Disclosures Predicting survival in metastatic colorectal cancer Daniel Sargent, PhD Mayo Clinic Consulting activities Amgen Pfizer Roche/Genentech Sanofi-Aventis Genomic Health Personalized Medicine - Decision Tools - Old: Clinical parameters Patient-based Age, PS, co-morbidities, experience with prior therapies, financial implications Tumor-based Stage, differentiation, number and sites of metastases New: Molecular Biomarkers Patient-based (Pharmacogenomics) Tumor-based Old Factor 1: Age Pooled analysis of four first-line advanced CRC trials testing 5-FU regimens Bolus 5FU +/- LV, or 5EU Studies conducted d from 1984-97 97 Any age > 18, PS -3 Total sample size: 1748 D Andre et al CCC 25 Toxicity by Age Group 1 Survival by Age 1 Survival by PS Grade 3 (%) 5 4 3 2 P=.7 P=.33 23.5% 2.9% 18.1% 16.4% 16.8% 13.% -65 66-7 71+ P=.21 t Survival Percent 8 6 4 Age<=65 66-7 71+ P=.25 t Survival Percent 8 6 4 PS= PS=1 PS=2,3 P<.1 1 2.% 4.8% 3.5% 2 2 Diarrhea Stomatitis Infection 1 2 3 from Randomization 1 2 3 from Randomization 1
What about? Pooled analysis Study Comparator Setting N regimen MOSAIC 1 5-FU/LV Adjuvant 2246 N9741 2 IFL 1 st Line 546 de Gramont 3 5-FU/LV 1 st Line 42 Rothenberg 4 5-FU/LV 2 nd Line 531 Total 3743 1 Andre et al, NEJM 24; 2 Goldberg et al, JCO 24; 3 de Gramont et al, JCO 2; 4 Rothenberg et al, JCO 23 e and Alive % Progression-Fre 1 9 8 7 6 5 4 3 2 1 PFS - First Line de Gramont, Goldberg studies Age < 7 Age > 7 1 HR=.72 1 2 e and Alive % Progression-Free 9 8 7 6 5 4 3 2 1 HR =.6 1 2 % Alive 1 9 8 7 6 5 4 3 2 1 OS - First line de Gramont, Goldberg studies Age < 7 Age > 7 1 HR=.73 1 2 % Alive 9 8 7 6 5 4 3 2 1 HR=.67 1 2 and Alive % Progression-Free 1 9 8 7 6 5 4 3 2 1 PFS 2 nd line Rothenberg Age < 7 Age > 7 1 5-FU/LV HR=.61 1 2 and Alive % Progression-Free 9 8 7 6 5 4 3 2 1 5-FU/LV HR=.69 1 2 Response rates by Study and Age ont N9741 De Gramo Rothenberg <7 >=7 <7 >=7 <7 >=7 OR=9.2 OR=2.3 OR=1.7 OR=1.6 OR=3 3.6 1 2 3 4 5 6 Response Rate OR=2.9 Rate, Gra ade 3+ 5 45 4 35 3 25 2 15 1 5 Adverse Events (Gr > 3) p =.4 p =.4 Age < 7 Age > 7 p =.8 Neutropenia Thrombocytopenia Fatigue 2
Adverse Events (Gr > 3) Age: Conclusions Rate, Grade 3+ 16 14 12 1 8 6 4 2 p =.37 p =.38 p =.38 Age < 7 Age > 7 p =.2 Neurotoxicity Diarrhea Nausea/Vomiting 6 Day Mortailty Among patients entered onto clinical trials Younger and older patients accrue the same benefit from 5-FU/LV and 4 Elderly patients do not experience clinically meaningful increased toxicity Age alone should not exclude an otherwise healthy elderly patient from receiving chemotherapy Old Factor 2: PS Pooled analysis of 9 First Line Phase III Trials Compare PS -1 to PS 2 patients Sargent et al, JCO 29 Trials Included Trial Treatment Arm N N, PS2 de Gramont LVFU2*, 42 44 Douillard LVFU2*, ifu/ir 385 27 FOCUS ifu/lv*, ifu/ox, ifu-ir 2,135 18 N9741 IFL*,, IROX 1159 58 OPTIMOX *+/-LVFU2 616 51 AIO 5FU/LV* +/- IRI 43 19 Porshen FUFOX*, CAPEOX 471 39 Saltz 5-FU/LV*, IFL 444 65 Tournigand *, FOLFIRI 226 26 Total 6286 59 *Defined Characteristics of 6,286 Patients 1 PFS in PS -1 vs. PS 2 PS -1 PS 2 9 8 7 PS -1 (median = 7.6 mos) PS 2 (median = 4.9 mos) No. Patients 5777 59 Age, Median 63 (19-88) 63 (24-84) (Range) Sex, % Male 64 58 sion Free % Progress 6 5 4 3 2 p-value <.1 HR: 1.52 (1.38-1.66) Study Arm, % 41 4 1 2 4 6 8 1 12 14 16 Months 3
Response Ra ate 6 5 4 3 2 1 RR Treatment by PS p-value <.1 p-value =.2 OR: 1.89 (1.69-2.12) OR: 1.99 (1.29-3.5) 51.6 38.9 36.8 25.6 Interaction p-value =.89 PS -1 PS 2 Treatment Free % Progression PFS Treatment by PS 1 p-value <.1 ~ PS -1 (median = 6.7 mos) 9 HR:.82 (.77-.86) Treatment ~ PS -1 (median = 8.4 mos) 8 ~ PS 2 (median = 4. mos) 7 6 Treatment ~ PS 2 (median = 6. mos) 5 p-value =.2 4 HR:.79 (.66-.96) 3 2 1 Interaction p-value =.68 2 4 6 8 1 12 14 16 Months Grade > 3 Toxicity by PS PS -1 PS 2 P-value Nausea* 8% 16% <.1 Vomiting* 8% 12%.6 Stomatitis* 2% 5%.11 Diarrhea* 17% 15%.32 N-penia* 34% 35%.51 6 day mortality 3% 12% <.1 PS: Conclusions Among patients entered onto clinical trials, PS 2 patients: Have a poorer prognosis Accrue similar benefit to PS -1 pts from superior Rx Benefit more from 5-FU infusion with oxaliplatin or irinotecan Have higher rates of nausea, vomiting, & 6-day mortality *Note: FOCUS Trial Excluded Trt Age PS # of Sites Hgb ANC Alk Phos AST Additional Clinical Factors IFL IROX OTHER 4 to 5 <4 5 to 6 6 to 7 >7 1 2 1 2 3+ 12 to 13.5 <11 11 to 12 >13.5 <41 41 to 53 53 to 7 >7 <85 85 to 125 125 to 25 >25 <2 2 to 3 3 to 45 >45 NONE PREDICTIVE OF TREATMENT BENEFIT New factor 1: KRAS.5 1 1.5 2 2.5 Hazard Ratio Sannoff, JCO 28 4
FS P 1..9.8.7.6.5.4.3.2.1 PFS by KRAS: Crystal KRAS wt (n=348) HR=.68; p=.17 mpfs C + FOLFIRI: 9.9 mo mpfs FOLFIRI: 8.7 mo. 2 4 6 8 1 12 14 16 18 Months Cetuximab + FOLFIRI FOLFIRI FS PF 1. KRAS mt (n=192).9 HR=1.7; p=.47.8 mpfs C + FOLFIRI: 7.6 mo mpfs FOLFIRI: 8.1 mo.7.6.5.4.3.2.1. 2 4 6 8 1 12 14 16 Months Cetuximab + FOLFIRI FOLFIRI Van Cutsem, NEJM 29 No surprise: also for Pmab Press release 8/6/9 New Factor 2: UGT1A1 Severe Neutropenia Risk: 7/7 vs 6/6 + 6/7 Genotypes n/n (%) Est. Odds Author 7/7 6/6 + 6/7 Ratio 95% CI Innocenti 3/6 (5%) 3/53 (6%) 16.7 2.3-12.6 Rouits 4/7 (57%) 1/66 (15%) 7.5 75 1.4 14-38.5 Marcuello a 4/1 (4%) 18/85 (21%) 2.5.6-9.7 Ando b 4/7 (57%) 22/111 (2%) 5.4 1.1-25.9 a Gr 3+ neutropenia. b Gr 4 leukopenia and/or Gr 3+ diarrhea. From Parodi et al, FDA Subcommittee presentation, November, 24 But not Irinotecan-specific: N9741 UGT1A1 genotype IFL (n=19) (n=285) IROX (n=13) All 6/6 (n=23) 7% 19% 1% 15% 6/7 (n=22) 11% 22% 15% 18% 7/7 (n=47) 18% 36% 55% 36% p-value*.46.11.4.7 McLeod, ASCO 26 Additional Potential Predictive Markers for Colon Cancer Treatment Drug Fluoropyrimidines Irinotecan Oxaliplatin EGFR Antibodies VEGF inhibitors General *FDA-recognized Marker TS, DPD*, TP, MSI, MTHFR expression/polymorphisms UGT polymorphisms*, MSI, transporter polymorphisms ERCC1, GST P1, XPD expression, transporter polymorphisms gene amplification/polymorphism, RAS mutation, BRAF mutation, ligand expression, PTEN expression, VEGF levels VEGF polymorphisms, ICAM polymorphisms/levels, E-selectin levels, HIF1, Glut-1, VEGFr gene expression Circulating tumor cells 5
Next Advance: identification of resistance markers in K-Ras Wildtype tumors Bypass Pathways and EGFR Resistance CRC K-Ras MT K-Ras WT Acknowledgment: Lee Ellis? B-Raf PI3K PTEN EGFR copy # EGFR MoAB Camp et al Clin Ca Res 25 B-raf mutations All pts had progressed on at least one line of therapy ~5% received monotherapy with EGFR MoAB ~5 MoAB with chemotherapy In patients with tumors with Mut B-raf, there were NO objective responses B-Raf Predicts for Benefit of Anti-EGFR Therapy in Patients with WT Ras and the Entire Cohort PFS OS Wild-type K-ras All Patients Nicolantonio et al. JCO 28 Bypass Pathways and EGFR Resistance PTEN Expression and PFS PTEN in the metastasis was predictive of efficacy. PTEN in the primary tumor was NOT predictive of efficacy. PTEN in the primary tumor and liver metastasis was concordant in only 6% of cases Logrank Test: p=.5 HR =,49 95% CI:.2-.75 PTEN+ PTEN- Camp et al Clin Ca Res 25 Loupakis et al. JCO 29 6
Challenges with PTEN Expression in primary tumors does not reflect expression in metastases Although it is not lost or mutated in CRC, its expression can be regulated by methylation or mirna Difficult to standardize IHC in different labs CAIRO -2 STUDY CAPOX + Bevacizumab +/- Cetuximab 755 Pts Tissue in 545 Supplemental Figure 1: Representative examples of PTEN positive (A, B) and negative (C, D) cases. The cases reported in A and C panels were evaluated at Ospedale Niguarda Ca Granda (Milan, Italy) whereas those in B and D at the Institute of Pathology in Locarno (Switzerland). Sartore-Bianchi et al. Cancer Res 29. EGFR copy number 7% no difference in PFS among arms Loss PTEN 42% no difference in PFS EGFR amplification and PTEN did NOT predict for response to chemo + cetuximab therapy Tol et al. Proc AACR, 29 Abstract 691 Bypass Pathways and EGFR Resistance PFS and PIK3CA Mutational Status in mcrc Patients Treated With Panitumumab and Cetuximab 11 pts > 85% received at least 1 prior Rx Cetuximab 13% Panitumumab 2% Cetuximab/Irinotecan 67% Camp et al Clin Ca Res 25 Sartore-Bianchi, A. et al. Cancer Res 29 Fig. 1 The Role of PREDICTIVE Markers for Efficacy of EGFR MoABs The sure thing Probably Yes Maybe, jury is still out K-Ras B-Raf PI3K mutations EGFR by IHC EGFR amplification PTEN Gene expression arrays No Cetuximab 16 Cetuximab/Irinotecan 184 Excluding patients from EGFR MoAB Rx by use of multiple predictive factors will greatly increase the efficacy of EGFR MoABS It is imperative to PROSPECTIVELY include biomarkers and tissue procurement in clinical trials When possible, biomarkers in primary tumors and liver metastasis should be compared Prenen, H. et al. Clin Cancer Res 29 7
Predicting Survival in Advanced CRC: Conclusions Clinical factors of Age, PS should not exclude patients No factors to predict differential efficacy of Oxaliplatin vs irinotecan KRAS mutation definite exclusion for EGFR inhibitor BRAF likely also will be validated, but lower prevalence Other markers unclear No markers for anti-vegf therapy 8