Thomas Jefferson University
Jefferson Digital Commons
Department of Radiation Oncology Faculty Papers
Department of Radiation Oncology
May 2008
Increasing tumor volume is predictive of poor
overall and progression-free survival: Secondary
analysis of the Radiation Therapy Oncology Group
93-11 phase I-II radiation dose-escalation study in
patients with inoperable non-small-cell lung cancer
Maria Werner-Wasik
Thomas Jefferson University, maria.werner-wasik@jefferson.edu
R. Suzanne Swann
Radiation Therapy Oncology Group Headquarters, Philadelphia, PA
Jeffrey Bradley
Washington University
Mary Graham
Phelps County Regional Medical Center, Rolla, MO
Bahman Emami
Loyola University Medical Center
Recommended Citation
Werner-Wasik, Maria; Swann, R. Suzanne; Bradley, Jeffrey; Graham, Mary; Emami, Bahman; Purdy,
James; and Sause, William, "Increasing tumor volume is predictive of poor overall and progressionfree survival: Secondary analysis of the Radiation Therapy Oncology Group 93-11 phase I-II
radiation dose-escalation study in patients with inoperable non-small-cell lung cancer" (2008).
Department of Radiation Oncology Faculty Papers. Paper 4.
http://jdc.jefferson.edu/radoncfp/4
This Article is brought to you for free and open access by the Jefferson Digital Commons. The Jefferson Digital Commons is a service of Thomas
Jefferson University's Center for Teaching and Learning (CTL). The Commons is a showcase for Jefferson books and journals, peer-reviewed scholarly
publications, unique historical collections from the University archives, and teaching tools. The Jefferson Digital Commons allows researchers and
interested readers anywhere in the world to learn about and keep up to date with Jefferson scholarship. This article has been accepted for inclusion in
See next page for additional authors
Let us know how access to this document benefits you
Follow this and additional works at: http://jdc.jefferson.edu/radoncfp
Part of the Radiology Commons
Authors
Maria Werner-Wasik, R. Suzanne Swann, Jeffrey Bradley, Mary Graham, Bahman Emami, James Purdy, and
William Sause
This article is available at Jefferson Digital Commons: http://jdc.jefferson.edu/radoncfp/4
INCREASING TUMOR VOLUME IS PREDICTIVE OF POOR OVERALL AND
PROGRESSION-FREE SURVIVAL: SECONDARY ANALYSIS OF THE
RADIATION THERAPY ONCOLOGY GROUP 93-11 PHASE I-II RADIATION
DOSE-ESCALATION STUDY IN PATIENTS WITH INOPERABLE NON–SMALLCELL LUNG CANCER
MARIA WERNER-WASIK, M.D.,a
R. SUZANNE SWANN,PH.D.,b
JEFFREY BRADLEY, M.D.,c
MARY GRAHAM, M.D.,d
BAHMAN EMAMI, M.D.,e
JAMES PURDY,PH.D.,f
WILLIAM SAUSE, M.D.g
a
Department of Radiation Oncology, Jefferson Medical College, Thomas Jefferson
University Hospital, Philadelphia, PA;
b
Radiation Therapy Oncology Group Headquarters, Philadelphia, PA;
c
Washington University, St. Louis, MO;
d
Phelps County Regional Medical Center, Rolla, MO;
e
Loyola University Medical Center, Maywood, IL;
f
University of California, Davis, Medical Center, Sacramento, CA
g
LDS Hospital, Salt Lake City, UT
Presented at the International Association for Study of Lung Cancer Meeting, Barcelona,
Spain, July 2-6, 2005.
Contact:
Maria Werner-Wasik, M.D.
Department of Radiation Oncology,
Kimmel Cancer Center
Jefferson Medical College
111 S. 11th St., Philadelphia, PA 19107
Tel: (215) 955-6705
Fax: (215) 955-0412
Email: maria.werner-wasik@jeffersonhospital.org
This paper has been peer reviewed. It is the final version prior to publication in the International
Journal of Radiation Oncology, Biology, Physics 70(2):385-390, 2008. The published version is
available at doi: 10.1016/j.ijrobp.2007.06.034. Copyright © 2008 Elsevier, Inc.
Abstract
Purpose:
Patients with non–small-cell lung cancer (NSCLC) in the Radiation Therapy Oncology
Group (RTOG) 93-11 trial received radiation doses of 70.9, 77.4, 83.8, or 90.3 Gy. The
locoregional control and survival rates were similar among the various dose levels. We
investigated the effect of the gross tumor volume (GTV) on the outcome.
Methods and Materials:
The GTV was defined as the sum of the volumes of the primary tumor and involved
lymph nodes. The tumor response, median survival time (MST), and progression-free
3
3
survival (PFS) were analyzed separately for smaller (≤45 cm ) vs. larger (>45 cm )
tumors.
Results:
3
3
The distribution of the GTV was as follows: ≤45 cm in 79 (49%) and >45 cm in 82
3
(51%) of 161 patients. The median GTV was 47.3 cm . N0 status and female gender were
3
associated with better tumor responses. Patients with smaller (≤45 cm ) tumors achieved
3
a longer MST and better PFS than did patients with larger (>45 cm ) tumors (29.7 vs.
13.3 months, p < 0.0001; and 15.8 vs. 8.3 months, p < 0.0001, respectively). Increasing
the radiation dose had no effect on the MST or PFS. On multivariate analysis, only a
smaller GTV was a significant prognostic factor for improved MST and PFS (hazard
ratio [HR], 2.12, p = 0.0002; and HR, 2.0, p = 0.0002, respectively). The GTV as a
continuous variable was also significantly associated with the MST and PFS (HR, 1.59, p
< 0.0001; and HR, 1.39, p < 0.0001, respectively).
Conclusions:
Radiation dose escalation up to 90.3Gy did not result in improved MST or PFS. The
tumor responses were greater in node-negative patients and women. An increasing GTV
was strongly associated with decreased MST and PFS. Future radiotherapy trials patients
might need to use stratification by tumor volume.
INTRODUCTION
The current American Joint Committee on Cancer staging system for the primary tumor
in lung cancer is based mostly on the tumor extent and involvement of the neighboring
structures (e.g., pleura, chest wall, mediastinum, bone, esophagus, and proximal airways)
rather than on tumor size or volume. A notable exception is Stage T1, in which a tumor
surrounded completely by lung parenchyma cannot exceed 3 cm in the largest dimension.
However, a Stage T2 tumor can measure 1.5 cm or 8 cm, as long as it invades the visceral
pleura only, with sparing of the other structures. Evidence has been accumulating (1-11)
that an increasing tumor volume has a significant effect on patient outcome, possibly
even overriding the T stage assignment. Other factors influencing the American Joint
Committee on Cancer stage assignment are nodal involvement and the presence of distant
metastases.
In a recently published Radiation Therapy Oncology Group (RTOG) Phase I-II student
(12) of radiation dose escalation for patients with inoperable non–small cell-lung cancer
(NSCLC), the observed locoregional control rates and survival rates were similar
between treatment groups, receiving escalated radiation doses (from 70.9 Gy to 90.3 Gy,
depending on the volume of lung receiving ≥20 Gy [V20]). A reasonable initial
hypothesis would be, however, to expect that smaller tumors should demonstrate
improved local control with greater radiation doses compared with larger tumors.
To investigate this hypothesis, we undertook a retrospective analysis of data from the
RTOG 93-11 clinical trial in an attempt to demonstrate any benefit of radiation dose
escalation for patients with smaller tumors and to determine any relationship between the
initial tumor volume and patient outcome.
METHODS AND MATERIALS
Patient population
The RTOG 93-11 study was a Phase I-II radiation dose escalation trial for patients with
inoperable Stage I-III NSCLC treated with three-dimensional (3D) radiotherapy alone,
without concurrent chemotherapy, although induction chemotherapy was allowed. The
primary objective of the study was to determine the treatment-related morbidity and to
determine the maximal tolerated radiation dose. The secondary objectives were to
determine the local control and overall survival (OS) rates. The patient population
consisted of subjects with NSCLC (inoperable Stage I, II, and IIIA and Stage IIIB;
supraclavicular nodes involvement was not allowed; Table 1). Patients were treated
according to the volumetric treatment planning computed tomography findings and the
gross tumor volume (GTV) included the primary tumor and any enlarged regional lymph
nodes (>1 cm) with a minimal 3D margin of 1 cm. Noninvolved nodal areas were not
irradiated, and no special effort was made to account for the respiratory motion, apart for
assessing motion with fluoroscopy. Patients were placed into dose-escalation groups
according to the V20 value in their radiotherapy (RT) plan, predicting the likelihood of
treatment-related pneumonitis (13). Patients with aV20 of <25% were assigned to Group
1 and received an escalated dose to 70.9, 77.4, 83.8, or 90.3 Gy. Patients with a V20 of
26–35% were assigned to Group 2 and received an escalated dose to 70.9, 77.4, or 83.8
Gy. Patients with a V20 of >35% were assigned to Group 3 and received an escalating
dose to 64.5, 70.9, or 77.4 Gy. All fraction sizes were 2.15 Gy. The study accrued
patients only to Groups 1 and 2. Group 3 enrollment was stopped because of poor
accrual.
Evaluation of local control, OS, and progression-free survival
A chest X-ray was obtained 4 weeks after RT completion. Computed tomography scans
of the chest were obtained at 6 and 12 months and repeated yearly thereafter. Local
control (complete response [CR] or partial response [PR] vs. stable or progressive
disease) was reported by the enrolling institutions. No central review of the follow-up
computed tomography scans was performed. OS and progression-free survival (PFS)
were reported as measured from the date of registration in the study.
Statistical analysis
The GTV was defined as the sum of the volumes of the primary tumor and involved
lymph nodes. In the 3D plans, the primary tumor volume and the involved nodal volume
were outlined as one structure; no data are available in the RTOG electronic database to
allow for separation of those two volumes. Therefore, in an attempt to at least partially
correct this deficiency, nodal status (N0 vs. N1 or N2 or N3) was analyzed as one of the
variables. This allowed for the separation of the effect of the tumor GTV vs. nodal GTV
(at least for Stage I, or N0, patients). OS was defined as death from any cause; an event
for PFS was local or regional progression, distant metastases, or death from any cause.
For the purpose of this investigation, tumor response, OS, and PFS were analyzed
3
3
separately for the smaller tumors (≤45 cm ) vs. larger tumors (>45 cm ), first among all
patients and, later, within each radiation dose level. GTV was also analyzed as a
continuous variable. The association of response (CR/PR vs. stable/ progressive disease)
and the GTV categorized by cutpoint was tested by Fisher’s exact test. OS and PFS were
estimated by the Kaplan-Meier method and tested using the log–rank test statistic.
Univariate and multivariate analyses of OS and PFS with the GTV and other prognostic
factors (age [<60 vs. ≥60], gender, Karnofsky performance status [90–100 vs. 70–80],
histologic type [nonsquamous vs. squamous], stage [I-II vs. IIIA-IIIB], previous
chemotherapy [yes vs. no], and maximal radiation dose to the lung) were done using the
Cox proportional hazards model. Multivariate modeling used the stepwise selection
method. When analyzed as a continuous variable, GTV was transformed using a log10
transformation to ensure normality. Patients with unknown tumor volumes were excluded
from this analysis.
RESULTS
Patient characteristics
A total of 176 patients were included in the original report of the study (12). Of the 176
patients, 161 had available data on GTV and tumor response and were the subject of this
secondary analysis. The patient characteristics are presented in Table 1. Overall, most
patients were older (>60 years) with a Karnofsky performance status between 70 and 80.
The patients in this analysis were approximately equally split between men and women
and those in Group 1 were more likely to have node-negative disease than were those in
Group 2. The distribution of the American Joint Committee on Cancer stage was Stage I
in 67, Stage II in 12, and Stage III in 48 patients in Group 1 and Stage I in 10, Stage II in
3, and Stage III in 35 patients in Group 2.
Tumor response, OS, and PFS
3
3
The GTV was ≤45 cm in 79 (49%) and >45 cm ,82 (51%) of 161 patients (median, 47.3;
3
range, 1.9–1,039.9 cm ); 14 patients had an unknown GTV. The tumor response rate
3
3
(CR/PR) was better for smaller tumors (≤45 cm ) than for larger tumors (>45 cm ; 87%
vs. 76%, respectively), as was stable/progressive disease (13% vs. 24%, respectively; p =
3
0.0691, Fisher’s exact test). Results using a cutoff point of 30 cm did not better
distinguish between those patients with a tumor response and those with stable or
3
progressive disease than using a cutoff point of 45 cm (p = 0.0642). A cutoff point of 60
3
cm did not discriminate between the two groups (p = 0.4139). When the GTV was
analyzed as a continuous variable, on univariate analysis, it was borderline statistically
significantly associated with tumor response (p = 0.0551); however, on multivariate
analysis, N stage (N0 vs. N1-N3) and female gender were the only significant variables
(p = 0.025 and p = 0.02, respectively). This can be explained by the greater rate of
responses (70%) in patients with N0 disease vs. N1-N3 (30%).
3
Patients with smaller tumors (≤45 cm ) achieved a longer median survival than did
3
patients with larger tumors (>45 cm ; 29.7 vs. 13.3 months, p < 0.0001; Fig. 1), as well as
better median PFS (15.8 vs. 8.3 months; p < 0.0001; Fig. 2).
3
3
When a different GTV was chosen as a cutoff point (30 cm or 60 cm ), patients with
3
3
smaller tumors (≤30 cm or ≤60 cm ) still achieved better OS (32.9 vs. 14.6 months for
3
3
30 cm , p = 0.0002; and 26.8 vs. 13.3 months for 60 cm , p = 0.0006), as well as better
3
3
PFS (15.5 vs. 9.0 months for 30 cm , p = 0.0031; and 14.7 vs. 8.7 months for 60 cm , p =
0.0023).
When the effect of GTV was analyzed on univariate analysis, a smaller GTV was
associated with improved OS, with significant hazard ratios (HRs) for cutoff points of 30
3
3
3
cm (HR, 2.15; p = 0.0002); 45 cm (HR, 2.14; p < 0.0001); and 60 cm (HR, 1.91; p =
0.0008), as well as for GTV analyzed as a continuous variable (HR, 1.59; p < 0.0001).
The other variables associated with improved OS on univariate analysis were female
gender (p = 0.0407) and nodal status (p = 0.067, borderline significance). The same
factors were significant for PFS on univariate analysis (data not shown).
On multivariate analysis of the factors associated with improved OS and PFS, only a
smaller tumor volume was significantly prognostic for both endpoints (HR, 2.12; p =
0.0002; and HR, 2.0; p = 0.0002, respectively) when GTV was analyzed as a continuous
variable. Age, gender, performance status, histologic type, N stage (N0 vs. N1-N3),
previous chemotherapy, and maximal radiation dose were not significant (Tables 2 and
3
3
3
3). The other GTV cutoff points (≤30 cm , ≤45 cm , and ≤60 cm ) retained their
statistically significant association with improved OS and PFS on multivariate analysis
and again were the only factors in the multivariate models using a stepwise selection
method.
Effect of radiation dose escalation on tumor response, OS, and PFS by tumor volume
The primary research hypothesis of this study was that higher radiation doses would lead
to increased efficacy in smaller tumors. Table 4 shows the frequencies and percentages of
patients with a CR/PR and stable or progressive disease for each radiation dose and GTV
3
combination using the 45 cm cutoff point. No evidence was found in these data that the
CR/PR rates increased as the radiation dose increases for the two categories of GTV (p =
0.2213). Increasing the radiation dose had no effect on OS or PFS (data not shown for
3
PFS) when examined separately for smaller vs. larger tumors when the 45-cm GTV
3
3
cutoff point was used (Table 5). The results for the 30-cm and 60-cm cutoff points were
similar (data not shown). The consistently statistically significant increase in the relative
3
risk of death for all doses to a GTV >45 cm can be attributed to the strong effect of a
larger GTV on OS rather than the radiation dose. However, the analysis was not powered
to detect a dose–tumor volume interaction, and it could not be ruled out on the basis of
this analysis.
DISCUSSION
The aim of RTOG 93-11 was to determine the dose-limiting toxicity of 3D RT. The
radiation dose was safely escalated to 83.8 Gy for patients with V20 <25% and to 77.4 Gy
for patients with a V20 of 25–36%. The 90.3-Gy dose level was too toxic. The observed
locoregional control was similar among the study arms, without evidence that the higher
doses eliminated or at least lowered the recurrence rates.
Our initial hypothesis was that patients with volumetrically smaller tumors would have
improved survival with radiation dose escalation but not patients with larger tumors.
However, we were not able to demonstrate that in this secondary analysis of the RTOG
93-11 trial, at least not with the small patient numbers that were available at each
radiation dose level tested. It could be that doses >83.8 Gy in standard fractions are
necessary to eliminate local failure. Additionally, the protracted overall treatment time of
7–9 weeks might have facilitated tumor repopulation and therefore attenuated any
radiation dose response. Finally, the PTV margins were tight (1–1.5 cm around the
GTV), which might have increased the likelihood for a marginal miss in mobile tumors,
obliterating any potential benefit of dose escalation.
Such a benefit has been suggested in the Memorial Sloan-Kettering Cancer Center
experience (4), with the observation of improved local control and survival in Stage III
3
NSCLC patients with large (>100 cm ) tumors treated with radiation doses >64 Gy
compared with those who received lower radiation doses.
A significant interaction between radiation dose and tumor size was shown in the
University of Michigan retrospective analysis (5) of 114 patients with medically
inoperable Stage I and II NSCLC treated with 3D conformal RT in a dose-escalation
study. Patients treated to a biologically equivalent dose of ≤79.2 Gy lived longer if their
3
tumors did not exceed 51.8 cm in volume. However, patients treated to a biologically
equivalent dose of >79.2 Gy had the same overall survival, irrespective of tumor volume.
With all the limitations of the retrospective study, a hypothesis has been raised that
radiation dose escalation can result in improved outcome in NSCLC, at least in nodenegative, early-stage tumors.
In the reports of highly hypofractionated (‘‘radioablative’’) RT using precise localization
techniques to account for tumor motion, very high local control rates have been achieved
in medically inoperable patients with Stage I NSCLC receiving 60 Gy in three fractions
of 20 Gy each (10) or other hypofractionated regimens (11). Such doses have not yet
been tested in Stage III NSCLC and might be too dangerous for large and central tumors.
We found that the increasing tumor volume, defined as the sum of the primary tumor
volume and the volume of the involved lymph nodes, was associated with a greater risk
3
of local failure, with significantly better control achieved with tumors <45 cm than with
3
the larger tumors. The 45-cm volume corresponds roughly to a spherical tumor diameter
of 4.4 cm. It must be remembered that the ‘‘tumor volume’’ in our analysis denoted a
sum of the volume of the primary tumor and the involved lymph nodes, if any. However,
in the multivariate analysis of the tumor volume studied as a continuous variable, it was
only the earlier nodal stage and female gender, not the tumor volume, that was associated
with better local control. In reality, those two variables (volume and nodal stage) overlap
to a large degree, because Stage I NSCLC is defined as a node-negative tumor measuring
≤3 cm in the largest dimension. Separate values for the primary GTV and the nodal GTV
were not available in the RTOG 93-11 study; therefore, we were unable to isolate their
respective influences on outcomes.
Because a rigorous evaluation of locoregional control was not performed in the RTOG
93-11 trial, local control was not assessed in an actuarial fashion and the radiographic
responses might not reflect the true biologic tumor elimination; using survival as an
endpoint is a more objective measure of the relevance of tumor volume. A strong
association of increasing tumor volume with worsened survival and PFS was observed in
our analysis, overriding other known prognostic factors for survival, such as lower
disease stage.
Such an association has been previously reported (1–9).In 207 patients with inoperable
NSCLC (Stage I-III) treated at the Washington University with 3D-conformal thoracic
RT (1), overall survival, cause-specific survival, and local tumor control were highly
correlated with the GTV, and the GTV (and pathologic findings) were predictive for
survival on multivariate analysis, but overall stage and nodal stage were not. Those
3
patients with tumor volumes not exceeding 33 cm appeared to have the best outcome.
Local response was evaluated volumetrically on 107 follow-up thoracic computed
tomography scans of 22 patients (19 with Stage III NSCLC) treated with definitive
3
thoracic RT (2).A volume of ≤63 cm and a diameter of ≤4 cm were significantly
associated with improved local control compared with larger volumes or diameters. In a
large series from Wuerzburg (6), 784 scans of 136 patients were evaluated
3
volumetrically, and a cutoff point of 100 cm for tumor volume was a discriminating
factor for local control, but not survival. In that study, the primary tumor volume and
nodal volume were measured separately. The total tumor volume (tumor plus nodes), as
well as primary tumor volume alone, was a significant prognostic factor for survival in a
Japanese group experience (7).
Because most of the studies cited in our report included a significant proportion of
patients with nodal involvement (N1-N3), the relative prognostic value of the ‘‘T’’ tumor
volume vs. the ‘‘N’’ nodal volume needs to be elucidated. One would expect that worse
survival and possibly lower local control would be associated with an increasing nodal
volume rather than the primary tumor volume. However, contradictory data have been
published on this issue. On univariate analysis of the factors associated with overall
survival and failure-free survival in a Phase I-II radiation dose-escalation trial (3), only
the increasing GTV (defined as tumor plus nodes), but not the nodal stage or the overall
stage, were predictive. Similarly, in the Japanese experience (7) of 71 patients with Stage
III NSCLC, on univariate analysis, the total tumor volume and the primary tumor volume
were significant and the nodal volume was not. On multivariate analysis, the total tumor
volume and primary tumor volume were both significant prognostic factors.
Investigators from Shanghai Medical University (8) created a prognostic index model
predicting for local control in patients with NSCLC treated with RT. Patients with a
smaller tumor volume (primary plus nodal), earlier clinical stage, and treated with higher
total irradiation dose with a shortened overall treatment time had better local control.
In a Classification and Regression Tree analysis of the Thomas Jefferson University’s
3
107 patients with Stage III NSCLC (9), an aggregate nodal volume >12.5 cm (sum of
volumes of the abnormal hilar and mediastinal lymph nodes), as well as a central tumor
location, but not the primary tumor volume, were associated with a greater risk of nodal
3
recurrence and shorter median survival time than a nodal volume of ≤12.5 cm (MST
13.9 months vs. 17.1 months, respectively). We are not aware of other reports that have
focused on the prognostic value of the involved nodal volume.
CONCLUSIONS
Our study is one of several publications demonstrating the importance of tumor volume
in patients receiving thoracic RT for NSCLC. It is not fully clear whether patients with
smaller tumors have better outcomes simply because of the lower number of clonogenic
cells or whether smaller tumors are inherently more biologically favorable; however, the
tumor volume may need to be considered in the staging system for lung cancer, once
user-friendly volume assessment becomes commonplace in diagnostic studies.
REFERENCES
1. Bradley JD, Ieumwananonthachai N, Purdy JA, et al. Gross tumor volume, critical
prognostic factor in patients treated with three-dimensional conformal radiation
therapy for non–small-cell lung carcinoma. Int J Radiat Oncol Biol Phys
2002;52:49–57.
2. Werner-Wasik M, Xiao Y, Pequignot E, et al. Assessment of lung cancer response
after non-operative therapy: Tumor diameter, bidimensional product, and
volume—A serial CT scan-based study. Int J Radiat Oncol Biol Phys 2001;51:56–
61.
3. Belderbos JS, Heemsbergen WD, DeJaeger K, et al. Final results of a phase I/II
dose escalation trial in non–small-cell lung cancer using three-dimensional
conformal radiotherapy. Int J Radiat Oncol Biol Phys 2006;66:126–134.
4. Rengan R, Rosenzweig KE, Venkatraman E, et al. Improved local control with
higher doses of radiation in large-volume Stage III non–small-cell lung cancer. Int
J Radiat Oncol Biol Phys 2004;60:741–747.
5. Zhao L, West B, Hayman JA, et al. High radiation dose may reduce the negative
effect of large gross tumor volume in patients with medically inoperable earlystage non–small-cell lung cancer. Int J Radiat Oncol Biol Phys 2007;68:103–110.
6. Willner J, Baier K, Caragiani E, et al. Dose, volume and tumor control prediction
in primary radiotherapy of non–small-cell lung cancer. Int J Radiat Oncol Biol
Phys 2002;52:382–389.
7. Basaki K, Abe Y, Kondo H, et al. Prognostic factors for survival in Stage III non–
small-cell lung cancer treated with definitive radiation therapy: Impact of tumor
volume. Int J Radiat Oncol Biol Phys 2006;64:449–454.
8. Chen M, Jiang GL, Fu XL, et al. Prognostic factors for local control in non-small
cell lung cancer treated with definitive radiation therapy. Am J Clin Oncol
2002;25:76–80.
9. Werner-Wasik M, Peqignot E, Garofola B, et al. Volume of involved mediastinal
lymph nodes and tumor location are predictive of tumor recurrence: Classification
and regression tree (CART) analysis of patients with Stage III non-small cell lung
cancer [Abstract]. Proc ASCO 2003;22:638.
10. Timmerman R, Papiez L, McGarry R, et al. Extracranial stereotactic
radioablation: Results of a phase I study in medically inoperable Stage I nonsmall cell lung cancer. Chest 2003; 124:1946–1955.
11. Onishi H, Araki T, Shirato H, et al. Stereotactic hypofractionated high-dose
irradiation for stage I nonsmall cell lung carcinoma: Clinical outcomes in 245
subjects in a Japanese multiinstitutional study. Cancer 2004;101:1623–1631.
12. Bradley J, Graham MV, Winter K, et al. Toxicity and outcome results of RTOG
93-11: A phase I-II dose-escalation study using three-dimensional conformal
radiotherapy in patients with inoperable non–small-cell lung carcinoma. Int J
Radiat Oncol Biol Phys 2005;61:318–328.
13. Graham MV, Purdy JA, Emami B, et al. Clinical dose–volume histogram analysis
for pneumonitis after 3D treatment for non– small-cell lung cancer. Int J Radiat
Oncol Biol Phys 1999;45: 323–329.
TABLES AND FIGURES
Table 1. Patient characteristics
___________________________________________________________________
Characteristic
Group 1 (n = 127)
Group 2 (n = 48)
___________________________________________________________________
Age (y)
<60
18 (14)
5 (10)
≥60
109 (86)
43 (90)
Gender (n)
Male
72 (57)
22 (46)
Female
55 (43)
26 (54)
KPS (n)
70–80
85 (67)
30 (63)
90–100
42 (28)
18 (37)
Histologic type (n)
Squamous cell carcinoma
51 (40)
21 (44)
Adenocarcinoma
42 (33)
17 (35)
Other
N stage (n)
34 (21)
10 (21)
N0
83 (65)
17 (35)
N1
10 (8)
6 (13)
N2
32 (25)
22 (46)
N3
2 (1)
3 (6)
___________________________________________________________________
Abbreviation: KPS = Karnofsky performance status.
Data in parentheses are percentages.
Table 2. Multivariate analysis of overall survival for different gross tumor volumes used
as cutoff point and as continuous variable
________________________________________________________________________
Modela
Comparison
Hazard ratio
pb
95% CI
________________________________________________________________________
3
GTV (cm )
<30 vs. ≥30
2.18
1.43–3.32
0.0003
3
GTV (cm )
≤45 vs. >45
2.12
1.43–3.13
0.0002
GTV (cm )
≤60 vs. >60
1.87
1.27–2.75
0.0015
GTVc
Continuous
1.59
1.33–1.91
<0.0001
3
________________________________________________________________________
Abbreviations: CI = confidence interval; GTV = gross tumor volume; KPS =
Karnofsky performance status.
a
Following covariates did not meet entry criteria for any multivariate model: age (<60
vs. ≥60 y), gender (female vs. male), KPS (90–100 vs. 70–80), histologic type
(nonsquamous vs. squamous), N stage (N0 vs. N1-N3), previous chemotherapy (no vs.
yes), or maximal dose to lung (continuous).
b
Chi-square test using Cox proportional hazards model; stepwise selection, with
entry level of 0.05 and exit level of 0.10.
c
GTV transformed using log10 to ensure normalcy.
Table 3. Multivariate analysis of progression-free survival for different gross tumor
volumes used as cutoff point and as continuous variable
________________________________________________________________________
Modela
Comparison
Hazard ratio
pb
95% CI
________________________________________________________________________
3
GTV (cm )
<30 vs. ≥30
1.74
1.20–2.53
0.0039
3
GTV (cm )
≤45 vs. >45
2.00
1.40–2.86
0.0002
GTV (cm )
≤60 vs. >60
1.65
1.16–2.36
0.0056
c
Continuous
1.39
1.18–1.64
<0.0001
3
GTV
________________________________________________________________________
Abbreviations as in Table 2.
a
Following covariates did not meet entry criteria for any multivariate model: age (<60
vs. ≥60 y), gender (female vs. male), KPS (90–100 vs. 70–80), histologic type
(nonsquamous vs. squamous), N stage (N0 vs. N1-N3), previous chemotherapy (no vs.
yes), or maximal dose to lung (continuous).
b
Chi-square test using Cox proportional hazards model; stepwise selection, with
entry level of 0.05 and exit level of 0.10.
c
GTV was transformed using log10 to ensure normality.
Table 4. Frequency of tumor response subdivided by radiation dose level and gross tumor
3
volume cutpoint of 45 cm
GTV ≤45 cm3
Dose 70.9 Gy
Dose 77.4 Gy
Dose 83.8 Gy
Dose 90.3 Gy
Dose 70.9 Gy
Dose 77.4 Gy
Dose 83.8 Gy
Dose 90.3 Gy
Incidence (n)
CR/PR
SD/PD
13 (93)
1 (7)
14 (82)
3 (18)
15 (88)
2 (12)
23 (88)
3 (12)
21 (70)
9 (3)
19 (68)
9 (32)
12 (92)
1 (8)
8 (89)
1 (11)
p*
0.2736
Abbreviations:
CR = complete response; PR = partial response; SD = stable disease; PD = progressive
disease; GTV = gross tumor volume.
Data in parentheses are percentages.
* Fisher’s exact test.
Table 5. Multivariate analysis of overall survival subdivided by radiation dose level
3
and gross tumor volume cutpoint of 45 cm
________________________________________________________________________________________________________
Hazard
Model
n
ratio
95% CI
p*
______________________________________________________________________
GTV ≤45 cm3, dose 83.8 Gy
GTV ≤45 cm3, dose 77.4 Gy
GTV ≤45 cm3, dose 70.9 Gy
GTV >45 cm3, dose 90.3 Gy
GTV >45 cm3, dose 83.8 Gy
GTV >45 cm3, dose 77.4 Gy
GTV >45 cm3, dose 70.9 Gy
17
17
14
9
13
28
30
1.60
1.10
1.57
4.20
3.83
2.41
2.61
Abbreviations as in Table 4.
3
Reference level: GTV ≤45 cm , dose 90.3 Gy.
* Chi-square test using Cox proportional hazards model.
0.65–3.93
0.43–2.82
0.63–3.91
1.52–11.64
1.53–9.60
1.06–5.48
1.17–5.84
0.3058
0.8432
0.3301
0.0058
0.0041
0.0361
0.0193
Fig. 1. Five-year overall survival rate for patients with gross tumor volume ≤45 cm
3
(solid curve) vs. those with gross tumor volume >45 cm (dotted curve).
Fig. 2. Five-year progression-free survival rate for patients with gross tumor volume ≤45
3
3
cm (solid curve) vs. those with gross tumor volume >45 cm (dotted curve).
This paper has been peer reviewed. It is the final version prior to publication in the International
Journal of Radiation Oncology, Biology, Physics 70(2):385-390, 2008. The published version is
available at doi: 10.1016/j.ijrobp.2007.06.034. Copyright © 2008 Elsevier, Inc.
3