JOURNAL OF CLINICAL ONCOLOGY
ORIGINAL REPORT
Impact of Intensity-Modulated Radiation Therapy Technique
for Locally Advanced Non–Small-Cell Lung Cancer: A
Secondary Analysis of the NRG Oncology RTOG 0617
Randomized Clinical Trial
Stephen G. Chun, Chen Hu, Hak Choy, Ritsuko U. Komaki, Robert D. Timmerman, Steven E. Schild,
Jeffrey A. Bogart, Michael C. Dobelbower, Walter Bosch, James M. Galvin, Vivek S. Kavadi, Samir Narayan,
Puneeth Iyengar, Clifford G. Robinson, Raymond B. Wynn, Adam Raben, Mark E. Augspurger, Robert M. MacRae,
Rebecca Paulus, and Jeffrey D. Bradley
ABSTRACT
Purpose
Although intensity-modulated radiation therapy (IMRT) is increasingly used to treat locally advanced
non–small-cell lung cancer (NSCLC), IMRT and three-dimensional conformal external beam radiation
therapy (3D-CRT) have not been compared prospectively. This study compares 3D-CRT and IMRT
outcomes for locally advanced NSCLC in a large prospective clinical trial.
Patients and Methods
A secondary analysis was performed to compare IMRT with 3D-CRT in NRG Oncology clinical trial
RTOG 0617, in which patients received concurrent chemotherapy of carboplatin and paclitaxel with
or without cetuximab, and 60- versus 74-Gy radiation doses. Comparisons included 2-year overall
survival (OS), progression-free survival, local failure, distant metastasis, and selected Common
Terminology Criteria for Adverse Events (version 3) $ grade 3 toxicities.
Results
The median follow-up was 21.3 months. Of 482 patients, 53% were treated with 3D-CRT and 47%
with IMRT. The IMRT group had larger planning treatment volumes (median, 427 v 486 mL; P = .005);
a larger planning treatment volume/volume of lung ratio (median, 0.13 v 0.15; P =.013);andmore
stage IIIB disease (30.3% v 38.6%, P = .056). Two-year OS, progression-free survival, local failure, and
distant metastasis–free survival were not different between IMRT and 3D-CRT. IMRT was associated
with less $ grade 3 pneumonitis (7.9% v 3.5%, P = .039) and a reduced risk in adjusted analyses (odds
ratio, 0.41; 95% CI, 0.171 to 0.986; P = .046). IMRT also produced lower heart doses (P , .05), and
the volume of heart receiving 40 Gy (V40) was significantlyassociatedwithOSonadjusted
analysis (P , .05). The lung V5 was not associated with any $ grade 3 toxicity, whereas the lung
V20 was associated with increased $ grade 3 pneumonitis risk on multi variable analysis (P = .026).
Conclusion
IMRT was associated with lower rates of severe pneumonitis and cardiac doses in NRG Oncology
clinical trial RTOG 0617, which supports routine use of IMRT for locally advanced NSCLC.
J Clin Oncol 35 :56-62. © 2016 by American Socie ty of Clini cal Oncology
INTRODUCTION
Non–small-cell lung cancer (NSCLC) is the
leading cancer killer in the United States,
1
and the
standard of care for locally advanced unresectable
NSCLC is concurrent radiation therapy (RT)
with cytotoxic chemotherapy.
2 -5
Historically,
locally advanced NSCLC has been treated with
three-dimensional conformal external beam
radiation therapy (3D-CRT) with concurrent
chemotherapy,
6 -8
but there has been increas-
ing use of intensity-modulated radiation ther-
apy (IMRT).
9-12
Although IMRT is more complex
to plan and deliver than 3D-CRT, its pote n-
tialclinicalbenefits have not previously been
assessed in a multi-i nstitutional prospective
clinical tr ial.
13
IMRT uses complex modulated radiation
beams that sculpt the radiation dose to precisely
conform to complex geometric targets, which
creates sharper radiation dose gradients between
Author affiliations appear at the end of this
article.
Published at ascopubs.org/journal/jco on
October 3, 2016.
Support information appears at the end
of this article.
Presented at the World Conference on
Lung Cancer, Denver, CO, September 6-9
2015; American Society of Radiation
Oncology Annual Meeting, San Antonio,
TX, October 18-21, 2015; and 2015 Best
of ASTRO, San Diego, CA, November
13-14, 2015.
Clinical trial information: NCT00533949.
Corresponding author: Stephen G. Chun,
MD, Department of Radiation Oncology,
University of Texas MD Anderson Cancer
Center, 1515 Holcombe Blvd, Houston,
TX 77030; e-mail: sgchun@mdanderson.
org.
© 2016 by American Society of Clinical
Oncology
0732-183X/17/3501w-56w/$20.00
ASSOCIATED CONTENT
See accompanying Oncology
Grand Rounds on page 6
Appendix
DOI: 10.1200/JCO.2016.69.1378
Data Supplement
DOI: 10.1200/JCO.2016.69.1378
DOI: 10.1200/JCO.2016.69.1378
56 © 2016 by American Society of Clinical Oncology
VOLUME 35
•
NUMBER 1
•
JANUARY 1, 2017
tumor and normal tissue than 3D-CRT. For these reasons, IMRT
can improve radiation coverage of tumors and enhance the
therapeutic ratio by avoiding adjacent organs at risk. IMRT has
a number of theoretical advantages over 3D-CRT for locally ad-
vanced NSCLC. Dosimetric studies have shown IMRT to reduce
the doses delivered to adjacent normal tissue, such as the lungs,
esophagus, and heart, by improving conformity of the RT dose
distribution.
14-18
A retrospective MD Anderson Cancer Center
study compared IMRT with 3D-CRT and found that IMRT pro-
vides equivalent survival to 3D-CRT despite IMRT patients having
significantly worse performance status and larger tumors.
19
A
SEER and a National Cancer Database study found IMRT to be
associated with improved overall survival (OS) compared with
3D-CRT in patients treated in the United States for stage III
NSCLC.
9,20
IMRT also resulted in favorable outcomes compared
with historic controls in a Memorial Sloan Kettering Cancer Center
study.
21
These findings have provided the impetus to evaluate
IMRT in the context of a large, prospective, multi-institutional
clinical trial for locally advanced NSCLC.
22
NRG Oncology clinical trial RTOG 0617 was a randomized
phase III trial that used a two-by-two factorial design to assess
theroleofRTdoseescalation(60v 74 Gy) and the addition
of cetuximab to weekly car boplat in and paclitaxel (car boplatin
and p ac l i t axel wi t h or w i t h ou t cetuximab) for locally advanced
NSCLC.
23
The use of IMRTor 3D-CRTwas one of the stratification
factors at random assignment, which resulted in a balanced use of
these techniques within the 60- and 74-Gy arms. The current study
is a secondary analysis comparing outcomes of IMRT versus
3D-CRT in NRG Oncology clinical trial RTOG 0617.
PATIENTS AND METHODS
Design, Setting, and Participants
All eligible patients enrolled in NRG Oncology clinical trial RTOG
0617 from November 2007 through November 2011were included in this
secondary analysis. The study details of NRG Oncology clinical trial RTOG
0617 were reported in the primary outcome manuscript.
23
The CONSORT
diagram for NRG Oncology clinical trial RTOG 0617 is shown in Figure 1.
All patients had histologically proven NSCLC and had American Joint
Committee on Cancer stage IIIA or IIIB disease with Zubrod performance
status 0 to 1.
Statistical Considerations
All eligible patients were included in this secondary analysis. For
patient and tumor characteristics, categorical data were compared with x
2
or Fisher exact test as appropriate, and continuous data were compared
with Wilcoxon rank sum test. Given the nature of the RT technique and the
potentially confounding impact of RT dose levels, stratified analyses were
used when applicable. The van Elteren test,
24
the stratified extension for
Wilcoxon rank sum test, was used to compare dosimetric parameters after
stratifying by RT dose levels. OS, progression-free survival (PFS), time to
local failure (LF), and time to distant metastasis (DM) were calculated
from the date of random assignment to the date of failure or last follow-up.
The rates of OS and PFS were estimated using the Kaplan-Meier method,
and the distributions of OS and PFS were compared using the log-rank
test
25
stratified by RT dose levels.
26
The development of LF and DM was
analyzed by using the cause-specific competing risks analysis method,
27
with deaths without LF or DM as competing events. The rates of LF and
DM were estimated using cumulative incidence function,
27
and the dis-
tributions of LF and DM were compared using the log-rank test, stratified
by RT dose levels. Cox proportional hazards regressions were used to
evaluate the impact of RT technique and other factors on all outcomes after
stratifying by RT dose levels.
27
All adverse events were graded with
Common Terminology Criteria for Adverse Events (version 3) criteria. The
association between adverse events or treatment interruptions and RT
techniques was assessed with Cochran-Mantel-Haenszel statistics (strati-
fied by RT dose levels and cetuximab usage) and logistic regression models.
All statistical tests were two-sided and P , .05 was considered statistically
significant. SAS 9.4 software (SAS Institute, Cary, NC) was used for all
statistical analyses.
RESULTS
Treatment assignment and patient characteristics by RT technique
were summarized and compared (Table 1). Due to stratification,
the use of IMRT and 3D-CRT were similar in the 60- and 74-Gy
dose arms. Marginally more patients had stage IIIB/N3 disease in
the IMRT group than in the 3D-CRT group (30.3% v 38.6%,
P = .056). Positron emission tomography (PET) staging was used
more often in the IMRT group than in the 3D-CRT group (88.2% v
94.3%, P = .019). Patients treated with IMRT were less likely to
have completed high school or to have education beyond high
school (P = .01). Otherwise, the 3D-CRT and IMRT groups were
not different with respect to other baseline prognostic factors and
characteristics (P . .05).
There were substantial differences in dosimetry and target
volumes between the 3D-CRT and IMRT plans after adjusting for
RT dose levels (Table 2). The median planning treatment volume
(PTV) size was greater in the IMRT group than in the 3D-CRT
group (427 v 486 mL, P = .005), and the PTV/volume of lung ratio
was significantly bigger in the IMRT group (0.13 v 0.15, P = .013).
IMRT provided better PTV coverage by 100% of the prescription
dose (median, 94.8% v 95.1%; P = .058), whereas it had a slightly
lower minimum dose to the PTV (median, 55.2 v 53.4 Gy;
P , .001). After stratifying by RT dose levels and PTV quar tiles,
3D-CRTalso had marginally higher mean lung doses (median, 18.1 v
17.7 Gy; P = .088) and marginally hig her esophageal doses
(median, 27.6 v 25.6 Gy; P = .078) than IMRT. The lung V20 was
not different between g roups (median, 30.5% v 29.9% ; P = .297),
although IMRTwas associated w it h a larger lung V5 than 3 D-CRT
(median, 54.8% v 61. 6%; P , .001). The maximum dose de-
livered to nontarget tissue outside the PTV was also significantly
lower in patients treated with IMRT (median, 69.9 v 69.55 Gy; P =
.026). Heart doses ( V20, V40, and V60) were significantly lower
in patients treated with IMRT (P , .05), despite the volume of
heart inside the PTV was not different (median, 2.05 v 3.56 mL;
P = .183).
With a median follow-up of 21.3 months, 3D-CRT and IMRT
did not have different 2-year rates of OS, PFS, LF, and DM (Table 3).
The heart V5, V20, V40, V60, and the size of the PTV were sig-
ni
ficantly (P , .05) associated with OS in univariable analysis
(Appendix Table A1, online only). After adjusting for RT technique,
age, and percent PTV covered by 100% of the prescription dose
(Appendix Tabl e A 2, online only), site accrual volume, and PET
staging, the heart V40 remained significantly associated with OS
(hazard ratio, 1.012; 95% CI, 1.005 to 1.02; P , .001).
The severe adverse effect profile of 3D-CRT and IMRT were
comp a red , d efined as treatment-associated $ grade 3 events
ascopubs.org/journal/jco © 2016 by American Society of Clinical Oncology 57
3D-CRT Versus IMRT for Locally Advanced NSCLC
(Table 4). IMRT was associated with less $ grade 3 pneumonitis
than 3D-CRT (7.9% v 3.5%, P = .039). The rates of $ grade 3
esophagitis and dysphagia, weight loss, and cardiovascular toxicity
in both groups were not different (P . .05). Although the lung V5
was signi ficantly larger in patients treated with IMRT, it was not
associated with any kind of $ grade 3 toxicity (Appendix Table A3,
online only).
To better understand the impact of radiation technique on $
grade 3 pneumonitis, further statistical analysis was performed. On
multivariable analysis (Table 5), IMRT remained associated with
Follow-Up
Patients randomly allocated
(N = 544)
Enrollment
Allocation/Analysis
Treatment
Randomly allocated to standard dose
(60 Gy)
Excluded from analysis
Withdrew consent
• CBC done > 2 weeks before
registration
• Unable to confirm eligibility
• N3 disease based on contralateral
adenopathy
• Prior malignancy < 3 years before
registration
• Contralateral disease at
registration
• N3 disease based on supraclavicular
adenopathy
• Pancoast tumor
• Thoracotomy done < 3 weeks before
registration
(n = 15)
(n = 3)
(n = 2)
(n = 2)
(n = 2)
(n = 2)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
♦ Consolidation chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per protocol
• > 115% of therapy
• Chemotherapy not given
(n = 101)
(n = 14)
(n = 6)
(n = 7)
(n = 1)
(n = 22)
♦ RT delivery
• Received 60 Gy
• Received < 60 Gy
• Received > 60 Gy
• Did not receive RT
(n = 142)
(n = 5)
(n = 1)
(n = 3)
♦ Treatment terminated due to
toxicity
(n = 29)
Allocated to standard dose (60 Gy)
and analyzable
(n = 151)
♦ Concurrent chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per protocol
• > 115% of therapy
• Chemotherapy not given
(n = 110)
(n = 26)
(n = 7)
(n = 4)
(n = 1)
(n = 3)
Randomly allocated to high dose
(74 Gy) (n = 121)
♦ Excluded from analysis
• Withdrew consent
• PET done > 6 weeks before
registration
• Planned brachial plexus dose
exceeded 66 Gy
• Brain MRI done > 6 weeks
before registration
• Contralateral disease at
registration
• Initial diagnosis > 12 weeks before
registration
• Hemoglobin < 10.0 g/dL at
time of registration
• Pancoast tumor
• Positive supraclavicular node
• Resectable/operable disease
• Serum creatinine < 60 mL/min at
time of registration
(n = 14)
(n = 2)
(n = 2)
(n = 2)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
♦ Treatment terminated due to
toxicity
(n = 25)
♦ Concurrent chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
(n = 78)
(n = 20)
(n = 1)
(n = 4)
(n = 1)
(n = 3)
♦ Consolidation chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
(n = 60)
(n = 10)
(n = 3)
(n = 6)
(n = 2)
(n = 26)
♦ RT delivery
• Received 74 Gy
• Received < 74 Gy
• Did not receive RT
(n = 85)
(n = 20)
(n = 2)
Allocated to high dose (74 Gy) and
analyzable
(n = 107)
Randomly allocated to standard dose
(60 Gy) plus cetuximab
(n = 147)
♦ Excluded from analysis
• Withdrew consent
• CBC > 2 weeks before
registration
• Initial diagnosis > 12 weeks
before registration
• Metastatic disease at
registration
• Malignant pleural effusion present
at time of registration
• N3 disease based on supraclavicular
adenopathy
• Unable to confirm eligibility
• Pulmonary function tests done > 12
weeks before registration
(n = 10)
(n = 1)
(n = 3)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 1)
(n = 110)
(n = 22)
(n = 1)
(n = 2)
(n = 2)
(n = 74)
(n = 24)
(n = 5)
(n = 9)
(n = 1)
(n = 24)
(n = 73)
(n = 38)
(n = 2)
(n = 5)
(n = 19)
(n = 125)
(n = 5)
(n = 2)
(n = 5)
(n = 91)
(n = 28)
(n = 10)
(n = 4)
(n = 1)
(n = 3)
♦ Concurrent cetuximab delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• Cetuximab not given
♦ Consolidation chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
♦ Consolidation cetuximab delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• Cetuximab not given
♦ RT delivery
• Received 60 Gy
• Received < 60 Gy
• Received > 60 Gy
• Did not receive RT
♦ Concurrent chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
Allocated to standard dose (60 Gy) plus
cetuximab and analyzable
(n = 137)
♦ Treatment terminated due to
toxicity
(n = 53)
Randomly allocated to high dose
(74 Gy) plus cetuximab
(n = 110)
♦ Excluded from analysis
• Withdrew consent
• Planned brachial plexus dose
exceeded 66 Gy
• Brain MRI done > 6 weeks before
registration
• COPD required hospitalization
within 30 days before registration
• Hemoglobin < 10.0 g/dL at time of
registration
• T2/N0 disease at time of
registration
(n = 10)
(n = 1)
(n = 4)
(n = 2)
(n = 1)
(n = 1)
(n = 1)
♦ Treatment terminated due to
toxicity
(n = 46)
♦ Concurrent chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
(n = 68)
(n = 18)
(n = 6)
(n = 1)
(n = 2)
(n = 5)
♦ Concurrent cetuximab delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• Cetuximab not given
(n = 78)
(n = 17)
(n = 1)
(n = 4)
♦ Consolidation chemotherapy delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• > 115% of therapy
• Chemotherapy not given
(n = 52)
(n = 18)
(n = 3)
(n = 6)
(n = 1)
(n = 20)
♦ Consolidation cetuximab delivery
• 85% to 115% of therapy
• < 85% of therapy, per protocol
• 70% to < 85% of therapy, not per
protocol
• < 70% of therapy, not per
protocol
• Cetuximab not given
(n = 45)
(n = 30)
(n = 2)
(n = 3)
(n = 20)
♦ RT delivery
• Received 74 Gy
• Received < 74 Gy
• Did not receive RT
(n = 75)
(n = 22)
(n = 3)
Allocated to high dose (74 Gy) plus
cetuximab and analyzable
(n = 100)
( n = 166)
Lost to follow-up
(n = 7)
Lost to follow-up
(n = 2)
Lost to follow-up
(n = 3)
Lost to follow-up
(n = 0)
Fig 1. CONSORT diagram of NRG Oncology clinical trial RTOG 0617. COPD, chronic obstructive pulmonary disease; MRI, magnetic resonance imaging; PET, positron
emissions tomography; RT, radiation therapy.
58
© 2016 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY
Chun et al
a statistically significant reduction in pneumonitis risk (odds ratio,
0.41; 95% CI, 0.17 to 0.99; P = .046), whereas stage IIIB disease
and lung V20 were associated with increased pneumonitis risk
(P , .05).NeitherthelungV5northemeanlungdosewas
significantly associated (P . .05) with $ grade 3 pneumonitis
(Appendix Table A4, online o nly).
Treatment interruptions and the administration of full doses
of concurrent chemotherapy were also compared between 3D-CRT
and IMRT (Appendix Table A5, online only). There were similar
rates of treatment interruptions due to adverse effects or illness
(17.7% v 17.5%, P = .969), and administration of full doses of
concurrent carboplatin (area under the curve, 2) and paclitaxel
(45 mg/m
2
) in the 3D-CRT and IMRT groups, respectively (70.1% v
66.7%, P =.388).
DISCUSSION
This secondary analysis of radiation technique in NRG On-
cology clinical trial RTOG 0617 aimed to change clinical
practice by clar ifying t he value of IMRT for locally advanced
NSCLC b ased on findings from a large prospective multi-
institutional tr ial. We found that patients treated with IMRT
in NRG Oncology clinical trial RTOG 0617 did not have dif-
ferent 2-year survival outcomes from 3D-CRT despite IMRT
having worse prognostic factors, such as larger tumors and more
American Joint Committee on Cancer stage IIIB disease.
Nevertheless, IMRT achieved equivalent lung V20s and better
PTV coverage than 3D-CRT. In turn, IMRT was associated
with a s ig ni ficantreductioninsevere$ grade 3 pneumonitis.
Moreover, IMRTwas able to reduce radiation doses delivered to the
heart, and heart doses were highly associated with OS on multi-
variable analysis. On the basis of these findings and in conjunction
with a recent study that showed IMRT to be associated with
improved quality of life in NRG Oncology clinical trial RTOG
0617,
28
we advocate for the routine use of IMRT in locally ad-
vanced NSCLC to reduce both severe lung toxicity and doses of
radiation delivered to the heart.
Despite larger tumors, IMRT resulted in significantly lower
rates of $ gra de 3 pneumon itis. The lung V20 is a classi c and
the most frequently described dosimetr ic parameter believed to
be a threshold dose that predicts probability of lung injur y.
29
However, a number of retrospective analyses have correlated
radiation pneumonitis w ith low-dose baths, such as the V5.
30-32
Low doses have not been found to predict pneumonitis in
patients w ith medically inoperable early-stage NSCLC treated
with stereotactic RT in a prospective Radiation Therapy On-
cology G ro u p tria l .
33
Partly as a consequence of using more
beam entry points, one of the hallmarks of IMRT is its ability to
improve conformity of the intermediate- and high-dose region
by spreading a low dose over a larger area, thereby increasing
parameters such as the lung V5. Despite significantly g reater
lung V5 values, IMRT was associated with a better lung toxicity
profile than 3D-CRT. The findings of this study provide no
suggestion that the lung V5 is a predictor of toxicity in the RTof
locally advanced NSCLC. Moreover, these results argue against
using the lung V5 for IMRT plan optimization because an at-
tempt to lower the V5 can potentially lead to less conformity of
the hig h-dose region and an inability to reduce intermediate
dose (V 20), both of which were important objectives confirmed
in this study.
In this study, patients treated with IMRT seem to have worse
socioeconomic circumstances than those treated with 3D-CRT.
Although socioeconomic variables such as income, health in-
surance status, and access to specialized care were not collected in
NRG Oncology clinical trial RTOG 0617, we observed significant
Table 1. Demographic and Clinical Characteristics
Characteristic 3D-CRT, % IMRT, % P
No. of patients 254 228
Radiation therapy dose level, Gy .637
60 57.1 59.2
74 42.9 40.8
Cetuximab assigned .953
Yes 47.6 47.4
No 52.4 52.6
Median age, years (range) 64 (37-82) 64 (38-83) .903*
Sex .966
Male 59.8 59.6
Female 40.2 40.4
Race .19
Native American 0.8 0
Asian 1.2 4.4
Black or African American 10.2 9.6
Pacific Islander or Native Hawaiian 0.4 0
White 86.2 85.1
Unknown 1.2 0.9
Ethnicity .357
Hispanic or Latino 3.9 1.8
Non-Hispanic or Latino 92.9 94.7
Unknown 3.1 3.5
Education status .01
Less than high school 16.1 12.3
High school 34.7 44.3
More than high school 41.7 30.7
Other/unknown 7.5 12.7
Zubrod performance status .266
0 59.8 54.8
1 40.2 45.2
Positron emission tomography staging .019
Yes 88.2 94.3
No 11.8 5.7
Histology .241
Squamous carcinoma 46.5 39.9
Adenocarcinoma 37 42.5
Large cell undiffe rentiated 3.5 1.8
Non–small cell not otherwise specified 13 15.8
T stage .331
Unknown 2.4 0.9
T1 19.3 16.2
T2 34.6 33.3
T3 19.3 21.9
T4 24.4 27.6
N stage .088
N0 4.7 7.5
N1 5.1 3.9
N2 83.1 75.9
N3 7.1 12.7
AJCC stage group .056
IIIA 69.7 61.4
IIIB 30.3 38.6
Abbreviations: 3D-CRT, three-dimensional conformal external beam radiation
therapy; AJCC, American Joint Commission on Cancer; IMRT, intensity-
modulated radiation therapy; N, clinical node stage; T, clinical tumor stage.
*P value from t test; otherwise, all other P values from x
2
test.
ascopubs.org/journal/jco © 2016 by American Society of Clinical Oncology 59
3D-CRT Versus IMRT for Locally Advanced NSCLC
differences in education status in patients treated with IMRT and
3D-CRT. Patients treated with IMRT were less likely to have
completed high school or attain education beyond high school. We
speculate that worse socioeconomic circumstances may account
for the larger tumor volumes and more advanced-stage tumors
seen in the IMRT group possibly due to barriers to health care
access, which leads to later diagnoses. Despite indications that
patients in the IMRT group had worse socioeconomic circum-
stances, these patients had a notably better severe toxicity profile,
and OS was not different from that of the 3D-CRT group. IMRT
could have mitigated a possible negative impact that socio-
economic status might have otherwise had on survival and co-
ordination of care.
The dose of radiation to the heart was shown to be an im-
portant predictor of survival in NRG Oncology clinical trial RTOG
0617,
23
and this secondary analysis shows that IMRT is able to
significantly reduce radiation doses delivered to the heart. Of note,
IMRT did not have different survival rates from 3D-CRT despite
treatment of larger and more advanced-stage tumors. Although this
study was not designed to determine the survival impact of radiation
doses to the heart, IMRT possibly mitigated the potential adverse
survival effect conferred by larger and more advanced tumors by
reducing radiation doses to the heart, such as the V40, which ac-
counts for similar survival between the 3D-CRT and IMRT groups.
However, longer follow-up may be needed to capture differences in
cardiac toxicity associated with IMRT. Although institutional accrual
status was previously shown to be associated with survival outcomes
in NRG Oncology clinical trial RTOG 0617 patients,
34
the heart V40
remains significantly associated with OS on multivariable analysis,
even with adjustment for institutional accrual status. Further
pending analyses of heart doses in NRG Oncology clinical trial
RTOG 0617 may provide critical insights into the effect of radiation
doses on specific anatomic regions of the heart as well as pertinent
heart radiation dose constraints.
Although survival outcomes appear to be equivalent be-
tween IMRT and 3D-CRT in this early analysis of outcomes in
Table 2. Dosimetric Factors of 3D-CRT Versus IMRT
Dosimetric Factor
3D-CRT IMRT
PMedian Q1-Q3 Median Q1-Q3
PTV volume, mL 426.7 298.1-586.5 486.2 347.6-677.3 .005*
Volume of lung excluding CTV, mL 3,331.4 2,676.7-4,045.0 3,215.7 2,754.6-4,020.0 .779*
PTV volume:lung volume ratio 0.13 0.09-0.19 0.15 0.10-0.21 .013*
Minimum dose to PTV, Gy 55.2 49.8-60.2 53.4 48.0-57.3 , .001†
Maximum dose to PTV, Gy 68.8 66.1-80.8 70.2 66.1-80.9 .256†
Dose to cover 95% of PTV, Gy 60.8 60.0-72.3 60.7 60.0-73.0 .088†
PTV covered by 100% Rx dose, % 94.8 87.0-96.4 95.1 92.1-97.0 .058*
Mean lung dose, Gy 18.1 15.4-20.6 17.7 14.4-20.1 .088†
Volume of lung, %
V5 54.8 43.3-65.9 61.6 52.1-70.4 , .001†
V20 30.5 25.3-35.1 29.9 24.0-34.7 .297†
Mean esophagus dose, Gy 27.6 22.1-32.8 25.6 20.2-32.6 .078†
Volume of esophagus, %
V20 47.6 39.4-56.9 46.8 36.7-56.7 .466†
V60 19.7 5.2-30.4 18.4 3.6-29.3 .927†
Volume of heart, %
V20 23.5 7.8-46.0 19.3 5.2-36.5 .049†
V40 11.4 1.7-25.9 6.8 0.6-15.5 .003†
V60 2.4 0.0-8.3 1.4 0.0-5.0 .045†
Volume of heart inside PTV, mL 2.05 0.00-16.46 3.56 0.00-16.73 .183*
Maximum dose outside PTV, Gy 69.9 66.3-80.8 69.55 65.6-79.9 .026†
Abbreviations: 3D-CRT, three-dimensional conformal external beam radiation therapy; CTV, clinical target volume; IMRT, intensity-modulated radiation therapy; PTV,
planning treatment volume; Q1, quartile 1; Q3, quartile 3; Rx, prescription; V, volume receiving radiation dose.
*P value from Wilcoxon test.
†P value from Wilcoxon test stratified by radiation therapy dose level (60 v 74 Gy).
Table 3. Outcomes at 2 Years by Radiation Therapy Technique
Outcome 3D-CRT, % (95% CI) IMRT, % (95% CI) P
Overall survival 49.4 (42.9 to 55.5) 53.2 (46.4 to 59.6) .597
Progression-free survival 27.0 (21.5 to 32.7) 25.2 (19.7 to 31.1) .595
Local failure 37.1 (31.0 to 43.1) 30.8 (24.8 to 36.9) .498
Distant metastases 49.6 (43.2 to 55.8) 45.9 (39.2 to 52.3) .661
NOTE. P values from a two-sided log-rank test stratified by radiation therapy
dose level (60 v 74 Gy).
Abbreviations: 3D-CRT, three-dimensional conformal external beam radiation
therapy; IMRT, intensity-modulated radiation therapy.
Table 4. CTCAE $ Grade 3 Radiation-Related Adverse Events of
3D-CRT and IMRT
$ Grade 3 Toxicity 3D-CRT, % (No.) IMRT, % (No.) P
No. of patients 254 228
Pneumonitis 7.9 (20) 3.5 (8) .039
Esophagitis/dysphagia 15.4 (39) 13.2 (30) .534
Weight loss 2.8 (7) 3.9 (9) .419
Cardiovascular 8.3 (21) 4.8 (11) .131
NOTE. P values from a Cochran-Mantel-Haenszel test stratified by radiation
therapy dose level (60 v 74 Gy) and cetuximab random assignment.
Abbreviations: 3D-CRT, three-dimensional conformal external beam radiation
therapy; CTCAE, Common Terminology Criteria for Adverse Events (version 3);
IMRT, intensity-modulated radiation therapy.
60 © 2016 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY
Chun et al