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

Download

Open document

Flip pages

Copy asset link

Request this asset

Transcript (if available)

Content

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
by
Chang-Zheng Zhong

___________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE (Biostatistics)
December 2015
Copyright 2015 Chang-Zheng Zhong

i

Table of Contents
Table of Contents ................................................................................................................. i
List of Figures ..................................................................................................................... ii
List of Tables ..................................................................................................................... iii
Abstract ............................................................................................................................... 1
1. Introduction ..................................................................................................................... 4
1.1 Age ......................................................................................................................................... 4
1.2 Gender .................................................................................................................................... 4
1.3 Ethnicity ................................................................................................................................. 5
1.4 Site of disease ......................................................................................................................... 5
1.5 CCG-782 (high-does methotrexate study) ............................................................................. 7
1.6 INT-0133 (ifosfamide study) .................................................................................................. 8
1.7 P9754 (Plot study of DOX and IFOS intensification) ............................................................ 9
2. Material and Methods ................................................................................................... 10
2.1 Description of Study ............................................................................................................. 10
2.1.1 Participants .................................................................................................................... 10
2.1.2 Tumor site ..................................................................................................................... 13
2.1.3 Gender ........................................................................................................................... 13
2.1.4 Age ................................................................................................................................ 13
2.1.5 Race ............................................................................................................................... 15
2.1.6 Histological Response ................................................................................................... 15
2.1.7 Local Progression .......................................................................................................... 16
2.2 Statistical Methods ............................................................................................................... 16
3. Results ........................................................................................................................... 17
3.1 Patient Characteristics .......................................................................................................... 17
3.2 Event free survival (EFS) ..................................................................................................... 18
3.3 Prognostic Factors ................................................................................................................ 34
4. Discussion ..................................................................................................................... 36
5. Reference ...................................................................................................................... 37

ii

List of Figures
Figure 1. CCG-782, Induction Scheme and Maintenance Schema for Grade III to IV
Responders, 1 Cycle (repeat twice). .............................................................................. 8
Figure 2. INT-0133, Protocol road map. ........................................................................ 9
Figure 3. P9754, Treatment Plan and Modifications. .................................................. 10
Figure 4. Consort diagram of the three studies of enrollment ..................................... 12
Figure 5. Kaplan-Meier Curve ..................................................................................... 20
Figure 6. Kaplan-Meier Curve by Study of Enrollment .............................................. 21
Figure 7. Kaplan-Meier Curve by Local Progression .................................................. 23
Figure 8. Kaplan-Meier Curve by Gender ................................................................... 24
Figure 9. Kaplan-Meier Curve by Age ........................................................................ 26
Figure 10. Kaplan-Meier Curve by Race ..................................................................... 28
Figure 11. Kaplan-Meier curve by Histological Response (Huvos grading) in the CCG-
782 and INT-0133 study .............................................................................................. 30
Figure 12. Kaplan-Meier Curve by Histological Response (non-Huvos grading) in the
P9754 study .................................................................................................................. 31
Figure 13. Kaplan-Meier Curve by Histological Response (Hubos and non-Huvos
grading) ........................................................................................................................ 33

iii

List of Tables
Table 1. The total observations in the investigation .................................................... 13
Table 2. Tumor site ...................................................................................................... 13
Table 3. Patient population by Gender ........................................................................ 13
Table 4a. Age in whole year ........................................................................................ 14
Table 5. Age ................................................................................................................. 15
Table 6. Race ............................................................................................................... 15
Table 7. Histological response ..................................................................................... 16
Table 8. Local tumor progression ................................................................................ 16
Table 9. Patient Characteristics .................................................................................... 17
Table 10. Kaplan-Meier Life Table ............................................................................. 19
Table 11. Kaplan-Meier Life Table by Study of Enrollment ....................................... 20
Table 12. Log-rank test for Study of Enrollment ......................................................... 21
Table 13. Kaplan-Meier Life Table by Local Progression .......................................... 22
Table 14. Log-rank test for Local Progression ............................................................ 23
Table 15. Kaplan-Meier Life Table by Gender ........................................................... 24
Table 16. Log-rank test for Gender .............................................................................. 25
Table 17. Kaplan-Meier Life Table by Age ................................................................. 25
Table 18. Log-rank test for Age ................................................................................... 26
Table 19. Kaplan-Meier Life Table by Race ............................................................... 27
Table 20. Log-rank test for Race ................................................................................. 28
Table 21. Kaplan-Meier Life Table by Histological Response (Huvos grading) in the
CCG-782 and INT-0133 study ..................................................................................... 29
Table 22. Log-rank test for Histological Response (Huvos grading) in the CCG-782 and
INT-0133 study ............................................................................................................ 30
Table 23. Kaplan-Meier Life Table by Histological Response (non-Huvos grading) in the
P9754 study .................................................................................................................. 31
Table 24. Log-rank test for Histological Response (non-Huvos grading) in the P9754
study ............................................................................................................................. 32

iv

Table 25. Kaplan-Meier Life Table by Histological Response (Hubos and non-Huvos
grading) ........................................................................................................................ 32
Table 26. Log-rank test for Histological Response (Hubos and non-Huvos grading) . 33
Table 27. Stepwise Proportional Hazards regression .................................................. 35

1

Abstract
Purpose
Three osteosarcoma clinical trials (INT-0133, P9754, and CCG-782) from the Children’s
Oncology Group (COG) were included to identify the association between treatment
failure (metastases and relapses) and the change in primary tumor size during
preoperative chemotherapy. We are interested in the induction duration, which was
defined as the time of start preoperative chemotherapy to definitive surgery. Definitive
surgery is considered a necessary part of osteosarcoma treatment and that, occasionally,
during the induction period, a patient can experience an adverse clinical outcome. Thus,
the occurrence of new sites of disease conventionally has required a change in treatment,
but the action to be taken if the primary tumor size, as assessed by imaging radiographic,
increases is not clear.
Patients and Methods
Since the enrollment criteria for the studies varied based on site and extent of disease,
criteria was established to select patients in this investigation. As a result, there were 991
patients who were under 30 years without metastasis. sEvent free survival (EFS) was the
primary outcome measure from definitive surgery. EFS was estimated with the Kaplan-
Meier method. The EFS hazard rate was compared across groups defined by prognostic
factors using the log-rank statistic. Prognostic factor significance and associated relative
risk of various patient characteristics measured were assessed with a proportional hazards
regression model with the characteristic of interest.
Results
For patients without metastases, and early disease progression during induction duration,
the EFS was 65.8% at 5 years after the definitive surgery. The EFS for patients with the
local tumor progression during induction duration was 50.9% compared to 67.1% for
patients without local tumor progression at 5 years. The log-rank test indicated there was

2

a statistically significant association between local tumor progression and EFS. The EFS
of histological response in the CCG-782 and INT-0133 study were 58.2% for the
standard response compared to 76.9% for the good response at 5 years. The log-rank test
indicated there were statistically significantly different between the histological
responses. In the P9754 study, the EFS of histological response were 59.5% for the
standard response compared to 75.4% for the good response at 5 years. The Log-rank test
identified there was a statistically significant difference between the two histological
responses. The relative risk for patients with local progression was 1.52 (1.0-2.1) times
higher than patients who without local progression after adjustment for the study of
enrollment, histological response and the interaction term of study of enrollment and
histological response.
Conclusion
Our finding revealed that the primary tumor progression during the induction duration is
an important prognostic factor for accessing the subsequent relative risk of osteosarcoma
cancer after adjustment for other prognostic factors. Local tumor progression strengthens
the prediction of treatment failure given that the increasing size of primary tumor during
induction therapy was not necessarily associated with poor histological response to
induction chemotherapy. In summary, these finding give a supplement/additive viewpoint
to evaluate the treatment plan during the preoperative period.

4

1. Introduction
Osteosarcoma (OS) is a pleomorphic malignant tumor of bone in which the proliferating
spindle cells produce osteoid or immature bone and is the most common primary bone
malignancy (56% of bone cancers) (Jaffe et al., 2009; Meyers et al., 2005). It can arise in
any bone but is most common in the metaphyses of long bones. The most common sites
are the distal femur, the proximal tibia, and the proximal humerus. The incidence rate of
osteosarcoma and 95% confidence interval for pediatric and adolescent (aged 0-19 years)
is 5.0 (4.6-5.6) per year per million persons (Jaffe et al., 2009). Approximately 15-20% of
osteosarcoma patients will present with clinically detectable metastatic disease (Pizzo,
2010). Unfortunately, only 15-20% of patients with osteosarcoma of the extremity
without metastases after diagnosis survived five years before the introduction of
chemotherapy (Meyers, 1987; Provisor et al., 1997). The overall 5-year survival rate for
osteosarcoma was 68%, without significant differences between genders (Jaffe et al.,
2009; Ward et al., 2014). Several factors have been identified to be associated with
osteosarcoma incident rate. These are described as below.
1.1 Age
Osteosarcoma has a bimodal age at diagnosis distribution. The first peak is at 10-14 years
old corresponding to puberty and the second occurs in in individuals 65 years of age or
older. Only 2% patients with osteosarcoma are younger than 5 years. As such, the disease
occurs most frequently in children and adolescents aged 10-20 years. (Jaffe et al., 2009;
Provisor et al., 1997)
1.2 Gender
According to most recent publication by the U.S. Cancer Statistics Working Group, the
incidence rate of osteosarcoma is higher in males than females. Recently, the
Surveillance, Epidemiology, and End Results Program (SEER) reported that male
incident rate is 5.4 per million persons per year compared with 4.0 per million persons
per year for females (Jaffe et al., 2009).

5

1.3 Ethnicity
The osteosarcoma incidence rate is higher in blacks than in Hispanics and whites: blacks,
6.8 per year per million persons; Hispanics, 6.5 per year per million persons; and whites,
4.6 per year per million persons (Jaffe et al., 2009). The relative 5-year survival rate is
higher in “other races category” (66%, including American Indian/Alaska Native,
Asian/Pacific Islander, Hispanic) than in whites (60.8%) and blacks (61%) from the
SEER report (Mirabello et al., 2009).
1.4 Site of disease
Osteosarcoma occurs in any bone, but often near the metaphyseal growth plates of the
long bones of the extremities. The most commonly sites are the femur (42%, with 75% of
those tumors in the distal femur), the tibia (19%, with 80% of those tumors in the
proximal tibia), and the humerus (10%, with 90% of those tumors in the proximal
humerus). Other locations are the pelvis (8%), the skull or jaw (8%) and ribs (1.25%)
(Jaffe et al., 2009).
Prior to the introduction of systemic chemotherapy, amputation was the only surgical
treatment available for osteosarcoma in 1970s, and after surgery 80% of patients was died
from the metastatic disease, mainly of the lungs (Meyers, 1987). This result prompted
several larger randomized, prospective trials designed to evaluate the role of adjuvant
chemotherapy in the treatment of osteosarcoma in 1980s. One of these studies from the
reported a highly statistically significant improvement in disease-free survival (DSF) for
the patients treated with adjuvant chemotherapy (Link et al., 1986). At the present time,
there is consensus that adjuvant chemotherapy is effective in the treatment of
osteosarcoma. Moreover, investigators at Memorial Sloan-Kettering Cancer Center
(MSKCC) initially reported the use of preoperative chemotherapy and followed by
surgery and postoperative chemotherapy. The use of an initial period of chemotherapy
allowed an histological assessment of the response to chemotherapy at the time the
primary tumor was resected. A detailed review from MSKCC revealed that histological
response to preoperative chemotherapy predicted outcome best when a limited period of
treatment with limited agents is used (Provisor et al., 1997). When patients were treated
for a longer period with more drugs prior to surgery, the correlation between histological

6

response and disease free survival (DFS) is attenuated (Meyers et al., 1992). Several
investigators have noted that patients who have a poor response to preoperative
chemotherapy have a greater subsequent probability for metastasis and death from
osteosarcoma.
Because the amount of necrosis is the specimen resected after preoperative chemotherapy
is associated significantly with risk for adverse clinical outcome, some investigators have
designed studies that have tested increasing therapeutic intensity in patients with a ‘poor
histological response’. The implications of other clinical observations during the
prepoperative chemotherapy period (‘induction therapy’) have not been investigated to
date. Of particular interest are the implications of changes in the radiographic assessment
of tumor size during induction therapy.
The observation of tumor progression at the primary tumor site without identification of
sites of new metastatic disease occurs infrequently in non-metastatic osteosarcoma
(Provisor, et al., 1987). Those investigators demonstrated that increasing size of primary
tumor during induction therapy was not necessarily associated with poor histological
response to induction chemotherapy. Because of the uncertainty regarding whether
increasing local tumor size during induction reflects tumor that is growing despite the
administration of putatively effective chemotherapeutic agents, or a local inflammatory
process, I aggregated data from three studies of childhood osteosarcoma to address this
issue.
In this investigation, the three clinical trials (INT-0133, P9754, and CCG-782) from the
Children’s Oncology Group (COG) were included to identify the association between
treatment failure (metastases and relapses) and the change in primary tumor size during
preoperative chemotherapy. We are interested in the induction duration, which was
defined as the time of start preoperative chemotherapy to definitive surgery. Definitive
surgery is considered a necessary part of osteosarcoma treatment and that, occasionally,
during the induction period, a patient can experience an adverse clinical outcome. Thus,
the occurrence of new sites of disease conventionally has required a change in treatment,
but the action to be taken if the primary tumor size, as assessed by imaging radiographic,

7

increases is not clear.
The three trials were:
1.5 CCG-782 (high-does methotrexate study)
CCG-782 opened for enrollment in 1983. Enrollment to this study was restricted to
patients, who had newly diagnosed histologically confirmed, nonmetastatic osteosarcoma
with a primary site in an appendicular bone, and were less than 22 years of age at
diagnosis, and had not previously received any therapy. Patients with a questionable
lesion on chest x-ray or CT scan that could not be definitely called a pulmonary
metastasis were eligible. Patients with parosteal, periosteal, multicentric, and
postradiation osteosarcoma were ineligible. Chemotherapy was instituted within 4 weeks
from the date of biopsy. The investigational review boards at the individual institutions
approved the protocol.
Following biopsy, in the absence of disease progression, all patients received four courses
of vincristine (VCR) and high-dose methotrexate (MTX) with leucovorin rescue; one
course of dactinomycin (BCD) was given between the second and third MTX courses.
Surgery was performed 7 to 10 days following completion of the fourth MTX course or at
any point during induction if radiographically confirmed tumor progression occurred.
After surgery, all patients without local tumor progression received consolidation
chemotherapy, which consisted of one course of BCD, four courses of vincristine and
high-dose MTX with leucovorin rescue, and one course of DOX. Following completion
of consolidation chemotherapy, patients were assigned to one of two maintenance
regimens based on the histological response induced the primary tumor by preoperative
chemotherapy. Patients who showed greater than 95% tumor necrosis (good response)
were to receive three cycles of maintenance therapy with BCD, high-dose MTX with
leucovorin rescue, vincristine, and DOX. High-dose MTX was deleted after the first cycle
of maintenance therapy. Those patients who had < 95% tumor necrosis (poor histological
response) were to receive three cycles of maintenance therapy with DOX, CDDP, and
BCD. Patients who underwent early definitive surgery because of radiographically
suspected local tumor progression were to receive postoperative chemotherapy with BCD,

8

CDDP, and DOX (Fig 1).

Figure 1. CCG-782, Induction Scheme and Maintenance Schema for Grade III to IV
Responders, 1 Cycle (repeat twice). Vincristine (VCR), high-dose methotrexate
(HDMTX), dactinomycin (BCD), doxorubicin (DOX)
1.6 INT-0133 (ifosfamide study)
INT-0133 study was initiated at 1993 and was a randomized trial of that included a
comparison of two chemotherapy regimens for the treatment of newly diagnosed,
previously untreated osteosarcoma. One regimen was for the administration of high dose-
MTX (HDMTX), DOX, and CCDP. The other regimen was for the administration of
these agents plus ifosfamide (IFO).
Patients were under 30 years old with newly diagnosed (no greater than one month from
diagnostic biopsy) fully malignant high-grade osteosarcoma of bone of any site. Patients
with low-grade osteosarcoma, parosteal or periosteal sarcoma were ineligible. We
considered only patients with initially non-metastatic disease in this analysis.
Chemotherapy was administered for 10 weeks prior to surgical resection of the primary
tumor and any metastatic disease, for that subgroup of patients with metastatic disease
that were considered eligible. Patients also were randomly assigned at enrollment either
to receive muramyl tripeptide (MTP-PE) with and after maintenance chemotherapy or to
receive maintenance chemotherapy alone. There were two chemotherapy arms, regimens
A (CCDP, DOX, and HDMTX) and B (IFO, DOX, HDMTX, and CCDP), and within
those regimens, patients were assigned randomly to receive liposomal MTP or not. Both
arms of this study were for longer periods of more intensive preoperative chemotherapy
than utilized on CCG-782 (Fig 2).

9

Figure 2. INT-0133, Protocol road map. Doxorubicin (DOX), cisplatin (CDDP),
high-dose methotrexate (HDMTX), liposomal mu-ramyl tripeptide
phosphatidylethanolamine (L-MTP-PE), ifosfamide (IFOS)
1.7 P9754 (Plot study of DOX and IFOS intensification)
P9754 was initiated at September 1999. Patients were under 30 years old at diagnosis of
newly diagnosed, previously untreated biopsy-proven moderate or high-grade
osteosarcoma of any site without metastasis. Patients with low-grade osteosarcoma,
parosteal or periosteal sarcoma were not eligible.
Three separate regimens were piloted sequentially within the framework of P9754. All
regimens included a preoperative induction period with post induction therapy assigned
according to tumor necrosis identified in the resection specimen from the primary tumor.
Induction therapy was modeled on that of INT-0133 to facilitate comparison of response
rates.
This protocol encompassed all three pilot studies, which were activated in an order
deemed appropriate by the COG Osteosarcoma Disease Committee. Originally the pilots
were designed to open sequentially. However, due to preliminary data analysis of INT-
0133, this sequence was changed such that pilot 3 opened before pilot 2. All patients were
enrolled on pilot 1 initially. Once accrual met, pilot 3 opened. When pilot 3 met its

10

accrual goal, pilot 2 was opened for enrollment (Fig 3).

Figure 3. P9754, Treatment Plan and Modifications.
2. Material and Methods
2.1 Description of Study
2.1.1 Participants
Since the enrollment criteria for the studies varied based on site and extent of disease, I
established criteria to select patients such that individuals meeting those criteria would be
eligible for any of the three studies. As a result, I selected only patients who were under
30 years at diagnosis of newly diagnosed, previously untreated biopsy-proven moderate
or high-grade appendicular osteosarcoma without metastasis.
Initially, there were 231 subjects in the CCG-782 study. Twenty-five (25) subjects that

11

lost to follow-up or had disease progression before the induction period were excluded.
Total 206 subjects from CCG-782 were included in this investigation (Fig 4). For INT-
0133 study, there were 793 subjects at the beginning. At first, 103 subjects with
metastasis and 14 subjects that were not eligible were excluded. Then 19 subjects with
early disease progression before the definitive surgery and 47 subjects that did not
undergo definitive surgery were also excluded. Six hundred ten (610) subjects remained
for the investigation from INT-0133 after these exclusions (Fig 4). For the P9754 study,
there were 211 subjects enrolled. Seven (7) subjects with early disease progression before
definitive surgery and 29 subjects that did not undergo definitive surgery were excluded.
Finally, there were 175 subjects included in the investigation (Fig 4) from P9754 study.
In total, 991 subjects were considered qualified in this investigation (Table 1).

CCG-782
N=231
Lost to Follow-up or had disease
progression (metastasis) before
induction period
N=25
CCG-782
N=206

12

Figure 4. Consort diagram of the three studies of enrollment

INT-0133
N=793
Metastatics
N=103
INT-0133
N=690
Not eligible
N=14
Disease progression
(metastasis) before
definitive surgery
N=19
INT-0133
N=657
INT-0133
N=676
Not receiving
definitive surgery
N=47
INT-0133
N=610
P9754
N=211
Disease progrssion
(metastasis) before
definitive surgery
N=7
P9754
N=204
Not receiving
definitive surgery
N=29
P9754
N=175

13

Table 1. The total observations in the investigation
| Freq. Percent Cum.
------------+-----------------------------------
CCG-782 | 206 20.79 20.79
INT-0133 | 610 61.55 82.34
P9754 | 175 17.66 100.00
------------+-----------------------------------
Total | 991 100.00

2.1.2 Tumor site
Primary tumor sites were categorized into three main subgroups, i.e. leg, arm, and others
(other sites and missing) among 991 subjects. There were 688 (69.42%) tumors were on
the leg, 211 (21.29%) on the arm, and 92 (9.29%) tumors were on the other sites (others).
Table 2. Tumor site
| P9754 INT-0133 CCG-782 | Total
-----------+---------------------------------+----------
Leg | 109 402 177 | 688
Arm | 13 169 29 | 211
Others | 53 39 0 | 92
-----------+---------------------------------+----------
Total | 175 610 206 | 991

2.1.3 Gender
There were 98 (56.00%) males and 77 (44.00) females in the P9754 study, 338 (55.41%)
males and 272 (44.59%) females in the INT-0133, and 110(53.40%) males and 96
(46.60%) females in CCG-782 from total 546 (55.10%) males and 445 (44.90%) females
in this investigation (Table 3).
Table 3. Patient population by Gender
| P9754 INT-0133 CCG-782 | Total
-----------+---------------------------------+----------
Male | 98 338 110 | 546
Female | 77 272 96 | 445
-----------+---------------------------------+----------
Total | 175 610 206 | 991

2.1.4 Age
Age at enrollment was computed in whole years (Table 4 a.b.). The average age was
13.23±4.20 (range: 1-30 whole year). Age was categorized into two groups: equal or less

14

than 16 years old and more than 16 years old (Table 5). The cut point was chosen
according to bimodal distribution of age at diagnosis of osteosarcoma. 809 (81.63%)
subjects were equal or under the 16 years old and 182 (18.37%) were above 16 years old.
Table 4a. Age in whole year
| P9754 INT-0133 CCG-782 | Total
-----------+---------------------------------+----------
1 | 0 1 0 | 1
3 | 1 2 1 | 4
4 | 4 3 2 | 9
5 | 3 10 3 | 16
6 | 4 10 2 | 16
7 | 5 28 2 | 35
8 | 11 24 12 | 47
9 | 2 26 11 | 39
10 | 15 50 14 | 79
11 | 10 47 20 | 77
12 | 18 69 13 | 100
13 | 18 57 30 | 105
14 | 19 61 18 | 98
15 | 18 59 22 | 99
16 | 12 57 15 | 84
17 | 10 36 19 | 65
18 | 4 22 8 | 34
19 | 9 15 9 | 33
20 | 3 7 2 | 12
21 | 1 3 3 | 7
22 | 1 5 0 | 6
23 | 4 3 0 | 7
24 | 0 5 0 | 5
26 | 0 2 0 | 2
27 | 1 1 0 | 2
28 | 1 1 0 | 2
29 | 0 4 0 | 4
30 | 1 2 0 | 3
-----------+---------------------------------+----------

Table 4b. Age in whole year
Age in Whole Year
-------------------------------------------------------------
Percentiles Smallest
1% 4 1
5% 7 3
10% 8 3
25% 11 3

50% 13
Largest
75% 16 29 Obs 991
90% 18 30 Mean 13.23108
95% 20 30 Std. Dev. 4.196208
99% 27 30 Variance 17.60816

15

Table 5. Age
| Freq. Percent Cum.
---------------------+-----------------------------------
≤16 Age Whole Year | 809 81.63 81.63
>16 Age Whole Year | 182 18.37 100.00
---------------------+-----------------------------------
Total | 991 100.00

2.1.5 Race
Race was categorized into three groups: white, Africa American, and others according the
five-year survival rate difference (Table 6). Six hundred eighty-one (681) (68.72%)
subjects were white, 127 (12.82%) were Africa America, and others were 183 (18.47%)
subjects.
Table 6. Race
| P9754 INT-0133 CCG-782 | Total
----------------------+---------------------------------+----------
White | 119 414 148 | 681
African American | 29 83 15 | 127
Others | 27 113 43 | 183
----------------------+---------------------------------+----------
Total | 175 610 206 | 991

2.1.6 Histological Response
The tumor specimen from patients who underwent definitive resection after induction
underwent evaluation of necrosis by an institutional pathologist. Histological response
was graded according to the method of Huvos (Meyers et al., 1992; Provisor et al., 1997)
in INT-0133 and CCG-782. Histological response was categorized into two groups: I-II
grading (standard response) and III-IV grading (good response).
For the P9754 study, a different grading method was applied to the histological response
rather than Huvos’ method. The good response was the complete or near complete cell
necrosis, with <2% viable tumor cell residual in the total tumor volume and standard
response was with >2% viable tumor cell residual in the total tumor volume. Thus, we
treated the histological response as a categorical variable, which included 4 groups,
Huvos I/II grading, Huvos III/IV grading, non-Huvos I/II grading, and non-Huvos III/IV
grading (Table 7).

16

Table 7. Histological response
| P9754 INT-0133 CCG-782 | Total
-----------------+---------------------------------+----------
Huvos I/II | 0 335 148 | 483
Huvos III/IV | 0 275 58 | 333
non-Huvos I/II | 113 0 0 | 113
non-Huvos III/IV | 62 0 0 | 62
-----------------+---------------------------------+----------
Total | 175 610 206 | 991

2.1.7 Local Progression
Increasing size of primary tumor during induction therapy was defined as local tumor
progression. The dichotomous local progression was summarized (Table 8).
Table 8. Local tumor progression
| P9754 INT-0133 CCG-782 | Total
-----------+---------------------------------+----------
No | 129 606 170 | 905
Yes | 46 4 36 | 86
-----------+---------------------------------+----------
Total | 175 610 206 | 991
2.2 Statistical Methods
The primary goals of the study were to be addressed in patients without detectable
metastases at diagnosis. We planned to assess relative risks associated with local
progression and others prognostic factors. Patients with local tumor progression would be
compared with patients with no local tumor progression after adjustment for other
prognostic factors.
Induction duration was defined as above that from the enrollment to the definitive surgery.
Event free survival (EFS) was the primary outcome measure. EFS was defined as the
time from definitive surgery until an adverse event or last patient contact, whichever
came first. Adverse events included disease progression, diagnosis of second malignant
neoplasm, or death before disease progression or second malignant neoplasm (SMN).
Patients without adverse events were censored at date of last contact.
EFS was estimated with the Kaplan-Meier method (Kalbfleisch and Prentice, 1980). The
EFS hazard rate was compared across groups defined by prognostic factors using the log-
rank statistic (Kalbfleisch and Prentice, 1980). When comparing treatment assignment, I

17

controlled the family-wise error rate by Scheffe procedure (Scheffé, 1959)
Prognostic significance and associated relative risk of various patient characteristics
measured at entry were assessed with a proportional hazards regression model with the
characteristic of interest as the only component (Kalbfleisch and Prentice, 1980). We
identified the characteristics to be included for the backwards-stepwise proportional
hazards regression at the 0.05 level or less significant level. Interactions between study
enrollment and prognostic factors were examined. We fitted all models using an
hierarchical approach, so that inclusion of a higher order term requires inclusion of all
lower order terms once the interaction was statistically significant. Confidence intervals
for relative risks were derived from the proportional hazards regression model.
3. Results
3.1 Patient Characteristics
Among 991 patients in this investigation from the three studies, 546 (55.10%) were male
and 445 (44.90%) were female (Table 9). Most of them were white with 681 (68.72%)
subjects, 127 (12.82%) were Africa American and 183 (18.47%) were other race. Age
whole year at enrollment >16 were 35 (18.37%) subjects and ≤16 were 809 (81.63%)
subjects.
Table 9. Patient Characteristics
| P9754 INT-0133 CCG-782 | Total | %
-------------------+---------------------------------+-----------+---------+
Tumor sites | | | |
Leg | 109 402 177 | 688 | 69.42|
Arm | 13 169 29 | 211 | 21.29|
Other | 53 39 0 | 92 | 9.28|
-------------------+---------------------------------+-----------+---------+
Sex | | | |
Male | 98 338 110 | 546 | 55.10|
Female | 77 272 96 | 445 | 44.90|
-------------------+---------------------------------+-----------+---------+
White | 119 414 148 | 681 | 68.72|
African American | 29 83 15 | 127 | 12.82|
Others | 27 113 43 | 183 | 18.47|
-------------------+---------------------------------+-----------+---------+
Age | | | |
=<16 Whole Year | 140 504 165 | 809 | 81.63|
>16 Whole Year | 35 106 41 | 182 | 18.37|
-------------------+---------------------------------+-----------+---------+
Histologic response| | | |

18

Huvos I/II | 0 335 148 | 483 | 48.74|
Huvos III/IV | 0 275 58 | 333 | 33.60|
non-Huvos I/II | 113 0 0 | 113 | 11.40|
non-Huvos III/IV | 62 0 0 | 62 | 6.26|
-------------------+---------------------------------+-----------+---------+
Local progression | | | |
No | 129 606 170 | 905 | 91.32|
Yes | 46 4 36 | 86 | 8.68|
-------------------+---------------------------------+-----------+---------+

| Histologic Response
Local Progression | III/IV I/II | Total
------------------+----------------------+----------
No | 380 525 | 905
Yes | 15 71 | 86
------------------+----------------------+----------
Total | 395 596 | 991
3.2 Event free survival (EFS)
For patients without metastases, and early disease progression during induction duration,
the EFS was 70.4% at 3 years and 65.8% at 5 years after the definitive surgery (Table 10
and Fig 5). The EFS stratified by the study of enrollment were 60.6%, 67.7%, and 65.2%
at 5 years for CCG-782, INT-0133 and P9754 respectively. As the result of log-rank test
by the study of enrollment, there was statistically significant different among the three
EFS (p=0.02) (Table 11 and Fig 6). The multiply comparisons adjusted for family-wise
error rate revealed that there was a statistically significant difference of EFS between the
study of CCG-782 and INT-0133 (p=0.01) (Table 12).
The EFS for patients with the local tumor progression during induction duration was
50.9% compared to 67.1% for patients without local tumor progression at 5 years (Table
13 and Fig 7). The log-rank test indicated there was a statistically significant association
between local tumor progression and EFS (p< 0.001) (Table 14). EFS for male and
female patients were 66.0% and 65.5% respectively at 5 years (Table 15 and Fig 8). Log-
rank test indicated EFS by sex was not statistically significantly different (p=0.80) (Table
16). The EFS for patients who were age ≤16 was 66.8% compared to 61.2% of patient
who were age >16 at 5 year (Table 17 and Fig 9). The log-rank result showed that there
was a statistically significantly difference between the two age groups (p=0.04) (Table
18). EFS among race groups were 65.2%, 72.0%, and 63.4% for white, Africa American,
and others respectively at 5 years (Table 19 and Fig 10). Log-rank test showed there were
not statistically significantly different among race (p=0.46) (Table 20). The EFS of
histological response (Huovs grading) in the CCG-782 and INT-0133 study were 58.2%

19

for the standard response compared to 76.9% for the good response at 5 years (Table 21
and Fig 11). The log-rank test indicated there were statistically significantly different
between the histological responses (p<0.0001) (Table 22). In the P9754 study, the EFS of
histological response (non-Huvos grading) were 59.5% for the standard response
compared to 75.4% for the good response at 5 years (Table 23 and Fig 12). The Log-rank
test identified there was a statistically significant difference between the two histological
responses (p=0.04) (Table 24). The EFS for the four histological responses were
summarized (Table 25 and Fig 13). The Log-rank test indicated that there were
statistically significantly different among the four histological grades. For more details,
Huvos I/II was statistically significantly different from Huvos III/IV and non-Huvos
III/IV (p< 0.0001 and 0.01 respectively). Non-Huvos I/II was statistically significantly
different from Huvos III/IV and non-Huvos III/IV (p< 0.001 and 0.04 respectively)
(Table 26).
Table 10. Kaplan-Meier Life Table
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
1 863 114 0.8839 0.0102 0.8621 0.9024
2 728 119 0.7610 0.0137 0.7330 0.7866
3 654 54 0.7039 0.0147 0.6740 0.7316
4 579 33 0.6673 0.0152 0.6364 0.6962
5 507 8 0.6577 0.0154 0.6266 0.6869
6 419 8 0.6463 0.0157 0.6147 0.6760
7 317 4 0.6392 0.0159 0.6072 0.6694
8 210 2 0.6345 0.0161 0.6020 0.6652
9 122 7 0.6067 0.0186 0.5692 0.6421
10 51 0 0.6067 0.0186 0.5692 0.6421
-------------------------------------------------------------------------------

20

Figure 5. Kaplan-Meier Curve

Table 11. Kaplan-Meier Life Table by Study of Enrollment
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
CCG-782
1 174 28 0.8624 0.0242 0.8069 0.9029
2 137 32 0.7011 0.0324 0.6324 0.7594
3 120 11 0.6438 0.0340 0.5728 0.7061
4 109 7 0.6056 0.0349 0.5335 0.6701
5 97 0 0.6056 0.0349 0.5335 0.6701
6 78 3 0.5854 0.0357 0.5121 0.6516
7 57 2 0.5693 0.0365 0.4946 0.6371
8 45 1 0.5569 0.0377 0.4798 0.6271
9 33 4 0.5050 0.0422 0.4196 0.5842
10 13 0 0.5050 0.0422 0.4196 0.5842
INT-0133
1 536 67 0.8892 0.0128 0.8613 0.9117
2 465 64 0.7822 0.0168 0.7470 0.8131
3 419 33 0.7258 0.0183 0.6882 0.7598
4 366 22 0.6862 0.0191 0.6470 0.7220
5 322 5 0.6765 0.0193 0.6370 0.7128
6 272 5 0.6649 0.0197 0.6247 0.7019
7 229 2 0.6593 0.0199 0.6186 0.6967
8 165 1 0.6564 0.0200 0.6155 0.6940
9 90 3 0.6369 0.0224 0.5912 0.6789
10 39 0 0.6369 0.0224 0.5912 0.6789
P9754
1 155 19 0.8908 0.0236 0.8342 0.9289
2 128 23 0.7569 0.0327 0.6857 0.8142
3 117 10 0.6971 0.0351 0.6222 0.7600
4 106 4 0.6723 0.0360 0.5962 0.7372
5 90 3 0.6522 0.0368 0.5750 0.7189
0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
Event-Free Survival From Definitive Surgery

21

6 72 0 0.6522 0.0368 0.5750 0.7189
7 33 0 0.6522 0.0368 0.5750 0.7189
8 2 0 0.6522 0.0368 0.5750 0.7189
--------------------------------------------------------------------------------

Figure 6. Kaplan-Meier Curve by Study of Enrollment

Table 12. Log-rank test for Study of Enrollment
| Events Events
| observed expected
---------+-------------------------
P9754 | 59 60.16
INT-0133 | 202 220.85
CCG-782 | 88 67.99
---------+-------------------------
Total | 349 349.00

chi2(2) = 7.53
Pr>chi2 = 0.0232

| Events Events
| observed expected
---------+-------------------------
P9754 | 59 56.41
INT-0133 | 202 204.59
---------+-------------------------
Total | 261 261.00

chi2(1) = 0.15
0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
P9754
INT-0133
CCG-782
Study of Enrollment
By Study of Enrollment
Event-Free Survival From Definitive Surgery

22

Pr>chi2 = 0.6958
| Events Events
| observed expected
--------+-------------------------
P9754 | 59 67.81
CCG-782 | 88 79.19
--------+-------------------------
Total | 147 147.00

chi2(1) = 2.18
Pr>chi2 = 0.1397

| Events Events
| observed expected
---------+-------------------------
INT-0133 | 202 221.76
CCG-782 | 88 68.24
---------+-------------------------
Total | 290 290.00

chi2(1) = 7.49
Pr>chi2 = 0.0062

Event | P9754 INT-0133 CCG-782 | Total
-----------+---------------------------------+----------
No | 116 406 118 | 640
Yes | 59 202 88 | 349
-----------+---------------------------------+----------
Total | 175 608 206 | 989

Table 13. Kaplan-Meier Life Table by Local Progression
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
No
1 800 95 0.8943 0.0103 0.8723 0.9127
2 680 105 0.7760 0.0140 0.7472 0.8020
3 610 51 0.7171 0.0151 0.6862 0.7456
4 538 30 0.6806 0.0158 0.6485 0.7104
5 469 7 0.6713 0.0159 0.6390 0.7015
6 385 8 0.6586 0.0163 0.6257 0.6894
7 300 4 0.6510 0.0165 0.6175 0.6823
8 202 2 0.6460 0.0168 0.6121 0.6778
9 116 6 0.6205 0.0192 0.5817 0.6568
10 50 0 0.6205 0.0192 0.5817 0.6568
Yes
1 64 19 0.7697 0.0464 0.6629 0.8464
2 49 14 0.5971 0.0543 0.4826 0.6942
3 45 3 0.5598 0.0550 0.4454 0.6594
4 42 3 0.5216 0.0555 0.4079 0.6235
5 39 1 0.5089 0.0556 0.3956 0.6113
6 35 0 0.5089 0.0556 0.3956 0.6113
7 18 0 0.5089 0.0556 0.3956 0.6113
8 9 0 0.5089 0.0556 0.3956 0.6113
9 7 1 0.4453 0.0768 0.2926 0.5871
10 2 0 0.4453 0.0768 0.2926 0.5871
-------------------------------------------------------------------------------

23

Figure 7. Kaplan-Meier Curve by Local Progression

Table 14. Log-rank test for Local Progression
| Events Events
| observed expected
-------------+-------------------------
No | 308 324.25
Yes | 41 24.75
-------------+-------------------------
Total | 349 349.00

chi2(1) = 11.49
Pr>chi2 = 0.0007

| Local Progression
Event | No Yes | Total
-----------+----------------------+----------
No | 597 43 | 640
Yes | 308 41 | 349
-----------+----------------------+----------
Total | 905 84 | 989

0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
No
Yes
Local Progression
By Local Progression
Event-Free Survival From Definitive Surgery

24

Table 15. Kaplan-Meier Life Table by Gender
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
Male
1 474 61 0.8866 0.0137 0.8567 0.9106
2 399 63 0.7677 0.0183 0.7295 0.8013
3 357 31 0.7071 0.0198 0.6662 0.7440
4 320 18 0.6706 0.0206 0.6285 0.7091
5 279 5 0.6598 0.0208 0.6173 0.6988
6 225 5 0.6468 0.0212 0.6036 0.6866
7 171 3 0.6366 0.0217 0.5925 0.6774
8 112 2 0.6282 0.0222 0.5830 0.6700
9 63 5 0.5888 0.0271 0.5337 0.6397
10 30 0 0.5888 0.0271 0.5337 0.6397
Female
1 390 53 0.8805 0.0154 0.8466 0.9074
2 330 56 0.7530 0.0205 0.7100 0.7906
3 298 23 0.7000 0.0219 0.6548 0.7405
4 260 15 0.6632 0.0227 0.6166 0.7055
5 229 3 0.6551 0.0229 0.6082 0.6979
6 195 3 0.6457 0.0232 0.5982 0.6891
7 147 1 0.6423 0.0233 0.5945 0.6859
8 99 0 0.6423 0.0233 0.5945 0.6859
9 60 2 0.6273 0.0251 0.5760 0.6743
10 22 0 0.6273 0.0251 0.5760 0.6743
-------------------------------------------------------------------------------

Figure 8. Kaplan-Meier Curve by Gender

0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
Male
Female
Sex
By Sex
Event-Free Survival From Definitive Surgery

25

Table 16. Log-rank test for Gender
| Events Events
| observed expected
-------+-------------------------
Male | 193 190.73
Female | 156 158.27
-------+-------------------------
Total | 349 349.00

chi2(1) = 0.06
Pr>chi2 = 0.8071

Event | Male Female | Total
-----------+----------------------+----------
No | 351 289 | 640
Yes | 193 156 | 349
-----------+----------------------+----------
Total | 544 445 | 989

Table 17. Kaplan-Meier Life Table by Age
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
<= 16 Age Whole Year
1 712 88 0.8903 0.0110 0.8666 0.9101
2 607 95 0.7706 0.0149 0.7398 0.7983
3 543 48 0.7089 0.0161 0.6759 0.7392
4 482 25 0.6750 0.0167 0.6410 0.7066
5 426 5 0.6678 0.0169 0.6335 0.6996
6 361 5 0.6593 0.0171 0.6246 0.6915
7 277 3 0.6533 0.0173 0.6183 0.6860
8 185 1 0.6502 0.0175 0.6148 0.6832
9 106 5 0.6269 0.0198 0.5867 0.6643
10 44 0 0.6269 0.0198 0.5867 0.6643
> 16 Age Whole Year
1 152 26 0.8549 0.0263 0.7942 0.8988
2 122 24 0.7182 0.0338 0.6457 0.7785
3 112 6 0.6822 0.0352 0.6076 0.7455
4 98 8 0.6325 0.0367 0.5558 0.6996
5 82 3 0.6120 0.0374 0.5343 0.6806
6 59 3 0.5850 0.0389 0.5047 0.6568
7 41 1 0.5714 0.0403 0.4884 0.6459
8 26 1 0.5563 0.0420 0.4702 0.6340
9 17 2 0.5044 0.0520 0.3987 0.6008
10 8 0 0.5044 0.0520 0.3987 0.6008
-------------------------------------------------------------------------------

26

Figure 9. Kaplan-Meier Curve by Age

Table 18. Log-rank test for Age
| Events Events
| observed expected
---------------------+-------------------------
<= 16 Age Whole Year | 275 289.48
> 16 Age Whole Year | 74 59.52
---------------------+-------------------------
Total | 349 349.00

chi2(1) = 4.25
Pr>chi2 = 0.0393

Event | <=16 Age >16 Age | Total
-----------+----------------------+----------
No | 532 108 | 640
Yes | 275 74 | 349
-----------+----------------------+----------
Total | 807 182 | 989

0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
<= 16 Age Whole Year
> 16 Age Whole Year
Age
By Age
Event-Free Survival From Definitive Surgery

27

Table 19. Kaplan-Meier Life Table by Race
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
White
1 598 74 0.8902 0.0120 0.8640 0.9115
2 508 82 0.7674 0.0163 0.7335 0.7976
3 455 44 0.7003 0.0178 0.6639 0.7335
4 409 24 0.6624 0.0184 0.6250 0.6971
5 370 6 0.6525 0.0186 0.6148 0.6875
6 322 3 0.6468 0.0187 0.6089 0.6821
7 243 4 0.6377 0.0190 0.5991 0.6735
8 159 2 0.6315 0.0193 0.5923 0.6680
9 96 3 0.6165 0.0208 0.5744 0.6557
10 40 0 0.6165 0.0208 0.5744 0.6557
African American
1 113 15 0.8819 0.0286 0.8118 0.9270
2 98 13 0.7791 0.0369 0.6963 0.8418
3 89 5 0.7382 0.0392 0.6520 0.8062
4 77 2 0.7197 0.0403 0.6317 0.7902
5 64 0 0.7197 0.0403 0.6317 0.7902
6 46 3 0.6822 0.0437 0.5881 0.7592
7 37 0 0.6822 0.0437 0.5881 0.7592
8 26 0 0.6822 0.0437 0.5881 0.7592
9 16 2 0.6115 0.0616 0.4796 0.7195
10 7 0 0.6115 0.0616 0.4796 0.7195
Others
1 154 25 0.8620 0.0256 0.8027 0.9046
2 125 24 0.7244 0.0336 0.6523 0.7841
3 112 5 0.6946 0.0348 0.6206 0.7569
4 95 7 0.6498 0.0364 0.5734 0.7160
5 75 2 0.6343 0.0372 0.5566 0.7020
6 53 2 0.6110 0.0393 0.5292 0.6828
7 39 0 0.6110 0.0393 0.5292 0.6828
8 27 0 0.6110 0.0393 0.5292 0.6828
9 12 2 0.5455 0.0569 0.4278 0.6488
10 6 0 0.5455 0.0569 0.4278 0.6488
-------------------------------------------------------------------------------

28

Figure 10. Kaplan-Meier Curve by Race
Table 20. Log-rank test for Race
| Events Events
| observed expected
--------------------------+-------------------------
White | 242 243.81
African American | 40 45.45
Others | 67 59.75
--------------------------+-------------------------
Total | 349 349.00

chi2(2) = 1.55
Pr>chi2 = 0.4611

| Events Events
| observed expected
--------------------------+-------------------------
White | 242 237.68
African American | 40 44.32
--------------------------+-------------------------
Total | 282 282.00

chi2(1) = 0.50
Pr>chi2 = 0.4798

| Events Events
| observed expected
---------+-------------------------
White | 242 248.07
Others | 67 60.93
0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
White
Black or African American
Others
Race
By Race
Event-Free Survival From Definitive Surgery

29

---------+-------------------------
Total | 309 309.00

chi2(1) = 0.75
Pr>chi2 = 0.3851
| Events Events
| observed expected
--------------------------+-------------------------
African American | 40 46.42
Others | 67 60.58
--------------------------+-------------------------
Total | 107 107.00

chi2(1) = 1.57
Pr>chi2 = 0.2101

Event | White Black Others | Total
-----------+---------------------------------+----------
No | 437 87 116 | 640
Yes | 242 40 67 | 349
-----------+---------------------------------+----------
Total | 679 127 183 | 989

Table 21. Kaplan-Meier Life Table by Histological Response (Huvos grading) in the
CCG-782 and INT-0133 study
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
I/II
1 396 79 0.8344 0.0170 0.7979 0.8649
2 317 72 0.6812 0.0214 0.6371 0.7211
3 282 25 0.6265 0.0223 0.5810 0.6685
4 246 16 0.5896 0.0228 0.5434 0.6329
5 214 3 0.5820 0.0230 0.5356 0.6255
6 180 5 0.5671 0.0233 0.5200 0.6113
7 148 2 0.5606 0.0235 0.5132 0.6052
8 112 1 0.5567 0.0237 0.5090 0.6016
9 69 5 0.5256 0.0262 0.4730 0.5755
10 29 0 0.5256 0.0262 0.4730 0.5755
III/IV
1 314 16 0.9516 0.0118 0.9222 0.9701
2 285 24 0.8782 0.0181 0.8376 0.9091
3 257 19 0.8189 0.0214 0.7726 0.8567
4 229 13 0.7762 0.0233 0.7265 0.8180
5 205 2 0.7693 0.0236 0.7191 0.8118
6 169 3 0.7571 0.0243 0.7056 0.8009
7 138 2 0.7464 0.0251 0.6933 0.7917
8 98 1 0.7398 0.0257 0.6854 0.7863
9 54 2 0.7182 0.0293 0.6562 0.7710
10 23 0 0.7182 0.0293 0.6562 0.7710
-------------------------------------------------------------------------------

30

Figure 11. Kaplan-Meier curve by Histological Response (Huvos grading) in the
CCG-782 and INT-0133 study
Table 22. Log-rank test for Histological Response (Huvos grading) in the CCG-782
and INT-0133 study
| Events Events
| observed expected
-------------+-------------------------
I/II | 208 158.54
III/IV | 82 131.46
-------------+-------------------------
Total | 290 290.00

chi2(1) = 34.13
Pr>chi2 = 0.0000

Event | I/II III/IV | Total
-----------+----------------------+----------
No | 273 251 | 524
Yes | 208 82 | 290
-----------+----------------------+----------
Total | 481 333 | 814

0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
I/II
III/IV
Histologic Response (Huvos Grading)
Histologic Response of CCG-782 and INT-0133
Event-Free Survival From Definitive Surgery

31

Table 23. Kaplan-Meier Life Table by Histological Response (non-Huvos grading) in
the P9754 study
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
I/II
1 98 14 0.8750 0.0313 0.7981 0.9240
2 77 18 0.7109 0.0431 0.6166 0.7860
3 69 8 0.6361 0.0460 0.5386 0.7184
4 62 2 0.6166 0.0466 0.5185 0.7006
5 51 2 0.5953 0.0474 0.4962 0.6812
6 41 0 0.5953 0.0474 0.4962 0.6812
7 19 0 0.5953 0.0474 0.4962 0.6812
8 2 0 0.5953 0.0474 0.4962 0.6812
III/IV
1 58 5 0.9194 0.0346 0.8170 0.9656
2 52 5 0.8387 0.0467 0.7210 0.9098
3 49 2 0.8058 0.0503 0.6834 0.8847
4 45 2 0.7715 0.0537 0.6447 0.8579
5 40 1 0.7540 0.0553 0.6252 0.8439
6 32 0 0.7540 0.0553 0.6252 0.8439
7 15 0 0.7540 0.0553 0.6252 0.8439
-------------------------------------------------------------------------------

Figure 12. Kaplan-Meier Curve by Histological Response (non-Huvos grading) in
the P9754 study

0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 2 4 6 8
Year
I/II
III/IV
Histologic Response (non-Huvos Grading)
Histologic Response of P9754
Event-Free Survival From Definitive Surgery

32

Table 24. Log-rank test for Histological Response (non-Huvos grading) in the P9754
study
| Events Events
| observed expected
-------------+-------------------------
I/II | 44 36.22
III/IV | 15 22.78
-------------+-------------------------
Total | 59 59.00

chi2(1) = 4.34
Pr>chi2 = 0.0373

| Histologic Response
Event | I/II III/IV | Total
-----------+----------------------+----------
No | 69 47 | 116
Yes | 44 15 | 59
-----------+----------------------+----------
Total | 113 62 | 175

Table 25. Kaplan-Meier Life Table by Histological Response (Hubos and non-Huvos
grading)
Beg. Survivor Std.
Time Total Fail Function Error [95% Conf. Int.]
-------------------------------------------------------------------------------
Huvos I/II
1 396 79 0.8344 0.0170 0.7979 0.8649
2 317 72 0.6812 0.0214 0.6371 0.7211
3 282 25 0.6265 0.0223 0.5810 0.6685
4 246 16 0.5896 0.0228 0.5434 0.6329
5 214 3 0.5820 0.0230 0.5356 0.6255
6 180 5 0.5671 0.0233 0.5200 0.6113
7 148 2 0.5606 0.0235 0.5132 0.6052
8 112 1 0.5567 0.0237 0.5090 0.6016
9 69 5 0.5256 0.0262 0.4730 0.5755
10 29 0 0.5256 0.0262 0.4730 0.5755
Huvos III/IV
1 314 16 0.9516 0.0118 0.9222 0.9701
2 285 24 0.8782 0.0181 0.8376 0.9091
3 257 19 0.8189 0.0214 0.7726 0.8567
4 229 13 0.7762 0.0233 0.7265 0.8180
5 205 2 0.7693 0.0236 0.7191 0.8118
6 169 3 0.7571 0.0243 0.7056 0.8009
7 138 2 0.7464 0.0251 0.6933 0.7917
8 98 1 0.7398 0.0257 0.6854 0.7863
9 54 2 0.7182 0.0293 0.6562 0.7710
10 23 0 0.7182 0.0293 0.6562 0.7710
non-Huvos I/II
1 98 14 0.8750 0.0313 0.7981 0.9240
2 77 18 0.7109 0.0431 0.6166 0.7860
3 69 8 0.6361 0.0460 0.5386 0.7184
4 62 2 0.6166 0.0466 0.5185 0.7006
5 51 2 0.5953 0.0474 0.4962 0.6812
6 41 0 0.5953 0.0474 0.4962 0.6812
7 19 0 0.5953 0.0474 0.4962 0.6812
8 2 0 0.5953 0.0474 0.4962 0.6812
non-Huvos III/IV
1 58 5 0.9194 0.0346 0.8170 0.9656
2 52 5 0.8387 0.0467 0.7210 0.9098
3 49 2 0.8058 0.0503 0.6834 0.8847
4 45 2 0.7715 0.0537 0.6447 0.8579
5 40 1 0.7540 0.0553 0.6252 0.8439
6 32 0 0.7540 0.0553 0.6252 0.8439
7 15 0 0.7540 0.0553 0.6252 0.8439
-------------------------------------------------------------------------------

33

Figure 13. Kaplan-Meier Curve by Histological Response (Hubos and non-Huvos
grading)

Table 26. Log-rank test for Histological Response (Hubos and non-Huvos grading)
| Events Events
| observed expected
-----------------+-------------------------
Huvos I/II | 208 157.92
Huvos III/IV | 82 130.92
non-Huvos I/II | 44 36.84
non-Huvos III/IV | 15 23.33
-----------------+-------------------------
Total | 349 349.00

chi2(3) = 38.62
Pr>chi2 = 0.0000

| Events Events
| observed expected
-------------+-------------------------
Huvos I/II | 208 158.54
Huvos III/IV | 82 131.46
-------------+-------------------------
Total | 290 290.00

chi2(1) = 34.13
Pr>chi2 = 0.0000

| Events Events
| observed expected
---------------+-------------------------
Huvos I/II | 208 204.21
non-Huvos I/II | 44 47.79
0.00
0.25
0.50
0.75
1.00
Estimated Proportion Event-Free
0 5 10 15
Year
Huvos I/II
Huvos III/IV
non-Huvos I/II
non-Huvos III/IV
Histologic Response
By Histologic Response
Event-Free Survival From Definitive Surgery

34

---------------+-------------------------
Total | 252 252.00

chi2(1) = 0.37
Pr>chi2 = 0.5419
| Events Events
| observed expected
-----------------+-------------------------
Huvos I/II | 208 194.42
non-Huvos III/IV | 15 28.58
-----------------+-------------------------
Total | 223 223.00

chi2(1) = 7.44
Pr>chi2 = 0.0064

| Events Events
| observed expected
---------------+-------------------------
Huvos III/IV | 82 98.73
non-Huvos I/II | 44 27.27
---------------+-------------------------
Total | 126 126.00

chi2(1) = 13.21
Pr>chi2 = 0.0003

Log-rank test for equality of survivor functions

| Events Events
| observed expected
-----------------+-------------------------
Huvos III/IV | 82 82.46
non-Huvos III/IV | 15 14.54
-----------------+-------------------------
Total | 97 97.00

chi2(1) = 0.02
Pr>chi2 = 0.8948

Log-rank test for equality of survivor functions

| Events Events
| observed expected
-----------------+-------------------------
non-Huvos I/II | 44 36.22
non-Huvos III/IV | 15 22.78
-----------------+-------------------------
Total | 59 59.00

chi2(1) = 4.34
Pr>chi2 = 0.0373

3.3 Prognostic Factors
Backward stepwise procedure was used to model potential prognostic factors with EFS
from definitive surgery. As the result, primary local progression, study of enrollment, and
histological response, and the interaction of study of enrollment and histological response
were associated with EFS from definitive surgery in the final predictive model (Table 27).
The relative risk for patients with local progression was 1.52 (1.0-2.1) times higher than
patients who without local progression after adjustment for the study of enrollment,

35

histological response and the interaction term of study of enrollment and histological
response (p=0.05).
Table 27. Stepwise Proportional Hazards regression
| Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
Local Progression |
Yes | 1.449013 .2721784 1.97 0.048 1.002732 2.093917
|
Age |
>16 Age Whole Yr | 1.28793 .1731611 1.88 0.060 .9895754 1.676237
|
Sex |
male | 1.022158 .1137397 0.20 0.844 .8218669 1.271261
|
Study |
INT-0133 | 1.699951 .5748772 1.57 0.117 .8761596 3.298296
P9754 | 1.479334 .6065974 0.95 0.340 .6622667 3.304452
|
Histologic response |
I/II | 4.119317 1.391393 4.19 0.000 2.124773 7.986158
|
Study*His_Resp |
INT-0133*I/II | .4081841 .1503936 -2.43 0.015 .1982599 .8403833
P9754*I/II | .4355861 .1962746 -1.84 0.065 .1801041 1.053476
--------------------------------------------------------------------------------------

| Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
--------------------+----------------------------------------------------------------
Local Progression |
Yes | 1.441601 .2681848 1.97 0.049 1.001137 2.075851
|
Age |
>16 Age Whole Yr | 1.295179 .1701997 1.97 0.049 1.00109 1.675661
|
Study |
INT-0133 | 1.697516 .5739175 1.57 0.118 .8750425 3.293054
P9754 | 1.475827 .6048948 0.95 0.342 .6609287 3.29546
|
Histologic response |
I/II | 4.114735 1.389632 4.19 0.000 2.122625 7.976466
|
Study*His_Resp |
INT-0133*I/II | .4087358 .1505705 -2.43 0.015 .1985529 .8414127
P9754*I/II | .4374386 .1968825 -1.84 0.066 .181054 1.056881
--------------------------------------------------------------------------------------

------------------------------------------------------------------------------------
Characteristics | No. of Patient Haz. Ratio [95% Conf. Interval] P-value
--------------------+---------------------------------------------------------------
Local Progression |
No | 905 1.0 --
Yes | 86 1.5 1.0 to 2.1 .05
--------------------+---------------------------------------------------------------
Age |
<=16 Age Whole Yr | 905 1.0 --
>16 Age Whole Yr | 1.3 1.0 to 1.7 .05
--------------------+---------------------------------------------------------------
Study | .28
CCG-782 | 206 1.0 --
INT-0133 | 610 1.7 0.9 to 3.3 .12
P9754 | 175 1.5 0.7 to 3.3 .34
--------------------+---------------------------------------------------------------
Histologic Response |
III/IV | 395 1.0 --
I/II | 596 4.1 2.1 to 8.0 <.00001

36

--------------------+---------------------------------------------------------------
Study*his_Res | .05
INT-0133* I/II | 335 0.4 0.2 to 0.8 .02
P9754* I/II | 113 0.4 0.2 to 1.1 .07
-------------------------------------------------------------------------------------
4. Discussion
The main goal of this investigation was to determinate whether the primary local tumor
progression during the induction duration was associated with risk for adverse clinical
outcome of osteosarcoma. Because the observation of tumor progression at the primary
tumor site without identification of sites of new metastatic disease occurs infrequently in
non-metastatic osteosarcoma (Provisor, et al., 1987), this investigation aggregated three
studies (CCG-782, INT-0133, and P9754) into a large dataset (991 subjects) to explore
this relationship. In this aggregated dataset, there was only 8.7% (86 subjects) with local
tumor progression. From our result, local tumor progression was statistically significantly
associated with EFS. EFS for patients who had local tumor progression was 67%
compared to 51% for patients who did not at 5 years.
It has been known that the histological response after preoperative chemotherapy is
significantly associated with risk for adverse clinical outcome. In this investigation, the
histological response from Huvos grading was significantly associated with adverse
clinical events in the study of CCG-782 and INT-013, as well as the histological response
from non-Huvos grading in the P9754 study. There was also a statistically significant
association between age and adverse clinical events.
To address the significance of the main effect of local tumor progression on adverse
events, other prognostic factors were considered as confounders, especially the
histological response, which had a statistically significantly strong association with
adverse clinical events. Accordingly, the stepwise proportional hazards regression model
with the characteristic of interest provided the prognostic significance and associated
relative risk of various patient characteristics measured were assessed. The relative risk of
an adverse analytic event for patients who had local progression was 1.5 times higher
than patients who without local progression after adjusted for the study of enrollment,
histological response and the interaction between study of enrollment and histological

37

response.
Our finding revealed that the primary tumor progression during the induction duration is
an important prognostic factor for accessing the subsequent relative risk of osteosarcoma
cancer after adjustment for other prognostic factors. This mean that local tumor
progression strengthens the prediction of treatment failure given the increasing size of
primary tumor during induction therapy was not necessarily associated with poor
histological response to induction chemotherapy. In summary, these finding give a
supplement/additive viewpoint to evaluate the treatment plan during the preoperative
period.
5. Reference
Jaffe, N., Bruland, Ø.S., Bielack, S.S., and Emmeking, W.F. (2009). Pediatric and
adolescent osteosarcoma, Vol 152 (New York ;London: Springer).
Kalbfleisch, J.D., and Prentice, R.L.j.a. (1980). The statistical analysis of failure time
data (New York: Wiley).
Link, M.P., Goorin, A.M., Miser, A.W., Green, A.A., Pratt, C.B., Belasco, J.B., Pritchard,
J., Malpas, J.S., Baker, A.R., Kirkpatrick, J.A., et al. (1986). The Effect of Adjuvant
Chemotherapy on Relapse-Free Survival in Patients with Osteosarcoma of the Extremity.
The New England journal of medicine 314, 1600-1606.
Meyers, P.A. (1987). Malignant bone tumors in children: osteosarcoma.
Hematology/oncology clinics of North America 1, 655.
Meyers, P.A., Heller, G., Healey, J., Huvos, A., Lane, J., Marcove, R., Applewhite, A.,
Vlamis, V., and Rosen, G. (1992). Chemotherapy for nonmetastatic osteogenic sarcoma:
the Memorial Sloan-Kettering experience. Journal of clinical oncology : official journal
of the American Society of Clinical Oncology 10, 5-15.
Meyers, P.A., Schwartz, C.L., Krailo, M., Kleinerman, E.S., Betcher, D., Bernstein, M.L.,
Conrad, E., Ferguson, W., Gebhardt, M., Goorin, A.M., et al. (2005). Osteosarcoma: A

38

Randomized, Prospective Trial of the Addition of Ifosfamide and/or Muramyl Tripeptide
to Cisplatin, Doxorubicin, and High-Dose Methotrexate. Journal of Clinical Oncology 23,
2004-2011.
Mirabello, L., Troisi, R.J., and Savage, S.A. (2009). Osteosarcoma incidence and survival
rates from 1973 to 2004: data from the Surveillance, Epidemiology, and End Results
Program. Cancer 115, 1531-1543.
Pizzo, P.A. (2010). Principles and Practice of Pediatric Oncology (6th Edition) (Wolters
Kluwer).
Provisor, A.J., Ettinger, L.J., Nachman, J.B., Krailo, M.D., Makley, J.T., Yunis, E.J.,
Huvos, A.G., Betcher, D.L., Baum, E.S., Kisker, C.T., et al. (1997). Treatment of
nonmetastatic osteosarcoma of the extremity with preoperative and postoperative
chemotherapy: a report from the Children's Cancer Group. Journal of clinical oncology :
official journal of the American Society of Clinical Oncology 15, 76-84.
Scheffé, H. (1959). The analysis of variance (New York, Wiley).
Ward, E., DeSantis, C., Robbins, A., Kohler, B., and Jemal, A. (2014). Childhood and
adolescent cancer statistics, 2014. CA: A Cancer Journal for Clinicians 64, 83-103.

Asset Metadata

Creator Zhong, Chang-Zheng (author)

Core Title 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

Contributor Electronically uploaded by the author (provenance)

School Keck School of Medicine

Degree Master of Science

Degree Program Biostatistics

Publication Date 08/25/2016

Defense Date 08/25/2015

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

Tag clinical trials,OAI-PMH Harvest,oncology,Osteosarcoma,survival analysis

Format application/pdf (imt)

Language English

Advisor Azen, Stanley (committee chair), Krailo, Mark (committee chair), Barkauskas, Donald (committee member)

Creator Email changzhz@usc.edu,qoqerwanjih@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-170475

Unique identifier UC11275214

Identifier etd-ZhongChang-3844.pdf (filename),usctheses-c40-170475 (legacy record id)

Legacy Identifier etd-ZhongChang-3844-0.pdf

Dmrecord 170475

Document Type Thesis

Format application/pdf (imt)

Rights Zhong, Chang-Zheng

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

Abstract (if available)

Abstract Purpose: Three osteosarcoma clinical trials (INT-0133, P9754, and CCG-782) from the Children’s Oncology Group (COG) were included to identify the association between treatment failure (metastases and relapses) and the change in primary tumor size during preoperative chemotherapy. We are interested in the induction duration, which was defined as the time of start preoperative chemotherapy to definitive surgery. Definitive surgery is considered a necessary part of osteosarcoma treatment and that, occasionally, during the induction period, a patient can experience an adverse clinical outcome. Thus, the occurrence of new sites of disease conventionally has required a change in treatment, but the action to be taken if the primary tumor size, as assessed by imaging radiographic, increases is not clear. ❧ Patients and Methods: Since the enrollment criteria for the studies varied based on site and extent of disease, criteria was established to select patients in this investigation. As a result, there were 991 patients who were under 30 years without metastasis. Event free survival (EFS) was the primary outcome measure from definitive surgery. EFS was estimated with the Kaplan-Meier method. The EFS hazard rate was compared across groups defined by prognostic factors using the log-rank statistic. Prognostic factor significance and associated relative risk of various patient characteristics measured were assessed with a proportional hazards regression model with the characteristic of interest. ❧ Results: For patients without metastases, and early disease progression during induction duration, the EFS was 65.8% at 5 years after the definitive surgery. The EFS for patients with the local tumor progression during induction duration was 50.9% compared to 67.1% for patients without local tumor progression at 5 years. The log-rank test indicated there was a statistically significant association between local tumor progression and EFS. The EFS of histological response in the CCG-782 and INT-0133 study were 58.2% for the standard response compared to 76.9% for the good response at 5 years. The log-rank test indicated there were statistically significantly different between the histological responses. In the P9754 study, the EFS of histological response were 59.5% for the standard response compared to 75.4% for the good response at 5 years. The Log-rank test identified there was a statistically significant difference between the two histological responses. The relative risk for patients with local progression was 1.52 (1.0-2.1) times higher than patients who without local progression after adjustment for the study of enrollment, histological response and the interaction term of study of enrollment and histological response. ❧ Conclusion: Our finding revealed that the primary tumor progression during the induction duration is an important prognostic factor for accessing the subsequent relative risk of osteosarcoma cancer after adjustment for other prognostic factors. Local tumor progression strengthens the prediction of treatment failure given that the increasing size of primary tumor during induction therapy was not necessarily associated with poor histological response to induction chemotherapy. In summary, these finding give a supplement/additive viewpoint to evaluate the treatment plan during the preoperative period.