Original Study

Adjuvant Chemotherapy Guided by Molecular

Profiling and Improved Outcomes in Early Stage,
NoneSmall-Cell Lung Cancer
Gavitt A. Woodard,1 Sue X. Wang,1 Johannes R. Kratz,1 Clara T. Zoon-Besselink,1
Chun-Yuan Chiang,2 Matthew A. Gubens,3 Thierry M. Jahan,3 Collin M. Blakely,3
Kirk D. Jones,4 Michael J. Mann,1 David M. Jablons1
Abstract
A 14-gene molecular assay was performed prospectively on stage IA, IB, and IIA nonsquamous nonesmall-cell lung
cancer patients after surgical resection. Five-year disease-free survival was 93.8% among low-risk patients, 91.7%
among high-risk patients treated with adjuvant chemotherapy, and 48.9% among molecular high-risk patients who
did not receive adjuvant chemotherapy (P [ .004). Using molecular testing in addition to National Comprehensive
Cancer Network criteria might better identify patients who benefit from adjuvant chemotherapy.
Introduction: Many early stage nonesmall-cell lung cancer (NSCLC) patients who are not considered candidates for
adjuvant treatment according to current guidelines do harbor occult metastasis, and have disease recurrence despite
complete resection. Although National Comprehensive Cancer Network (NCCN) guidelines suggest clinicopathologic
characteristics to identify high-risk patients for adjuvant intervention, molecular profiling more accurately predicts
5-year survival. Early evidence of clinical benefit from application of this molecular-based management strategy,
however, has not been reported. Patients and Methods: An internationally validated, prognostic, 14-gene quantitative
polymerase chain reaction expression assay was used to stratify risk prospectively in 100 consecutive patients with
stage IA, IB, and IIA nonsquamous NSCLC. KaplaneMeyer estimates, log rank analysis, and Cox regression were
used to compare disease-free survival (DFS) between high-risk patients who did or did not elect adjuvant chemo-
therapy. Results: Forty-eight patients (48%) were deemed high-risk according to molecular testing and 36 (36%) met
NCCN high-risk criteria; risk designations were discordant in 34 (34%) of all patients. Estimated 5-year DFS was 48.9%
among molecular high-risk patients who did not undertake adjuvant chemotherapy, 93.8% among untreated molecular
low-risk patients, and 91.7% in molecular high-risk patients who did undergo chemotherapy (P ¼ .004). In contrast, DFS
was only 75.2% in untreated NCCN low-risk patients, and 61.9% in untreated NCCN high-risk patients (P ¼ .183).
Conclusion: This prospective, nonrandomized study provides initial evidence that high-risk designation according to
the 14-gene prognostic assay also predicts benefit from adjuvant chemotherapy for very early stage NSCLC, and
further supports the superiority of molecular stratification over current NCCN criteria at identifying high-risk patients.

Clinical Lung Cancer, Vol. 19, No. 1, 58-64 ª 2017 Elsevier Inc. All rights reserved.
Keywords: Cisplatin-based adjuvant therapy, NCCN guidelines, Prognostic genetic signature, Risk stratification,
Tumor genetic profile

Introduction stage I and II NSCLC have a 5-year overall survival rate of only
Nonesmall-cell lung cancer (NSCLC) remains a deadly disease 53% despite a complete surgical resection. Disease recurrence is
with poor prognosis, even among early stage patients. Patients with responsible for a large percentage of all mortality,1 suggesting that as

1
Division of Adult Cardiothoracic Surgery, Department of Surgery, University of Submitted: Mar 7, 2017; Revised: May 2, 2017; Accepted: May 16, 2017; Epub: May
California, San Francisco, San Francisco, CA 31, 2017
2
Hong Kong Institute of Cell and Molecular Medicine, Hong Kong
3
Division of Hematology and Oncology, Department of Medicine Address for correspondence: David M. Jablons, MD, Department of Surgery,
4
Department of Pathology, University of California, San Francisco, San Francisco, CA University of California at San Francisco, 500 Parnassus Ave, Room MUW-424, San
Francisco, CA 94143-1724
Fax: 415-353-9530; e-mail contact: david.jablons@ucsf.edu

58 - Clinical Lung Cancer January 2018

1525-7304/$ - see frontmatter ª 2017 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.cllc.2017.05.015
many as half or more of these patients harbor occult metastasis
despite their “early stage” designation. Multiple large-scale, pro-
spective studies have shown that survival can be improved and
recurrence rates lowered in patients at a higher risk of recurrence (ie,
certain stage IIA and all stage IIB and III patients) with adjuvant
cisplatin-based chemotherapy.2-5 Although 30% to 50% of stage I
and IIA patients die within 5 years of resection, many do not receive
adjuvant treatment because of a failure of previous prospective
studies, all on the basis of conventional Tumor, Node, Metastases
(TNM) risk stratification, to show a benefit in these patients.
Improved methods are therefore needed to discriminate between
patients in the earliest stages of NSCLC who have likely been cured
using surgical resection alone and those who are truly at high risk of
recurrence and likely to benefit from adjuvant chemotherapy. Na-
tional Comprehensive Cancer Network (NCCN) guidelines
currently recommend that this discrimination be on the basis of a
combination of TNM stage (IB and IIA) combined with a list of 6
clinicopathologic criteria (poorly differentiated tumors, vascular
invasion, wedge resection, tumor size > 4 cm, visceral pleural
involvement, and unknown lymph node status because of lack of
sampling).6 These criteria, however, have never been validated to
identify patients who benefit from adjuvant chemotherapy; they
might lead to overtreatment of some patients with low risk of occult
metastasis and undertreatment of others with biologically aggressive
tumors prone to metastasis and recurrence.7,8
A Clinical Laboratory Improvement Amendments (CLIA)-certi-
fied, 14-gene, quantitative polymerase chain reaction (PCR)-based
expression assay has been clinically validated in approximately 1500
patients, and was reported to predict mortality risk more accurately
than NCCN clinicopathologic criteria in early stage, nonsquamous
NSCLC, in large-scale, independent, blinded international
studies9,10 as well as prospectively in a clinical setting.11 This mo-
lecular stratification might provide a molecular precision approach
to identifying early stage, nonsquamous NSCLC patients who,
because of to their more aggressive molecular phenotype, are most
likely to harbor micrometastatic disease and most likely to benefit
from adjuvant chemotherapy. We conducted a prospective, non-
randomized, single-institution pilot study to assess whether plat-
inum doublet adjuvant chemotherapy is associated with improved
outcomes in patients identified as high-risk using 14-gene molecular
stratification.
Patients and Methods
A commercially available, CLIA-certified 14-gene quantitative
PCR expression assay (Encore Clinical, Brisbane, CA) was previ-
ously developed and internationally validated to risk-stratify patients
into low-, intermediate-, and high-risk groups for risk of 5-year
mortality after complete surgical resection. The complete methods
for RNA extraction from formalin-fixed, paraffin-embedded tumor
tissue specimens and subsequent quality controlled, quantitative
reverse transcription PCR, as well as a complete description of the
analytical validation of the assay, have been previously reported.9,12
This molecular prognostic test has been conducted prospectively at
an independent CLIA-certified laboratory on the formalin-fixed,
paraffin-embedded tumor tissues of all patients who undergo
resection of nonsquamous NSCLC at the University of California,
San Francisco (UCSF) since 2011, to confirm, in a routine clinical
setting and in a prospective fashion, the assay’s prognostic value and
assess its possible predictive clinical utility. The first 100 consecutive
patients with stage IA, IB, or IIA NSCLC resected at UCSF be-
tween 2011 and 2015 were included in the current analysis. Patients
with previous lung cancers, patients with multiple simultaneous
lung cancers, and patients who received neoadjuvant chemotherapy
were excluded.
Tumor samples were sent for molecular testing after finalization
of the surgical pathology report. All patients were discussed at a
multidisciplinary tumor board, and high-risk patients, on the basis
of either molecular testing or NCCN criteria, were referred to a
thoracic oncologist for discussion of adjuvant chemotherapy. Mo-
lecular intermediate- as well as high-risk patients were considered to
be at higher risk of recurrence, and were therefore grouped together
and designated as molecular high-risk. The decision to undergo
adjuvant chemotherapy was made on an individual basis by all
patients and their oncologists. Patients were followed prospectively
with the primary end point being lung cancer recurrence and
disease-free survival (DFS). Disease recurrence was chosen as the
primary end point because it is closely associated with 5-year
mortality.
Patient demographic information and longitudinal follow-up
data were collected using manual review of the clinical electronic
medical record, pathology reports, progress notes, and outside re-
cords. This study was approved by the UCSF Committee on Hu-
man Research, and all patients signed consent for research
participation and molecular tumor testing.
Statistical analysis was performed using SPSS Statistics Standard
(version 22.0 for Macintosh; IBM Corp, Armonk, NY). Unpaired t
tests were performed for continuous variables. We used c2 tests for
categorical variables if expected frequencies were  5, and a 2-tailed
Fisher exact probability test for categorical variables if expected
frequencies were < 5. A 2-sided a < 0.05 was defined as statistically
significant. Stratified KaplaneMeier (KM) analysis with a censored
data set and the log rank test were used to evaluate associations
between molecular and NCCN risk stratification and the end point
of DFS. Univariate and multivariate Cox proportional hazard
regression modeling were performed with the covariates of age, sex,
smoking history, pathologic stage, type of surgical resection, NCCN
risk, molecular risk, and adjuvant chemotherapy.
Results
Molecular testing was performed prospectively on 100 consecu-
tive stage IA, IB, and IIA nonsquamous NSCLC patients after
complete surgical resection between 2011 and 2015. Median age
was 67.7 years, 40% were male, and 69% were former or current
smokers. Fifty-eight (58%) were stage IA, 32 (32%) stage IB, and
10 (10%) stage IIA. Median follow-up was 23 months, with 11
recurrences (11%) during the follow-up period. Most patients 64
(64%) underwent lobectomy; only 1 (1%) patient required pneu-
monectomy, 1 (1%) required bilobectomy, and the remainder 34
(34%) underwent segmental resection (Table 1).
Molecular prognostic testing stratified 52 patients (52%) as low
risk for 5-year mortality and a total of 48 patients (48%) as either
intermediate risk (n ¼ 22) or high risk (n ¼ 26). NCCN criteria
identified 36 patients (36%) as candidates for adjuvant chemo-
therapy, and a total of 13 (13%) patients received adjuvant
Clinical Lung Cancer January 2018 - 59
 Adjuvant Chemotherapy in Molecular High-Risk Patients Improves Early Stage NSCLC Outcomes

Table 1 Clinical Characteristics and Molecular Risk Stratification

Molecular Low Molecular High

Characteristic All Patients (n [ 100) Risk (n [ 52) Risk (n [ 48) P
Mean Age ± SE, y 67.7  1.1 67.8  1.4 67.6  1.8 .9358
Men 40 (40) 19 (37) 21 (44) .5415
Women 60 (60) 33 (63) 27 (56) .5415
Asian 16 (16) 8 (15) 8 (17) .9999
History of Smoking 69 (69) 33 (63) 36 (75) .2799
Pathologic Stage .0001
IA 58 (58) 40 (77) 18 (38)
IB 32 (32) 8 (15) 24 (50)
IIA 10 (10) 4 (8) 6 (13)
Surgical Resection .0187
Segmental resection 34 (34) 24 (46) 10 (20)
Lobectomy 64 (64) 28 (54) 36 (75)
Bilobectomy 1 (1) 0 (0) 1 (2)
Pneumonectomy 1 (1) 0 (0) 1 (2)
NCCN Risk Stratification .0017
Low-risk features 64 (64) 41 (79) 23 (48)
High-risk featuresa 36 (36) 11 (21) 25 (52)
Adjuvant Chemotherapy <.0001
None 87 (87) 52 (100) 35 (73)
Received adjuvant chemotherapy 13 (13) 0 (0) 13 (27)
Median Follow-Up (IQR), mo 23 (10-42) 19 (10-39) 28 (9-42) .4014
5-Year DFS Rate, % e 93.8 58.8 .0060

Data are presented as n (%) except where otherwise stated.

Abbreviations: IQR ¼ interquartile range; NCCN ¼ National Comprehensive Cancer Network; SE ¼ standard error.
a
National Comprehensive Cancer Network guidelines recommend adjuvant chemotherapy for stage IB and node-negative IIA tumors with “high risk” clinicopathologic characteristics: poorly
differentiated tumors, vascular invasion, wedge resection, tumors > 4 cm, visceral pleural involvement, and incomplete lymph node sampling.

Figure 1 Disease-Free Survival (DFS) Analysis According to Molecular Risk and National Comprehensive Cancer Network (NCCN)
Criteria. (A) Molecular Stratification. Molecular Low-Risk Patients Had 5-Year DFS of 93.8%, Which Was Significantly Better
Than 58.8% DFS of Molecular High-Risk Patients (Log Rank P [ .006). (B) NCCN Stratification. Recurrence Rates Did Not
Differ Significantly Among Patients With NCCN High-Risk Clinicopathologic Features (5-Year DFS of 73.1%) Versus Patients
Without NCCN High-Risk Features (DFS of 75.7%; Log Rank P [ .112)

60 - Clinical Lung Cancer January 2018

 Gavitt A. Woodard et al
Table 2 Cox Univariate and Multivariate Regression Analysis of Recurrence: All Patients (n [ 100)

Single Variable Regression Multivariable Regression

Variable HR 95% CI P HR 95% CI P
Age
<65 y 1 1
65 y 2.83 0.61-13.14 .185 1.05 0.17-6.58 .957
Sex
Male 1 1
Female 1.64 0.49-5.53 .423 2.59 0.61-11.09 .198
Smoking History
Never smoker 1 1
Former or current smoker 1.10 0.38-3.18 .866 0.70 0.15-3.62 .728
Pathologic Stage
IA 1 1
IB 2.78 0.74-10.40 .129 1.22 0.05-28.03 .902
IIA 3.22 0.58-17.78 .181 1.75 0.25-12.14 .573
Surgical Resection
Segmental resection 1 1
Lobectomy or greater 6.23 0.78-49.71 .084 3.31 0.35-31.65 .298
NCCN Risk
Low risk 1 1
High risk 2.54 0.77-8.35 .125 2.32 0.23-23.18 .475
Molecular Risk
Low risk 1 1
High risk 3.18 1.14-8.89 .028 8.70 1.04-72.92 .046
Adjuvant Chemotherapy
No adjuvant chemotherapy 1 1
Adjuvant chemotherapy 0.55 0.07-4.28 .564 0.09 0.01-1.06 .055

Abbreviations: HR ¼ hazard ratio; NCCN ¼ National Comprehensive Cancer Network.

chemotherapy, all of whom were molecular high-risk. Refusal of
adjuvant chemotherapy was most often related to patient preference.
NCCN guideline recommendations for adjuvant chemotherapy and
recommendations for adjuvant chemotherapy on the basis of tumor
molecular prognosis were discordant in 34 (34%) of all patients
(Table 1).
There were no significant differences in age, sex, race, smoking
history, or follow-up times between molecular low-risk patients and
molecular high-risk patients. Most molecular low-risk patients 40 of
52 (77%) were stage IA, 8 of 52 (15%) were stage IB, and 4 of 52
(8%) were stage IIA, whereas molecular high-risk tumors were 18 of
48 (38%) stage IA, 24 of 48 (50%) stage IB, and 6 of 48 (15%)
stage IIA. Although, as expected, there was correlation between
TNM stage and molecular profile (more stage IA tumors tended to
be molecular low-risk [40 of 58; 69%], whereas more stage IB and
IIA tumors tended to be molecular high-risk [30 of 42; 71%]),
molecular risk did not simply recapitulate stage, because molecular
low- and high-risk patients were identified in all stages.
At a median follow-up of 23 months, there was a recurrence rate
of 11% among all patients. Among patients who did have disease
recurrence, the median time to recurrence was 19 months. Almost
all recurrences (10 of 11; 91%), however, occurred in molecular
high-risk patients. This translated to a 21% recurrence rate among
high-risk patients, which was significantly greater than the 2%
recurrence rate among molecular low-risk patients (Fisher exact test,
P ¼ .003). Also consistent with the molecular assay’s documented
success in discriminating high- and low-risk patients, the KM
estimate of 5-year DFS was significantly higher among low-risk
patients (93.8%) than among all molecular high-risk patients
(58.8%; log rank P ¼ .006; Figure 1A). Multivariate proportional
hazards modeling also suggested that, among risk factors tested, only
molecular high-risk remained a significant predictor of recurrence
when controlling for other variables, including NCCN risk category
and clinical stage (Table 2).
Furthermore, among molecular high-risk patients, adjuvant
chemotherapy was associated with a significant difference in DFS
(Table 3). Molecular high-risk patients who received adjuvant
chemotherapy had outcomes that were similar to the low-risk pa-
tients (5-year DFS of 91.7%), whereas high-risk patients who were
not treated with chemotherapy were found to have 5-year DFS of
only 48.7% (log rank P ¼ .004; Figure 2A). Although the 48
molecular high-risk patients did not represent a large enough
cohort for robust proportional hazards modeling, a multivariate
model that adjusted for risk factors with the greatest numerical
hazard ratio (HR) in univariate analysis of the entire study popu-
lation revealed a consistent trend toward reduction of recurrence in

Clinical Lung Cancer January 2018 - 61

 Adjuvant Chemotherapy in Molecular High-Risk Patients Improves Early Stage NSCLC Outcomes

Table 3 Molecular High-Risk Patients Stratified According to Adjuvant Chemotherapy

Molecular High Adjuvant Chemotherapy No Adjuvant

Variable Risk (n [ 48) (n [ 13) Chemotherapy (n [ 35) Pa
Mean Age ± SE, y 67.6  1.8 63.1  3.9 69.3  1.9 .2409
Men 21 (44) 8 (62) 13 (37) .1923
Women 27 (56) 5 (38) 22 (63) .1923
Asian 8 (17) 2 (15) 6 (17) .9999
History of Smoking 36 (75) 9 (69) 27 (75) .7101
Pathologic Stage .0780
IA 18 (38) 2 (15) 16 (46)
IB 24 (50) 10 (77) 14 (40)
IIA 6 (13) 1 (8) 5 (14)
Surgical Resection .2403
Segmental resection 10 (20) 1 (8) 9 (26)
Lobectomy 36 (75) 12 (92) 24 (69)
Bilobectomy 1 (2) 0 (0) 1 (3)
Pneumonectomy 1 (2) 0 (0) 1 (3)
NCCN Risk Stratification .0524
Low risk features 23 (48) 3 (23) 20 (57)
High risk featuresb 25 (52) 10 (77) 15 (43)
Median Follow-Up (IQR), mo 28 (9-42) 34 (12-42) 27 (9-37) .7424
5-Year DFS Rate, % 58.8 91.7 48.9 .0040

Data are presented as n (%) except where otherwise stated.

Abbreviations: IQR ¼ interquartile range; NCCN ¼ National Comprehensive Cancer Network; SE ¼ standard error.
a
Adjuvant chemotherapy versus no adjuvant chemotherapy.
b
National Comprehensive Cancer Network guidelines recommend adjuvant chemotherapy for stage IB and node-negative IIA tumors with “high risk” clinicopathologic characteristics: poorly
differentiated tumors, vascular invasion, wedge resection, tumors > 4 cm, visceral pleural involvement, and incomplete lymph node sampling.

Figure 2 Disease-Free Survival (DFS) Analysis on the Basis of Adjuvant Chemotherapy. (A) Molecular Stratification. There Were No
Molecular Low-Risk Patients Treated With Adjuvant Chemotherapy. Molecular Low-Risk Patients Had 5-Year DFS of 93.8%,
and High-Risk Patients Treated With Adjuvant Chemotherapy Had a Similar 5-Year DFS of 91.7%. In Contrast, High-Risk
Patients Who Did Not Receive Chemotherapy Had Frequent Disease Recurrence and Had 5-Year DFS of Only 48.9% (Log Rank
P [ .004). (B) National Comprehensive Cancer Network (NCCN) Stratification. There Were 3 NCCN Low-Risk Patients Treated
With Adjuvant Chemotherapy; None Had Disease Recurrence (Longest Disease-Free Follow-Up of 40 Months). DFS Was
75.2% Among NCCN Low-Risk Patients Who Did Not Receive Chemotherapy, 90.0% Among NCCN High-Risk Patients Treated
With Adjuvant Chemotherapy, and 61.9% Among NCCN High-Risk Patients Who Were Not Treated With Adjuvant
Chemotherapy (Log Rank P [ .183)

62 - Clinical Lung Cancer January 2018

 Gavitt A. Woodard et al
Table 4 Cox Univariate and Multivariate Regression Analysis of Recurrence: Molecular High-Risk Patients (n [ 48)

Single Variable Regression Multivariable Regression

Variable HR 95% CI P HR 95% CI P
Age
<65 y 1 1
65 y 1.93 0.40-9.32 .411 1.06 0.41-2.71 .907
Sex
Male 1 1
Female 1.39 0.72-2.69 .329 1.80 0.85-3.79 .122
Pathologic Stage
IA 1 1
IB 0.93 0.23-3.74 .915 0.63 0.21-1.92 .419
IIA 1.44 0.25-8.36 .687 1.15 0.46-2.91 .761
Surgical Resection
Segmental resection 1 1
Lobectomy or greater 1.79 0.23-14.20 .581 2.07 0.23-18.58 .516
Adjuvant Chemotherapy
No adjuvant chemotherapy 1 1
Adjuvant chemotherapy 0.26 0.03-2.03 .197 0.12 0.01-1.42 .092

Abbreviation: HR ¼ hazard ratio.

high-risk patients who received adjuvant treatment (HR, 0.12;
P ¼ .09; Table 4).
In contrast, NCCN high-risk criteria were less successful than the
molecular assay at prognostic differentiation of high- and low-risk
patients, as evidenced by: (1) poorer discrimination of DFS
among all NCCN high- and NCCN low-risk patients (log rank P ¼
.112; Figure 1B); (2) an estimated DFS of only 75.2% among
NCCN low-risk patients who, fully consistent with current guide-
lines, did not receive chemotherapy (log rank P ¼ .183; Figure 2B);
and (3) failure of NCCN risk category to achieve significance, as did
the molecular assay, as a predictor of outcome in either univariate or
multivariate proportional hazards modeling (Table 2).
Discussion
Adjuvant chemotherapy has been shown to improve survival
among early-stage NSCLC patients who are at greatest risk of 5-
year mortality (ie, TNM stages II and III), and therefore who
are at greatest risk of harboring occult micrometastatic disease.2-4
Those studies, however, failed to yield conclusive evidence of
benefit from adjuvant intervention in stage I and certain stage IIA
patients. Current NCCN guidelines therefore base recommenda-
tions for adjuvant chemotherapy in these patients on histopatho-
logic and clinical criteria for which there is little or no clinical
evidence of benefit. Furthermore, as many as 30% of stage IA
patients will also eventually succumb to occult metastasis that
might have been eradicated by postoperative chemotherapy; none
of those stage IA patients are recommended to undergo that
potentially life-saving chemotherapy on the basis of current
NCCN criteria. Many stage IB and IIA patients with an even
higher risk of early mortality currently refuse adjuvant interven-
tion, possibly influenced by lackluster data from past clinical trials
in which risk stratification was accomplished using conventional
staging criteria alone.
In this prospective, single-institution pilot study, the clinical
utility of molecular prognostic testing of early stage NSCLC to
supplement pathologic stage and NCCN guidelines was shown in
the context of adjuvant chemotherapy recommendations. This
study provides preliminary evidence that molecular testing of early-
stage nonsquamous NSCLC, followed by adjuvant chemotherapy in
molecular high-risk patients, might prevent a significant number
of recurrences and improve outcomes. As has been shown
previously,9-11 molecular risk scores better differentiated prospective
recurrence rates than did NCCN risk criteria in our cohort of 100
patients. Perhaps more important, molecular high-risk patients who
received adjuvant chemotherapy had significantly better estimates of
5-year DFS than those who did not.
Interestingly, adjuvant chemotherapy was associated with a
reduction in HR for recurrence with a borderline P value of .06 in
multivariate analysis, although that analysis included the large
number of low-risk patients, none of whom received adjuvant
treatment. These particular data from this relatively small cohort
suggest that, in contrast to previous large-scale randomized studies
in which chemotherapy was given indiscriminately to stage I pa-
tients, selective administration only to very early stage patients
designated as high-risk according to molecular analysis might more
readily yield an observable benefit in clinical outcomes.
Molecular testing identifies a population of high-risk patients that is
different from those identified using clinicopathologic NCCN char-
acteristics. In this study, 23 of 64 (36%) of patients classified as low-risk
according to NCCN criteria were identified as high-risk for recurrence
on the basis of molecular testing. In light of the higher recurrence rate
observed among untreated NCCN low-risk patients versus untreated
patients designated as low-risk according to the molecular test, these
data suggest that a significant number of early-stage NSCLC patients
might currently be undertreated with regard to adjuvant chemo-
therapy. Conversely, 11 of 36 (31%) of patients who met NCCN

Clinical Lung Cancer January 2018 - 63

 Adjuvant Chemotherapy in Molecular High-Risk Patients Improves Early Stage NSCLC Outcomes
high-risk criteria, and for whom NCCN guidelines would recommend
adjuvant chemotherapy, were molecular low-risk. None of those
discordant patients received adjuvant chemotherapy, and none had
disease recurrence.
Despite these intriguing observations, the current study is limited
by the lack of randomization to adjuvant chemotherapy. In addi-
tion, the relatively short median follow-up of 23 months as well as
the relatively small cohort of patients contributed to a limited
number of recurrence events, and therefore a relatively low statistical
power. However, on the basis of the analysis presented in this
report, this low number of events might also have been influenced
by a decision to undertake adjuvant chemotherapy in molecular
high-risk patients who would not have been considered candidates
for early intervention according to conventional criteria. Numeric
but statistically nonsignificant imbalances were observed between
treated and untreated molecular high-risk patients in terms of sex,
clinical stage, type of resection, and NCCN risk category. The di-
rection of these imbalances were mixed with regard to their expected
influences on clinical outcome; multivariate analysis further indi-
cated that, when adjusting for these factors, application of chemo-
therapy—which occurred selectively in molecular high-risk
patients—was associated with a borderline significant reduction in
the HR for recurrence. In the even smaller cohort of molecular
high-risk patients, adjustment for these factors in multivariate
regression analysis supported the same trend toward improvement
of DFS with adjuvant chemotherapy.
A larger, multicenter, randomized trial to evaluate the effect of
adjuvant therapy on disease-free and overall survival in molecular
high-risk patients is currently being organized, but data from this
larger, randomized study will not be available for years. Because of
the immediate need to improve outcomes in early stage NSCLC, we
undertook this smaller analysis of our single institution experience
integrating this molecular assay into clinical management, knowing
it would be limited by a relatively small sample size and lack of
randomization. Despite these limitations, we observed significantly
better DFS in molecular high-risk patients who elected to be treated
with adjuvant chemotherapy.
Conclusion
Adjuvant chemotherapy was associated with improved DFS in
stage I to IIA nonsquamous NSCLC patients designated as high-risk
according to a 14-gene prognostic assay. Although small and non-
randomized, this study therefore provides preliminary evidence that
a tumor’s 14-gene risk stratification has clinical utility in making
decisions regarding adjuvant chemotherapy.
Clinical Practice Points
 A CLIA-certified, commercially available 14-gene molecular
assay, used prospectively after complete surgical resection, was
64 - Clinical Lung Cancer January 2018
prognostic of recurrence in stage IA, IB, and IIA nonsquamous,
NSCLC patients.
 A molecular high-risk designation was also predictive of a sig-
nificant improvement in survival with adjuvant chemotherapy.
 Using molecular testing to supplement NCCN criteria might
better inform adjuvant chemotherapy recommendations.
Acknowledgments
The authors thank Jane Crockard, BS, Shayne Cardozo, BSN,
and Courtney Cook, BSN, for contributing to data collection in this
study.
Molecular testing was provided free of charge by Encore Clinical.
Disclosure
D.M.J., M.J.M., and J.R.K. have an equity interest in and
consulting relationship with Encore Clinical, which operates the
CLIA-certified laboratory that provides the 14-gene assay, and are
coinventors of technology licensed from the University of Califor-
nia. The remaining authors have stated that they have no conflicts
of interest.
References
1. Groome PA, Bolejack V, Crowley JJ, et al. The IASLC Lung Cancer Staging
Project: validation of the proposals for revision of the T, N, and M descriptors and
consequent stage groupings in the forthcoming (seventh) edition of the TNM
classification of malignant tumours. J Thorac Oncol 2007; 2:694-705.
2. Winton T, Livingston R, Johnson D, et al. Vinorelbine plus cisplatin vs.
observation in resected nonesmall-cell lung cancer. N Engl J Med 2005; 352:
2589-97.
3. Douillard JY, Rosell R, De Lena M, et al. Adjuvant vinorelbine plus cisplatin
versus observation in patients with completely resected stage IB-IIIA nonesmall-
cell lung cancer (Adjuvant Navelbine International Trialist Association [ANITA]):
a randomised controlled trial. Lancet Oncol 2006; 7:719-27.
4. Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J.
Cisplatin-based adjuvant chemotherapy in patients with completely resected
nonesmall-cell lung cancer. N Engl J Med 2004; 350:351-60.
5. Pignon JP, Tribodet H, Scagliotti GV, et al. Lung adjuvant cisplatin evaluation:
a pooled analysis by the LACE Collaborative Group. J Clin Oncol 2008; 26:
3552-9.
6. National Comprehensive Cancer Network. NCCN NoneSmall-Cell Lung Cancer
Clinical practice guidelines. Version 4, Available at: http://www.nccn.org/
professionals/physician_gls/pdf/nscl.pdf 2017. Accessed: February 17, 2017.
7. Kobayashi N, Toyooka S, Soh J, et al. Risk factors for recurrence and unfavorable
prognosis in patients with stage I nonesmall-cell lung cancer and a tumor diameter
of 20 mm or less. J Thorac Oncol 2007; 2:808-12.
8. Hsu CP, Hsia JY, Chang GC, et al. Surgical-pathologic factors affect long-term
outcomes in stage IB (pT2 N0 M0) nonesmall-cell lung cancer: a heteroge-
neous disease. J Thorac Cardiovasc Surg 2009; 138:426-33.
9. Kratz JR, He J, Van Den Eeden SK, et al. A practical molecular assay to predict
survival in resected non-squamous, nonesmall-cell lung cancer: development and
international validation studies. Lancet 2012; 379:823-32.
10. Kratz J, Van den Eden SK, He J, Jablons DM, Mann MJ. A prognostic assay to
identify patients at high risk of mortality despite small, node-negative lung tumors.
JAMA 2012; 308:1629-31.
11. Woodard GA, Gubens MA, Jahan TM, et al. Prognostic molecular assay might
improve identification of patients at risk for recurrence in early-stage nonesmall-
cell lung cancer. Clin Lung Cancer 2014; 15:426-32.
12. Kratz JR, Pham PT, Mulvihill MS, et al. Analytical validation of a practical mo-
lecular assay prognostic of survival in non-squamous nonesmall-cell lung cancer.
Diagn Mol Pathol 2013; 22:65-9.