavy censoring of EFS events. Generally, ASPS patients in the high-risk
group had the poorest survival (Figure 2.1D) in all risk stratum. For OS, surprisingly, ASPS patients had
a better survival and longer survival time compared to non-ASPS patients in both high and low risk strata,
although the differences were not statistically significant (Figure 2.2). This might due to the small sample
size of ASPS patients as well as smaller number of death events, and further led to the non-significant p-
value for overall survival between ASPS and non-ASPS in Table 5.
To build up the Cox proportional hazard model among ASPS patients, age at enrollment and tumor
invasiveness were used in the model selection process for EFS. FNCLCC grade and treatment group
were selected to construct OS. For both EFS and OS, risk group was the only predictor. For EFS, using
intermediate risk group as reference, high risk had a higher hazard rate and low risk had a lower hazard
rate (Table 8). For OS, both low- and high-risk strata had lower hazard rate compared to intermediate risk
group (Table 8), which could also be visualized in Figure 3.2. Similar to the non-significant p-values of
log-rank test of overall survival (Table 5), this unusual pattern of hazard ratio and survival probabilities
might due to the small number of patients as well death events in ASPS. On the other hand, with only 24
ASPS patients being in the multivariate Cox-regression model with a three-level covariate (risk group)
might not be feasible for hazard ratio comparison.
5. Figures and Tables
Table 1. Overall survival time for patients alive at last contact
Overall survival time for
patients alive at last contact Alive n
Median
(years)
Min
(years)
Max
(years)
ASPS (Total n=24) 18 5.75 0.86 8.55
non-ASPS (Total n=505) 393 6.51 0.01 10.98
13

All patients (Total n=529) 411 6.50 0.01 10.98

Table 2. Comparison of baseline characteristics between ASPS and non-ASPS patients
Characteristics ASPS (n=24) non-ASPS (n=505) p-value*
Age at enrollment 0.4146
0–9 years 4 (16.7%) 142 (28.1%)
10–17 years 15 (62.5%) 286 (56.6%)
≥18 years 5 (20.8%) 77 (15.3%)
Sex 0.8347
Male 12 (50.0%) 233 (46.1%)
Female 12 (50.0%) 272 (53.9%)
Race 0.1235
White 14 (58.3%) 361 (71.5%)
Black or African
American
7 (29.2%) 73 (14.4%)
Other 2 (8.3%) 22 (4.4%)
Unknown 1 (4.2%) 49 (9.7%)
Ethnicity 0.5867
Hispanic or Latino 5 (20.8%) 75 (14.9%)
Not Hispanic or Latino 19 (79.2%) 411 (81.4%)
Unknown 0 (0.0%) 19 (3.8%)
Tumor POG histologic
grade
0.0006
**

Low 0 (0.0%) 146 (28.9%)
High 24 (100.0%) 359 (71.1%)
Tumor FNCLCC
histologic grade

Low 18 (75.0%) 274 (54.3%) 0.1008
Indeterminate 0 (0.0%) 1 (0.2%)
High 6 (25.0%) 230 (45.5%)
Primary tumor size 0.3893
<=5 cm 11 (45.8%) 184 (36.4%)
>5 cm 13 (54.2%) 321 (63.6%)
Primary tumor depth 0.0982
Deep 23 (95.8%) 414 (82.0%)
Superficial 1 (4.2%) 91 (18.0%)
Invasiveness 0.5319
Invasive 12 (50.0%) 288 (57.0%)
Non-invasive 12 (50.0%) 217 (43.0%)
Treatment Arm
Assigned
0.3073
14

A: No adjuvant treatment 9 (37.5%) 196 (38.8%)
B: Adjuvant radiotherapy 2 (8.3%) 15 (3.0%)
C: Adjuvant
chemotherapy +
radiotherapy
3 (12.5%) 108 (21.4%)
D: Neoadjuvant chemo-
radiotherapy
10 (41.7%) 186 (36.8%)
Metastatic 0.0002
**

No 13 (54.2%) 436 (86.3%)
Yes 11 (45.8%) 69 (13.7%)
Un-resected tumor 1.0000
No 16 (66.7%) 325 (64.4%)
Yes 8 (33.3%) 180 (35.6%)
Microscopic margins 0.8378
Negative 8 (33.3%) 161 (31.9%)
Indeterminate 10 (41.7%) 186 (36.8%)
Positive 6 (25.0%) 158 (31.3%)
Risk group <0.0001
**

Low 11 (45.8%) 211 (41.8%)
Intermediate 2 (8.4%) 225 (44.6%)
High 11 (45.8%) 69 (13.7%)
*All p-values were calculated by exact conditional test of proportions (Fisher’s exact test) due to the
small sample size in ASPS. **p-values less than 0.05 were marked.

Table 3. Associations between each baseline characteristics among ASPS patients (n=24)
Age at
enroll-
ment
Sex Race Eth-
nicity
FNCL
CC
grade
Tumor
size
Tumor
depth
Inva-
sive-
ness
Metas-
tases
Un-
resected
tumor
Risk
group
Sex 0.3760
Race 0.4464 1.0000
Eth-
nicity
1.0000 1.0000 0.1415
FNCLC
C grade
1.0000 1.0000 0.0568 1.0000
Primary
tumor
size
0.0515 1.0000 0.3147 0.6299 0.3572
Primary
tumor
0.3750 1.0000 0.0417 0.2083 1.0000 1.0000
15

depth
Inva-
siveness
0.2649 0.6843 0.5736 1.0000 0.1550 0.0006 1.0000
Metas-
tases
0.2616 1.0000 0.2296 0.0411 0.6494 <.0001 1.0000 0.0123
Un-
resected
tumor
0.0363 1.0000 0.4379 0.6311 1.0000 0.0020 1.0000 0.1930 0.0078
Risk
group
0.1552 1.0000 0.1087 0.0078 0.7977 <.0001 0.0833 0.0011 <.0001 0.0035
Treat-
ment
group
0.1033 1.0000 0.1772 0.4262 0.6082 <.0001 0.2083 0.0010 <.0001 <.0001 <.0001
* Since all 24 ASPS patients had high POG grade tumor, POG grade was not listed in the table. Non-
informative p-values were marked in grey. Marginally significant p-values (between 0.05 and 0.10) were
marked in green. Significant p-values (≤0.05) were marked in red. P-values were included in the table for
comparison within ASPS patients, rather than estimates of association.
Table 4. Statistically significant pairwise comparisons between associated variables among ASPS patients
(n=24)
Two-way comparison pair

among ASPS (p value)
pair1 (p=0.0363) Un-resectability
Age at enrollment No Yes
0–9 years 4 (100.0 %) 0 (0.0%)
10–17 years 11 (73.3%) 4 (26.7%)
≥18 years 1 (20.00%) 4 (80.0%)
pair2 (p=0.0411) Metastases
Ethnicity No Yes
Hispanic or Latino 5 (100%) 0 (0%)
Not Hispanic or Latino 8 (42.1%) 11 (57.9%)
pair3 (p=0.0078) Risk group
Ethnicity Low Intermediate High
Hispanic or Latino 0 (0.00%) 2 (40.00%) 3 (60.00%)
Not Hispanic or Latino 11 (57.89%) 0 (0.00%) 8 (42.11%)
pair4 (p=0.0006) Invasiveness
tumor size Invasive Non-Invasive
5 cm 1 (9.09%) 10 (90.91%)
>5 cm 11 (84.62%) 2 (15.38%)
16

pair5 (p<0.0001) Metastases
tumor size No Yes
5 cm 11 (100.00%) 0 (0.00%)
>5 cm 2 (15.38%) 11 (84.62%)
pair6 (p=0.0020) Un-resectability
tumor size No Yes
5 cm 11 (100.00%) 0 (0.00%)
>5 cm 5 (38.46%) 8 (61.54%)
pair7 (p=0.0123) Metastases
Invasiveness No Yes
Invasive 3 (25.00%) 9 (75.00%)
Non-invasive 10 (83.33%) 2 (16.67%)
pair8 (p=0.0011) Risk group
Invasiveness Low Intermediate High
Invasive 9 (75.00%) 2 (16.67%) 1 (8.33%)
Non-invasive 2 (16.67%) 0 (0.00%) 10 (83.33%)
pair9 (p=0.0010) treatment group
Invasiveness
A: No adjuvant
treatment
B: Adjuvant
radiotherapy
C: Adjuvant
chemotherapy
+
radiotherapy
D:
Neoadjuvant
chemo-
radiotherapy
Invasive 1 (8.33%) 0 (0.00%) 3 (25.00%) 8 (66.67%)
Non-invasive 8 (66.67%) 2 (16.67%) 0 (0.00%) 2 (16.66%)
pair10 (p=0.0078) Un-resectability
Metastatic No Yes
No 12 (92.31%) 1 (7.69%)
Yes 4 (36.36%) 7 (63.64%)

Table 5. Log-rank test stratified by risk group to compare survival outcomes between ASPS and non-
ASPS
Chi-square DF p-value
EFS 9.5729 1 0.0020
OS 0.9715 1 0.3243

17

A. B.
C. D.
Figure 2.1 Kaplan-Meier Plot of Event-Free Survival (n=529). A: Comparison between ASPS and non-
ASPS in low risk stratum (log-rank p=0.5717). B: Comparison between ASPS and non-ASPS in
intermediate risk stratum (log-rank p=0.0058). C: Comparison between ASPS and non-ASPS in high risk
stratum (log-rank p=0.0010). D. Combined plot of event-free survival curve between ASPS and non-
ASPS.

Table 6.1 Median Survival time of ASPS and non-ASPS in EFS.
Risk group
Median EFS time for
ASPS (years)
95% Confidence
Interval for median
EFS time for ASPS
(based on log-log
transformation)
Median EFS time
for non-ASPS
(years)
95%
Confidence
Interval for
median EFS
time for non-
ASPS (based
on log-log
transformation)
High risk 0.49 0.15–1.14 1.03 0.85–1.63
Intermediate
risk
1.28 0.54–2.03 Not reached 8.96–∞
Low risk Not reached 5.15–∞ Not reached Not calculable
18

A. B.
C. D.
Figure 2.2 Kaplan-Meier Plot of Overall Survival (n=529) A: Comparison between ASPS and non-ASPS
in low risk stratum (log-rank p=0.4557). B: Comparison between ASPS and non-ASPS in intermediate
risk stratum (log-rank p=0.1883). C: Comparison between ASPS and non-ASPS in high risk stratum
(log-rank p=0.2656). D. Combined plots of overall survival curve between ASPS and non-ASPS.

Table 6.2 Median Survival time of ASPS and non-ASPS in OS.
Risk group
Median OS time for ASPS
(years)
95% Confidence
Interval for median
OS time for ASPS
(based on log-log
transformation)
Median OS time
for non-ASPS
(years)
95% Confidence
Interval for
median OS time
for non-ASPS
(based on log-
log
transformation)
High risk 5.59 1.64–∞ 2.37 1.48–3.36
Intermediate Not reached 1.22–∞ Not reached 9.87–∞
19

risk
Low risk Not reached Not calculable Not reached Not calculable

Table 7.1 Log-rank test stratified by risk group for each covariate in EFS for ASPS patients*.
EFS Covariate Stratified Log-rank test chi-square DF p-value
Age at enrollment 5.2253 2 0.0733
Gender 0.8238 1 0.3641
Race 1.7040 3 0.6360
Ethnicity 1.1788 1 0.2776
FNCLCC grade 0.1547 1 0.6941
Tumor depth 1.0000 1 0.3173
Invasiveness 1.7612 1 0.1845
Un-resected tumor 0.0420 1 0.8377
Treatment group 1.3095 2 0.5196
*Tumor size and metastases were not listed because the log-rank test statistics were not informative. Red
font indicates comparisons with p < 0.2.
Table 7.2 Log-rank test stratified by risk group for each covariate in OS for ASPS patients*.
OS Covariate Stratified Log-rank test chi-square DF p-value
Age at enrollment 2.7648 2 0.2510
Gender 0.3656 1 0.5454
Race 1.9562 2 0.3760
FNCLCC grade 1.8292 1 0.1762
Tumor depth 1.0000 1 0.3173
Invasiveness 0.1079 1 0.7426
Un-resected tumor 0.8845 1 0.3470
Treatment group 2.4423 1 0.1181
*Ethnicity, tumor size and metastases were not listed because the log-rank test statistics were not
informative. Red font indicates comparisons with p < 0.2 for consideration in the Cox proportional
hazards regression model.

Table 8. Estimated predictors among patients with ASPS
Outcome Variables Parameter
Estimate (SE)
Hazard
ratio(HR)
95% CI of
HR
EFS Risk group Low –2.86 (1.23) 0.06 0.01–0.64
Intermediate
(Reference
group)
1 — —
High 1.11 (0.89) 3.02 0.53–17.20
20

EFS
p-value
0.0025
OS Risk group Low –18.96
(3811.00)
0.00 ∞
Intermediate
(Reference
group)
1 — —
High –0.45 (1.12) 0.64 0.07–5.77
OS
p-value
0.9229

Figure 3.1 Kaplan-Meier Plot of Event-Free Survival among patients with ASPS by risk group

Figure 3.2 Kaplan-Meier Plot of Overall Survival among patients with ASPS by risk group
21

6. References
(1) Ferrari, Andrea et al. Non-Metastatic Unresected Paediatric Non-Rhabdomyosarcoma Soft Tissue
Sarcomas: Results of a Pooled Analysis from United States and European Groups. European journal of
cancer (Oxford, England : 1990) 47.5 (2011): 724–731. PMC. Web. 14 Sept. 2018.
(2) Gurney JG, Young JL, Roffers SD, Smith MA, Bunin GR. Soft Tissue Sarcomas. In: Ries LA, Smith
MA, Gurney JG, Linet M, Tamra T, Young JL, editors. Cancer incidence and survival among children
and adolescents: United States SEER Program 1975–1995, National Cancer Institute, SEER Program.
NIH Pub. No. 99-4649. Bethesda, Maryland: 1999. pp. 111–24.
(3) P. Reichardt, T. Lindner, D. Pink, P. C. Thuss-Patience, A. Kretzschmar, B. Dörken Chemotherapy in
alveolar soft part sarcomas. What do we know? Eur J Cancer. 2003 Jul; 39(11): 1511–1516.
(4) C. A. Portera, Jr, V. Ho, S. R. Patel, K. K. Hunt, B. W. Feig, P. M. Respondek, A. W. Yasko, R. S.
Benjamin, R. E. Pollock, P. W. Pisters: Alveolar soft part sarcoma: clinical course and patterns of
metastasis in 70 patients treated at a single institution, Cancer. 2001 Feb 1; 91(3): 585–591
(5) D. Orbach, B. Brennan, M. Casanova, C. Bergeron, V. Mosseri, N. Francotte, M. Van Noesel, A. Rey,
G. Bisogno, G. Pierron, et al. Paediatric and adolescent alveolar soft part sarcoma: A joint series from
European cooperative groups. Pediatr Blood Cancer. 2013 Nov; 60(11): 1826–1832. Published online
2013 Jul 16. doi: 10.1002/pbc.24683
(6) Kayton ML, Meyers P, Wexler LH, et al. Clinical presentation, treatment, and outcome of alveolar
soft part sarcoma in children, adolescents, and young adults. J Pediatr Surg 2006; 41:187–193.
(7) Sheri L. Spunt, et al. Protocol of Risk-Based Treatment for Non-Rhabdomyosarcoma Soft Tissue
Sarcomas (NRSTS) in Patients under 30 Years of Age. Children’s Oncology Group (ARST0332).
(8) Khoury, Joseph D. et al. Grading of Non-Rhabdomyosarcoma Soft Tissue Sarcoma in Children and
Adolescents: A Comparison of Parameters Used for the FNCLCC and POG Systems. Cancer 116.9
(2010): 2266–2274. PMC. Web. 14 Sept. 2018.
(9) John P. Klein, Melvin L. Moeschberger: Survival analysis, Techniques for Censored and Truncated
Data, Springer.

Abstract (if available)

Abstract Data from ARST0332, a risk-based treatment study of non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) in pediatric patients from the Children’s Oncology Group (COG), was used to explore the risk factors and survival outcome among patients diagnosed with childhood alveolar soft part sarcoma (ASPS). There were a total 24 of ASPS patients and 505 non-ASPS patients eligible and evaluable for analysis. The ASPS patients were not statistically significantly different in demographics, such as age at enrollment, sex, race and ethnicity, compared with non-ASPS patients. On the other hand, ASPS patients were more likely to be metastatic, and classified as high-risk group consequently. Log-rank tests stratified by risk group were used to compare the event-free survival (EFS) and overall survival (OS) between ASPS and non-ASPS patients. Statistically significant differences between ASPS and non-ASPS were seen in EFS but not for OS. The small number of death events among ASPS patients in certain risk stratum might explain the unusual pattern of OS. Univariate log-rank test stratified by risk group was used to screen for covariates to be considered for multivariate Cox-proportional hazard model. In both EFS and OS, risk group was the only predictor in the Cox-proportional model. In EFS, high risk group compared to intermediate risk group had a larger hazard rate. On the other hand, no statistical meaningful conclusion was found in OS, likely due to too few events.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A novel risk-based treatment strategy evaluated in pediatric head and neck non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) patients: a survival analysis from the Children's Oncology Group study...

PDF

Lymph node metastases in non-rhabdomyosarcoma soft tissue sarcoma

PDF

Survival of children and adolescents with low-risk non-rhabdomyosarcoma soft tissue sarcomas (NRSTS) treated with surgery only: an analysis of 234 patients from the Children’s Oncology Group stud...

PDF

Distant metastatic nonrhabdomyosarcoma soft tissue sarcomas in children and adolescents: clinical features and survival outcome among patients in Children's Oncology Group Phase 3 Study ARST0332

PDF

Extremity primary tumors in non-rhabdomyosarcoma soft tissue sarcoma: survival analysis

PDF

Infants in non-rhabdomyosarcoma soft tissue sarcoma

PDF

Comparative study of the POG and FNCLCC grading systems in non-rhabdomyosarcoma soft tissue sarcomas

PDF

Body wall primary tumors in non-rhabdomyosarcoma soft tissue sarcoma

PDF

An analysis of disease-free survival and overall survival in inflammatory breast cancer

PDF

An assessment of necrosis grading in childhood osteosarcoma: the effect of initial treatment on prognostic significance

PDF

Predictive factors of breast cancer survival: a population-based study

PDF

Effects of demographic and tumor characteristics on outcomes in children with malignant peripheral nerve sheath tumors

PDF

Native American ancestry among Hispanic Whites is associated with higher risk of childhood obesity: a longitudinal analysis of Children’s Health Study data

PDF

The evaluation of the long-term effectiveness of zero/low fluoroscopy workflow in ablation procedures for the treatment of paroxysmal and persistent atrial fibrillation

PDF

An assessment of impact of early local progression on subsequent risk for the treatment failure in adolescent and young adult patients with non-metastatic osteosarcoma

PDF

Air pollution and breast cancer survival in California teachers: using address histories and individual-level data

PDF

Body size and the risk of prostate cancer in the multiethnic cohort

PDF

Disparities in gallbladder, intra-hepatic bile duct, and other biliary cancers among multi-ethnic populations: a California Cancer Registry study

PDF

Pharmacogenetic association studies and the impact of population substructure in the women's interagency HIV study

Asset Metadata

Creator Hu, Jiaqi (author)

Core Title Risk factors and survival outcome in childhood alveolar soft part sarcoma among patients in the Children’s Oncology Group (COG) Phase 3 study ARST0332

School Keck School of Medicine

Degree Master of Science

Degree Program Applied Biostatistics and Epidemiology

Publication Date 11/12/2018

Defense Date 11/10/2018

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

Tag event-free survival,log-rank test,non-rhabdomyosarcoma soft tissue sarcomas,OAI-PMH Harvest,overall survival

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Barkauskas, Donald (committee chair), Alonzo, Todd A. (committee member), Franklin, Meredith (committee member)

Creator Email hujiaqi082@gmail.com,jiaqih@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-105980

Unique identifier UC11676899

Identifier etd-HuJiaqi-6956.pdf (filename),usctheses-c89-105980 (legacy record id)

Legacy Identifier etd-HuJiaqi-6956.pdf

Dmrecord 105980

Document Type Thesis

Format application/pdf (imt)

Rights Hu, Jiaqi

Type texts

Source 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 a...

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