1. Molecular Markers and Testing Strategies
in Advanced Non-Small Cell Lung Cancer
Howard (Jack) West, MD
Swedish Cancer Institute
Seattle, WA
Challenging Cases
in Breast & Lung Cancer
San Francisco, CA
October 22, 2011

2. Why do molecular testing?
• To improve clinical outcomes
– Give best treatment first (timing of EGFR TKI Rx)
– To provide access to agent (crizotinib for ALK-
positive)
– To identify subsets who might benefit from targeted
therapy (cetuximab?)
• To facilitate clinical research
– May improve patient outcomes
– Better understanding of molecular oncology

3. IPASS: Gefitinib vs. Carbo/Paclitaxel as
First Line Rx in Asian Never- or Light Ex-Smokers
R Carbo/Paclitaxel IV
Advanced NSCLC
No Prior Systemic Therapy A every 3 weeks
Never-/Light Former Smoker N
N = 1217 Gefitinib 250 mg/day
D
• Primary Endpoint: Progr-Free Survival (PFS)
• Biomarker analysis
Mok, NEJM 2009

4. IPASS: Objective Response Rate by
EGFR Mutation Status
Gefitinib
71.2% Carboplatin / paclitaxel
EGFR M+ odds ratio (95% CI) = 2.75
(1.65, 4.60), p=0.0001
Overall 47.3% EGFR M- odds ratio (95% CI) = 0.04
response (0.01, 0.27), p=0.0013
rate (%)
23.5%
1.1%
(n=132) (n=129) (n=91) (n=85)
Odds ratio >1 implies greater chance of response on gefitinib
Mok, NEJM 2009

5. IPASS Study: OS
Overall Population
Probability of Survival
1.0
HR: 0.90 (95% CI: 0.79-1.02; P = .109)
0.8
Gefitinib (n = 609)
0.6 Carboplatin/paclitaxel (n = 608)
0.4
0.2
0
0 4 8 12 16 20 24 28 32 36 40 44 48 52
Time Since Random Assignment (Mos)
EGFR Mutation Positive EGFR Mutation Negative
Probability of Survival
Probability of Survival
1.0 HR: 1.00 (95% CI: 0.76-1.33; P = .990) 1.0 HR: 1.18 (95% CI: 0.86-1.63; P = .309)
0.8 Gefitinib (n = 132) 0.8 Gefitinib (n = 91)
Carboplatin/paclitaxel (n = 129) Carboplatin/paclitaxel (n = 85)
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 4 8 12 16 20 24 28 32 36 40 44 48 52 0 4 8 12 16 20 24 28 32 36 40 44 48 52
Time Since Random Assignment (Mos) Time Since Random Assignment (Mos)
Fukuoka M, et al. J Clin Oncol. 2011;29:2866-2874.

6. IPASS Study: PFS by EGFR Status
• EGFR mutation status most predictive, EGFR gene
amplification also significantly predictive, but less
Fukuoka M, et al. J Clin Oncol. 2011;29:2866-2874.

7. Prevalence of EGFR Mutations
Phenotype of NSCLC Patient Prevalence of EGFR Mutation
All 10% to15%
Caucasian never-smokers ~35%
Asian never-smokers ~65%
Jackman DM et al. ASCO 2008 Annual Meeting. Abstract 8035.

8. MSKCC Continuum of Tobacco Exposure
and EGFR Mutations
• Looking for exon 19 or 21 mutations in tumors from 265 pts
with detailed smoking history
Pham, JCO 2006

9. Prevalence of EGFR Mutations by Smoking
Status: Lung AdenoCa at MSKCC
• Testing for exon 19 or 21 mutation in 2142 adenocarcinomas at
MSKCC
• Rate: EGFR mut’ns seen in 52% of never-smokers, 15% of former
smokers, and 6% of current smokers
D’Angelo, J Clin Oncol 2011

10. Prospective Trials of EGFR TKIs vs. Chemo
in EGFR Mutation (Exons 19, 21) Population
RR PFS (mo) OS (mo)
Trial N Rx
TKI Chemo TKI Chemo TKI Chemo
Maemondo Gefitinib vs.
230 74% 31% 10.8 5.4 30.5 23.6
NEJ002 Carbo/Pac
Mistudomi Gefitinib vs.
172 62% 32% 9.2 6.3 30.9 N.R.
WJTOG3405 Cis/Doce
Zhao Erlotinib vs.
165 83% 36% 13.1 4.6 N.R. N.R.
OPTIMAL Carbo/Gem
Rosell Erlotinib vs.
174 58% 15% 9.4 5.2 N.R. N.R.
EURTAC Plat Doublet
Maemondo, NEJM 2010; Mistudomi, Lancet Oncol 2010;
OPTIMAL, Lancet Oncol 2011; Rosell, ASCO 2011

11. TORCH Study:
Chemo  Erlotinib vs. Erlotinib  Chemo
Gridelli, ASCO 2010, #7508 Standard arm
R Cis/Gemcitabine Erlotinib 150 mg/d
Advanced NSCLC
No clinical or A x 6 cycles
molecular selection N Experimental arm
N = 760, Cis/Gemcitabine
closed by DSMB D Erlotinib 150 mg/d x 6 cycles
Order of therapy matters: get best treatment in at first opportunity

12. TORCH Study: Efficacy Analysis
Efficacy Outcome Erlotinib → Chemo → HR P
Chemo Erlotinib (95% CI) Value
(n = 380) (n = 380)
1.40
Median OS, mos 7.7 10.9 .002
(1.13-1.73)
Median PFS,* mos 2.2 5.7 Not reported
Objective response,† % 18 32
 With first-line treatment
• CR <1 1
• PR 9 27 Not reported
 With second-line treatment
• CR 1 <1
• PR 9 6
*Assessment of first-line treatment only.
†Intent-to-treat population.
Gridelli C, et al. ASCO 2010. Abstract 7508.

13. KRAS Mutations and Resistance
to EGFR Inhibitors
RR OS
Trial N Agent
Wild Type Mutant Wild Type Mutant
INTEREST 275 Gefitinib 10% 0% 7.5 mos 7.8 mos
Gefitinib or
Jackman 116 5% 0% 11.8 mos 13 mos
erlotinib
Gefitinib or
Massarelli 70 10% 0% 9.4 mos 5 mos
erlotinib
Shepherd 206 Erlotinib 10.2% 5% 7.5 mos 3.7 mos
Miller* 101 Erlotinib 32% 0% 21 mos 13 mos
*Patients with BAC.
Douillard. ASCO. 2008 (abstr 8001); Jackman. ASCO. 2008 (abstr 8035);
Massarelli. Clin Cancer Res. 2007;13:2890; Shepherd. ASCO. 2007 (abstr 7571);
Miller. J Clin Oncol. 2008;26:1472.

14. Waterfall Plot of Response on Erlotinib in
Advanced BAC, with Molecular Correlates
• Most but not all of the best responders carry EGFR mutations
• Many with minor response and even some with PR have neither EGFR
mutation nor EGFR gene amplification
• Several KRAS mutation patients have stable disease or even minor responses
(which very likely correlate with modest clinical benefit)
• Selecting on basis of EGFR mutations or FISH positivity would filter out many
beneficiaries; selecting by KRAS would also eliminate many w/SD or MR
Miller, JCO 2008

15. INTEREST Trial: A Wide Range of Very
Unhelpful Molecular Markers
Worldwide trial, second line docetaxel vs. gefitinib, N = 1466
OS DFS
Douillard, JCO 2010

16. EML4-ALK Translocations in NSCLC
Soda et al., Nature 448: 561-566, 2007
EML4-ALK frequency:
~4% (64/1709)
Primarily adenoCa, minimal
or no smoking history
Bang, ASCO 2010
#2 (Plenary), then
NEJM

17. 48 yo Female Never Smoker with Stage IV NSCLC
Positive for EML4-ALK
Pre-Treatment After 2 cycles PF-02341066

18. 77% of Patients with ALK-positive NSCLC
Remain on Crizotinib Treatment
• Duration of treatment
(median: 5.7 months)
0–3 mo 13 pts
>3–6 mo 29 pts
Individual patients
>6–9 mo 24 pts
>9–12 mo 9 pts
>12–18 mo 4 pts
>18 mo 3 pts
• Reasons for discontinuation
– Related AEs 1
– Non-related AEs 1
– Unrelated death 2
– Other 2
– Progression 13
0 3 6 9 12 15 18 21
Treatment duration (months)
N=82; red bars represent discontinued patients

19. The Majority of ALK-Positive Patients are
Never-Smokers
Never smoker
Light smoker
Smoker
ALK-Positive
Shaw et al. ASCO 2010

20. Mutation Status By Smoking History
Caucasians
18% 2%
ALK
EGFR
WT/WT
≤10 pack-years >10 pack-years
(N=255) (N=232)
Shaw et al. ASCO 2010

21. Mutation Status By Smoking History
Asians
8%
EGFR
KRAS
ALK
WT/WT/WT
1.2%
Never-smokers Ever-smokers
(N=127) (N=82)
Wong et al., Cancer 2009;115:1723-33.

22. Mutation Status in
Asian Never-Smokers
N=52
Sun et al. JCO 2010;28:4616-20

23. FLEX: Study Design
Adv NSCLC R Cis d1/Vin d1,8
EGFR IHC 1 +
No prior Rx
A
No brain mets N Cis d1/Vin d1,8
(N = 1,125) D + weekly cetuximab
• Stratification • Primary end point: OS
• – ECOG PS: 0/1 vs. 2
• – Stage: IIIB (wet) vs. IV • Secondary end points: PFS, RR, QOL, safety
Pirker et al, Lancet 2010

24. ITT (n=1125) Median OS 1-year survival
▬ CT + cetuximab 11.3 mo 47%
(n=557)
▬ CT 10.1 mo 42%
Overall survival (%)
(n=568)
HR=0.871 [95% CI 0.762–0.996]; p=0.044
Months
Pirker, Lancet 2010

25. EGFR Expression in FLEX Study
O‘Byrne,
Lancet Oncol 2011
Low EGFR Expression High EGFR Expression
Parameter
CT CT + Cetuximab CT CT + Cetuximab
Median OS
10.3 9.8 9.6 12.0
(months)

26. FLEX High EGFR IHC, by Histology
Adenocarcinoma (N = 135) Squamous (N = 144)

27. NCCN Guidelines (October, 2011)
• Patients with non-squamous advanced
NSCLC should be tested for EGFR mutation
and ALK rearrangement
• Treatment in this setting should be guided by
results of this testing
Caveat:
NCCN expertise: cancer treatment, independent of cost
NCCN non-expertise: health care economics/value of Rx

28. ASCO Guidelines (Sept, 2011) Re:
EGFR inhibitors
• Most patients should not receive EGFR TKI
as part of 1st line treatment
• For those with EGFR mut’n, EGFR TKI alone
may be recommended
• For patients receiving cis/vinorelbine,
cetuximab may be added

29. ASCO Guidelines (Sept, 2011) Re:
Molecular Testing
• ASCO recognizes that most patients with NSCLC may not
have any special molecular tests…these molecular tests
remain investigational, and selecting treatment based on
molecular tests has not been shown to improve a patients’s
overall length of live.
• Therefore, ASCO does not recommend using any routine
molecular analysis of tumor tissue to guide treatment
decisions at this time. For patients with an EGFR mutation,
erlotinib or gefitinib may be the best first-line therapy, but
may also work well as a second or third-line treatment.
• Larger tissue samples recommended when performing Bx,
to facilitate future research and maximize eligibility in trials.

30. My Conclusions:
Clinical Management Decisions
• EGFR TKIs
– EGFR TKI shouldn’t be 1st line Rx in unselected pts.
– If you know EGFR mut’n +, I’d give earlier >> later
– Importantly, EGFR wild type doesn’t mean no benefit
from EGFR TKI in maintenance or later
• ALK positivity is KEY to access to crizotinib 
major benefit
• EGFR IHC possibly very helpful in selecting for
cetuximab

31. Setting the Balance:
Where Do You Fit on the Spectrum?
Test enriched Test everyone to
population miss NOBODY
$5K/pos test
Never-smoker, adeno
Test 12%, find 40%
Adeno (any smoking Hx)
Test 45%, find 90%
Adeno (any) + non-squam light smokers,
Test 67%, find 95%
All non-squamous adv NSCLC
Test 75%, find 97%
All advanced NSCLC
Test 100%, find 100%
$60K/pos test
(All numbers approximate)

32. My Conclusions: Who To Test?
• Selective vs. everything for everyone?
– I favor a selective approach based on histology &
smoking status, not (yet) testing for all patients
• Remember fallibility of histology assignment
– EGFR mut’n pts can still get comparable benefit as
maintenance or 2nd line
– Stronger argument for ALK testing (pos = access)
– This is a clinical judgment, with ASCO and NCCN
endorsing different conclusions
– The only wrong answer is a dogmatic one