1. Ready for Prime Time Johanna Bendell, MD Director, GI Cancer Research Sarah Cannon Research Institute

2. Disclosures No conflicts to disclose PRESENTED BY: Johanna C. Bendell, MD

3. Learning Objectives After reading and reviewing this material, the participant should be able to: Know the biomarkers currently under evaluation for patients with colorectal cancer Know where these biomarkers have clinical application PRESENTED BY: Johanna C. Bendell, MD

4. What are we currently using and what is to come? KRAS BRAF Rash Mismatch Repair What does the future hold? –EGFR ligands –PI3 Kinase and PTEN –C-met

5. FIRST, SOME OF THE BACKGROUND… PRESENTED BY: Johanna Bendell, MD

6. Slide Courtesy Axel Grothey, Mayo Clin Proc. 2008;83(7):825- 846 Inhibition of apoptosis Cell growth & proliferation Survival gene transcription Protein translation Angiogenesis EGF EGF R

7. Evaluating Predictive Biomarkers: Trial Design Patient Population R Biomarker positive Biomarker negative Receive treatment Do not receive treatment Receive treatment Do not receive treatment

8. EGFR Pathway Signaling in CRC P P P P Grb2 Sos Ras Raf MEK ERK KRAS mutation (40%–50%) BRAF mutation (10%) Mutually exclusive EGFR Survival Proliferation Angiogenesis Metastasis EGFR = epidermal growth factor receptor.

9. Third-Line Panitumumab vs Best Supportive Care: Impact of KRAS Mutations—PFS The quantitative interaction test comparing the magnitude of the relative treatment effect on PFS between wild type and mutant KRAS was statistically significant (P<.0001). When treatment arms were combined, overall survival was longer in patients with wild-type KRAS vs patients with mutant KRAS. Amado et al. ASCO GI, 2008. Abstract 278. PFS by KRAS Status a and Treatment b Mutated KRAS Wild-Type KRAS Proportion Event-Free (%) Events/N(%) Median in weeks Mean in weeks 76/84(90)7.49.9 95/100(95)7.310.2 PMAB + BSC Events/N(%) Median in weeks Mean in weeks 115/124(93)12.319.0 114/119(96)7.39.3 PMAB + BSC BSC Alone P<0.0001 BSC Alone

10. NCIC CO.17 Trial Karapetis et al. NEJM 2008; 359(17):1757-1765 Overall Survival: Mutant Overall Survival: Wild-type Not prognostic here (BSC patients)

11. StudyRegimenKras Wild TypeKras Mutant + EGFR- EGFR+ EGFR- EGFR CRYSTAL VanCutsem 2009FOLFIRI +/- C9.9 mo.8.4 mo. 7.4 mo. (FOLFIRI + C) 7.7 mo. (FOLFIRI) HR 0.696; P=0.0012P=0.26 OPUS Bokemeyer 2009FOLFOX +/- C7.7 mo.7.2 mo.5.5 mo.8.6 mo. HR 0.57 [0.35-0.90]; P=0.0163HR 1.83 [1.09-3.05]; P=0.0192 COIN Maughan 2009 FOLFOX/XELOX +/- C 8.6 mo. 6.5 mo.6.9 mo. P=0.6P=0.46 PRIME Douillard 2009FOLFOX +/- P9.6 mo.8.0 mo.7.3 mo.8.8 mo. P=0.02 181 Price 2010FOLFIRI +/- P5.9 mo.3.9 mo.5.0 mo.4.9 mo. HR 0.73 [0.59-0.90]; P=0.004HR 0.84 [0.68-1.06]; P=0.14 Pmab Amado 2008Pmab v. BSC12.3 wks.7.3 wks.7.4 wks.7.3 wks. HR 0.45 [0.34-0.59]; P<0.0001 HR 0.99 [0.73-1.36] Cetuximab Jonker 2007 Cetuximab v. BSC 3.7 mo.1.9 mo.1.8 mo. HR 0.40 [0.30-0.54]; P<0.001HR 0.99; P = 0.96 PFS: Selected Randomized Trials with EGFR inhibitors

12. StudyRegimenKras Wild TypeKras Mutant + EGFR- EGFR+ EGFR- EGFR CRYSTAL VanCutsem 2009FOLFIRI +/- C 57.3%39.7%31.3% (FOLFIRI + C) 36.1% (FOLFIRI) P=0.34 OPUS Bokemeyer 2009FOLFOX +/- C 57%34% 53% P=0.0027P=0.029 COIN Maughan 2009 FOLFOX/XELOX +/- C 59%50%40%41% P=0.015P=0.87 PRIME Douillard 2009FOLFOX +/- P 55%48%40% P=0.068P=0.98 181 Price 2010FOLFIRI +/- P 35%10%13%14% Pmab Amado 2008Pmab v. BSC 17%0% Cetuximab Jonker 2007 Cetuximab v. BSC 12.8%0%1.2%0% RR: Selected Randomized Trials with EGFR inhibitors

13. But there is not only one way to mutate a KRAS… KRAS mutation Amino acid substitution Nucleic acid substitutionRelative Incidence % Absolute Incidence % Codon 12 mutations Aspartate (G12D)G35A32.513 Valine (G12V)G35T22.59 Cysteine (G12C)G34T8.83 Serine (G12S)G3A7.83 Alanine (G12A)G35C6.43 Arginine (G12R)G34C0.90.4 Codon 13 mutations Aspartate (G13D)G38A19.58 Other mutations1.80.7 KRAS mutation frequency in CRC patients:  40% Normanno N, et al. Nat Rev Clin Onc 2009;6:519-27

14. Tejpar et al., ASCO 2011

15. We might just have to consider the type of mutation Tejpar et al., ASCO 2011 Pooled analysis of OPUS and CRYSTAL

16. KRAS is an active biomarker Helps us decide who is not a candidate for EGFR inhibitor therapy But there may be more to the story And another example of how we cannot assume too quickly…

17. BRAF Mutations in CRC BRAF is primary effector of KRAS signaling BRAF mutations: –Occur most frequently in exon 15 (V600E) –Found in 4%-14% of patients with CRC –Mutually exclusive with KRAS mutations Raf MEK Erk P PP P Tumor cell proliferation and survival EGF Tumor Cell Ras Yarden. Nat Rev Mol Cell Biol. 2001;2:127; Di Nicolantonio. J Clin Oncol. 2008;26:5705; Artale. J Clin Oncol. 2008;26:4217.

18. HR, hazard ratio 32251612852220 3824146633100 0 0 CT CT + cetuximab Pooled analysis of OS in patients with KRAS WT / BRAF Mutant tumors Probability of overall survival 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 18061224603036424854 Time (months) Number of patients 3493172682251631208063194 3813502832121491076346172 0 0 CT CT + cetuximab KRAS wt/BRAF wt HR [95% CI]: 0.840 [0.710–0.993] p=0.041 FOLFIRI / FOLFOX4 + cetuximab: (n=349) median 24.8 months FOLFIRI / FOLFOX4: (n=381) median 21.1 months KRAS wt/BRAF mt HR [95% CI]: 0.633 [0.378–1.060] p=0.079 FOLFIRI / FOLFOX4 + cetuximab: (n=32) median 14.1 months FOLFIRI / FOLFOX4: (n=38) median 9.9 months Van Cutsem E, et al. N Engl J Med 2009;360:1408-17; Bokemeyer C, et al. J Clin Oncol 2009;27:663-71

19. CRYSTAL trial update: outcome in KRAS wild-type/ BRAF mutated mCRC KRAS wt/BRAF wt (n=566) KRAS wt/BRAF mt (n=59) FOLFIRI (n= 289) FOLFIRI + Cetuximab (n= 277) FOLFIRI (n=33) FOLFIRI + Cetuximab (n=26) mOS (mo)21.625.110.314.1 HR [95% CI] p-value a 0.83 [0.687–1.004] 0.0549 0.91 [0.507–1.624] 0.7440 mPFS (mo) 8.810.95.68.0 HR [95% CI] p-value a 0.68 [0.533–0.864] 0.0016 0.93 [0.425–2.056] 0.8656 RR (%) [95% CI] 42.6 [36.8–48.5] 61.0 [55.0–66.8] 15.2 [5.1–31.9] 19.2 [6.6–39.4] p-value b <0.00010.9136 a Stratified log-rank test; b Cochran-Mantel-Haenszel test Van Cutsem et al, JCO 2011

20. BRAF inhibitor: PLX4032 PLX4032 is an oral inhibitor of the BRAF mutant kinase which shows pronounced activity in BRAF mutant melanoma. Kopetz et al ASCO 2010 reported on an expansion cohort of patients with BRAF mutant mCRC (n=21) treated with PLX4032 at MTD of 960 mg bid 1 confirmed partial response and 4 minor responses (≥ 10% shrinkage) Adapted from slide courtesy Axel Grothey

21. Is BRAF ready for prime time? Sort of… It does not help predict who should and should not receive EGFR inhibitors But it does tell us if our patients have a significantly poorer prognosis But does this have treatment implications? –Maybe less likely to do chemotherapy holidays? –More frequent tumor assessment? –Referral for trials? Raf inhibitors? Combination trials?

22. Phase III Study 1 st Line FOLFOX4 ± Pmab: Efficacy by Skin Toxicity Analysis of efficacy and survival by worst grade skin toxicity severity in panitumumab + FOLFOX4 arms Pmab = panitumumab. Douillard J, et al. ASCO 2010. Abstract 3528. KRAS wt KRAS mt Grade 2–4 Grade 0–1 Grade 2–4 Grade 0–1 Median PFS, mo (95% CI)11.1 (9.4–12.9) 6.0 (5.2–8.7)7.6 (7.2–9.2)6.1 (5.1–7.4) Hazard ratio 0.562 (95% CI, 0.399–0.793) P =.001 0.661 (95% CI, 0.471–0.928) P =.0168 Median OS, mo (95% CI)28.3 (23.9–NE)11.5 (9.1–20.2)17.0 (14.9–20.2)10.1 (7.2–13.3) Hazard ratio 0.455 (95% CI, 0.318–0.651) P <.0001 0.595 (95% CI, 0.417–0.849) P =.0042 ORR, %62 39 4436 P value.003.219

23. Phase III Study 2 nd Line FOLFIRI ± Pmab: Pmab PFS by Skin Toxicity Severity ST = skin toxicity. Price TJ, et al. ASCO 2010. Abstract 3529. ST grade 2–4 was associated with longer PFS compared with ST grade 0–1, regardless of KRAS tumor status Events n (%) Median (95% CI) months Grade 2–4172/208 (83)7.4 (6.1–8.3) Grade 0–174/84 (88)4.5 (3.7–6.3) HR (Gr2–4:0–1) = 0.592 (95% CI, 0.447, 0.786) Log-rank P value = 0.0003 Events n (%) Median (95% CI) months Grade 2–4152/160 (95)6.0 (5.5–7.4) Grade 0–168/70 (97)2.8 (2.0–3.7) HR (Gr2–4:0–1) = 0.460 (95% CI, 0.340, 0.623) Log-rank P value <.0001 KRAS Safety Set, Panitumumab* 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 02468101214161820222426 Progression-free Survival Probability Months KRAS wt 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 02468101214161820222426 Progression-free Survival Probability Months KRAS mt

24. Phase III Study 2 nd Line FOLFIRI ± Pmab: Pmab OS by Skin Toxicity Severity ST grade 2–4 was associated with longer OS compared with ST grade 0–1, regardless of KRAS tumor status Events n (%) Median (95% CI) months Grade 2–4130/208 (63)16.5 (14.7–19.5) Grade 0–166/84 (79)9.0 (6.7–12.2) HR (Gr2–4:0–1) = 0.415 (95% CI, 0.302, 0.572) Log-rank P value <.0001 Events n (%) Median (95% CI) months Grade 2–4116/160 (73)13.7 (12.5–15.0) Grade 0–161/70 (87)7.3 (5.3–8.6) HR (Gr2–4:0–1) = 0.408 (95% CI, 0.294, 0.564) Log-rank P value <.0001 Survival Probability 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 02468101214161820222426 Months KRAS wt 283032 Survival Probability 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 02468101214161820222426 Months KRAS mt 283032 KRAS Safety Set, Panitumumab* Price TJ, et al. ASCO 2010. Abstract 3529.

25. Phase III Study 2 nd Line FOLFIRI ± Pmab: Pmab ORR by Skin Toxicity Severity ST grade 2–4 was associated with higher ORR compared with ST grade 0–1 only in patients with KRAS wt tumors 37 24 42 14 13 14 0 10 20 30 40 50 60 70 80 90 100 Wild-typeMutant All treated Grade 0–1 Grade 2–4 P =.003 a Odds ratio (95% CI) = 2.46 (1.31, 4.61) P = 1.000 a Odds ratio (95% CI) = 1.03 (0.42, 2.72) n = 84n = 202n = 286n = 70n = 154n = 224 Response Rate (%) Panitumumab KRAS Safety & Central Tumor Response Set* a Association between severity and objective response rate over study period (ORR) (odds ratio Grade 2–4 vs 0–1, adjusted for stratification factors) *Landmark analysis among patients with a PFS time ≥28 days Price TJ, et al. ASCO 2010. Abstract 3529.

26. Rash and α-EGFR Response  EVEREST study explored dose escalation in patients who did not develop rash to cetuximab Initial irinotecan/ cetuximab >G1 Rash Standard-dose cetuximab ≤G1 Rash Standard-dose cetuximab High-dose cetuximab Van Cutsem et al, GI symposium 2007, ab 237 Slide courtesy Bert O’Neil

27. EVEREST Results Standard DoseHigh DoseNonrandomized PR + CR16%30%25% PFS3.94.84.1 OS10.08.68.7 Conclusions:  Dose escalation was safe  Response increase was interesting  KRAS results are pending – Numbers small when accounting for KRAS Van Cutsem et al, GI symposium 2007, ab 237 Slide courtesy Bert O’Neil

28. Rash Making the best out of a bad thing Should we actually treat to rash? Does the appearance of a rash tell us something about the patient’s tumor itself? –Prognostic? Can we use this to further develop therapies?

29. Carethers et al., 1999; Arnold et al., 2003. Defective MMR: Adjuvant colon cancer Characterized by presence of MSI and loss of MLH1, MSH2, MSH6 or PMS2 expression –Microsatellites (tandem dinucleotide repeats) in genome can cause errors during replication that are fixed by mismatch repair mechanism –Mutations in MMR can lead to increased mutations and higher risk of colon cancer –Seems to give better prognosis once cancer established ~15% of sporadic CC, > 90% loss of MLH1 Clinical correlations: Right-sided, female, early stage, better prognosis Tumors: Poorly differentiated, signet-ring cell, lymphocytic infiltration, near diploid dMMR cells resistant to 5-FU

30. Pooled data on MMR status and adjuvant therapy TrialTreatmentN% Stage II% dMMR 7848525FU/LEV11730%14% INT 00355FU/LEV21550%18% 8746515FU/LV6619%12% GIVIO5FU/LV18352%16% FFCD5FU/LV15466%19% NCIC5FU/LV29261%15% Total102752%16% Sargent, ASCO 2008

31. Overall Survival by Treatment, stage II dMMR patients HR: 3.15 (1.07-9.29) p=0.03 N = 55 N = 47 Untreated 93% Treated 75% 5 yr OS P-value = 0.014 for treatment by MMR status interaction Sargent, ASCO 2008

32. MMR/MSI status in Stage II disease We should be checking 5-FU/LV is still standard for average risk Stage II disease if treatment is given Retrospective data strongly suggests these patients should not receive adjuvant chemotherapy We do not know how the addition of oxaliplatin might impact this data

33. QUASAR: Validation of Prespecified Assay Kerr. ASCO 2009. Abstract 4000. 48 genes associated with recurrence risk and 66 predictive of 5-FU/LV benefit FINAL ASSAY 7 recurrence genes 6 treatment benefit genes 5 reference genes RECURRENCE SCORE (0-100) TREATMENT SCORE (0-100) Multivariate analysis 761 candidate genes in 1,851 patients

34. QUASAR Results: Clinical/Pathological Covariates and Recurrence Pre-specified Multivariate Analysis, Surgery Alone Patients (n=605) Kerr ASCO 2009

35. 22% (16%-29%) 18% (13%-24%) 12% ( 9% -16%) Kaplan-Meier Estimates (95% CI) of Recurrence Risk at 3 years QUASAR Results: Recurrence Risk in Pre-specified Recurrence Risk Groups (n=711) Comparison of High vs. Low Recurrence Risk Groups using Cox Model: HR = 1.47 (p=0.046) Years Recurrence Risk Group High Intermediate Low Proportion Event Free 0.0 0.2 0.4 0.6 0.8 1.0 012345 Recurrence Risk Group Range of RS Proportion of patients Low<3043.7% Intermediate30-4030.7% High≥4125.6% Kerr ASCO 2009

36. QUASAR Results: Prediction of Differential 5FU/LV Benefit for Treatment Score and Recurrence Score Continuous Treatment Score and Treatment Benefit with 5FU/LV –Treatment Score by Treatment Interaction for RFI: interaction p = 0.19 Selected Secondary Analyses –Treatment Score by Treatment Interaction not significant when adjusted for prognostic covariates –Treatment Score by Treatment Interaction not significant for DFS (interaction p=0.12) or OS (interaction p=0.15) RS as a predictor of Chemotherapy Benefit: interaction p=0.69 –No significant difference in PROPORTIONAL benefit of chemotherapy was observed at low RS and at high RS Kerr ASCO 2009

37. Where does the Recurrence Score stand? It does predict patients at increased risk of disease recurrence BUT it does not predict who should get treatment in stage II with 5-FU alone treatment The answer likely lies in a prospective randomized trial – but this is a tough prospect C-07 data with oxaliplatin therapy presented at this year’s ASCO It will be interesting to see MOSAIC data

38. 1 Singh, AB et al. Cell Signal; 17:1183-1193,2005 2 Shelly, M et al. J Biol Chem; 273:10496-10505,1998 3 Khambata-Ford, S. et al. J Clin Oncol; 25:3230-3237, 2007 Amphiregulin/Epiregulin EGFR ligands: –1 in C. Elegans –4 in Drosophila –7 in mammals: EGF, TGF-α, HB-EGF, amphiregulin (AREG), betacellulin, epiregulin (EREG) and epigen 1 –EREG and AREG bind more weakly to EGFR than EGF but much more potently and prolonged –EREG preferentially activates heterodimers 2 High gene expression levels of EREG and AREG predict response to cetuximab 3 –High levels define tumors that are EGFR-dependent? Tabernero, ASCO 2009

39. EREG Expression in CRC Treated with Cetuximab NCIC CTG CO.17: Phase III clinical trial –Best supportive care (BST) ± cetuximab –572 patient tumor samples available –This trial is able to identify PROGNOSTIC variables –EREG expression detected by quantitative RT-PCR –High EREG and KRAS wild-type status present in 44% of patients –Median OS 9.9 vs 5.0 months for cetuximab vs best supportive care (HR, 0.46; P < 0.001) IHC: immunohistochemistry; RT-PCR: real-time polymerase chain reaction Jonker D, et al. ASCO 2009. Abstract 4016.

40. EREG Is a predictive marker of cetuximab efficacy Low EREG by minimum-P threshold Cetuximab + BSC BSC alone HR 0.93 [0.51-1.71], P = 0.81 Proportion alive 0 20 40 60 80 100 Time from randomization (months) 0 30 26 2 25 18 4 16 15 6 13 10 8 8 5 5 3 High EREG by minimum-P threshold Proportion alive 20 40 60 80 100 Time from randomization (months) 0 84 85 2 80 73 4 76 54 6 66 26 8 43 19 10 28 14 12 18 10 14 8 5 Cetuximab + BSC HR 0.46 [0.32-0.65], P < 0.0001 BSC alone 0 Jonker D, et al. ASCO 2009. Abstract 4016.

41. Combimarker: K-ras Wild-type PLUS EREG High –All comers n = 394 (100%)HR: 0.7 –K-ras wild-typen = 230 (58%) HR: 0.55 –Combimarkern = 169 (44%) HR: 0.46 Could use of the Combimarker effectively “stack the deck” to choose patients who would benefit from cetuximab use in earlier lines of therapy? Jonker. ASCO 2009. Abstract 4016.

42. Fecher, L. A. et al. J Clin Oncol; 25:1606-1620 2007 PI3K/AKT/mTOR Pathway PI3K / mTOR / AKT Pathway

43. BKM120 Phase I: clinical efficacy by radiologic assessment and overall response Best percentage change from baseline in sum of longest diameters PD SDNASD PDSD NA SD NA PR n=24 patients had measurable disease and baseline and post-baseline target lesion assessments Altered (mutation/expression)WT / Normal PIK3CA PTEN KRAS Resp a a Response as per investigator; NA, not available 100 50 150100 802580 100 25805080 12.5 150100 150100 Dose (mg) Pt 123456789101112131415161718192021222324 -90 -70 -50 -30 -10 10 30 50 BreastColorectalOther Bendell, et al. JCO 2011

44. Met Pathway Appleman (2011) JCO ePub

45. Rationale for MetMAb in mCRC Prevalence of Met expression in CRC was estimated to be 40%-50% using the Dx positivity criteria used in NSCLC Published data suggest that high Met and HGF expression are poor prognostic factors in CRC Stein, et al. Nat Med. 2009;15(1):59-67. P=0.1827 Low MET High MET 1.0 0.8 0.6 0.4 0.2 0.0 0 25 50 75 100 125 Metastasis-free Survival, Based on MET Time, Months Low MET High MET 1.0 0.8 0.6 0.4 0.2 0.0 Overall Survival Years “Low”: low HGF/ low MET (n=32) “High”: high HGF/ high MET (n=27) P=0.039 Median follow-up 6 years Kammula, et al. Cancer Lett. 2007;248(2):219-228. 0 2 4 6 8 10 12

46. MetMAb plus erlotinib in Met Dx+ patients Probability of progression free PFS: HR=0.53OS: HR=0.37 Time to progression (months) 0369121518 0.0 0.2 0.4 0.6 0.8 1.0 Placebo + erlotinib 3.8 26 MetMAb + erlotinib 12.6 16 Median (mo) HR (95% CI) Log-rank p-value No. of events Overall survival (months) 036912151821 Probability of survival 0.0 0.2 0.4 0.6 0.8 1.0 Placebo + erlotinib 1.5 27 Median (mo) HR (95% CI) Log-rank p-value No. of events MetMAb + erlotinib 2.9 20 0.53 (0.28–0.99) 0.04 0.37 (0.19–0.72) 0.002 Spigel, et al. ASCO 2011

47. Activity ARQ197 + cetuximab Eng et al. ASCO GI 2011 Median # prior therapies: 2 (range 1-4)

48. Primary Endpoint Overall Response Rate Panitumumab + Placebo (n = 48) Panitumumab + Rilotumumab (AMG 102) (n = 48) Panitumumab + Ganitumab (AMG 479) (n = 46) Objective Response - n (%) Complete Response (CR) Partial Response (PR) Stable Disease (SD) Progressive Disease (PD) Unevaluable/Not done 10 (21) 0 (0) 10 (21) 17 (35) 16 (33) 5 (10) 15 (31) 0 (0) 15 (31) 19 (40) 11 (23) 3 (6) 10 (22) 0 (0) 10 (22) 18 (39) 15 (33) 3 (6) Disease control rate a - % (95% CI)56 (41-71)71 (56-83)61 (45-75) Duration of response - median months (95% CI) 3.7 (3.6-NE)5.1 (3.7-5.6)3.7 (3.6-5.8) Posterior probability of Odds Ratio > 1 b 0.930.63 a Disease control rate = CR + PR + SD b OR is calculated based on ORR; an OR > 1 favors the combination arm over panitumumab alone NE, not estimable Responses were required to be confirmed at least 4 weeks after response criteria were first met VanCutsem et al. ASCOGI 2011, Eng et al., ASCO 2011

49. We are ready for prime time Biomarkers tell us about the individual patient’s tumor Now we have therapies that can focus on the specific pathways found to be abnormally activated or inactivated We are just at the beginning of this era of treatment options for patients We have to remember not to assume too quickly We have to be diligent about having biomarker studies included in large clinical trials so we can continue to learn how to best use them