Understanding predictive and prognostic markers Professor Aimery de Gramont Chairman of ARCAD Foundation and GERCOR, Paris FRANCE
Understanding predictive and prognostic markers Aimery de Gramont
Prognostic marker Effect of marker on clinical endpoint M C Buyse, Eur J Cancer Suppl 5, 2007.
Chibaudel B, et al. The Oncologist 2011
QUASAR Results: Recurrence Score, T Stage, and MMR Deficiency are Key Independent Predictors of Recurrence in Stage II Colon Cancer Risk of recurrence at 3 years 45% 40% 35% 30% 25% 20% 15% 10% 5% 0% 0 10 20 30 40 50 60 70 Recurrence Score T4 stage (13%) T3 and MMR proficient (76%) MMR deficient (11%) Kerr D, ASCO 2009 O Connell JCO 2010
Predictive marker Trt Effect of treatment on clinical endpoint depends on the marker M C Buyse, Eur J Cancer Suppl 5, 2007.
HER2 and trastuzumab mechanism of action trastuzumab HER2 receptor Trastuzumab Inhibits HER2-mediated signalling in HER2-positive tumors Prevents HER2 activation by blocking extracellular domain cleavage Activates antibody-dependent cellular cytotoxicity
TOGA Herceptin gastric Cancer. Primary end point: OS Event 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 FC + T FC 11.1 13.8 Events 167 182 Median OS 13.8 11.1 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 HR 0.74 95% CI 0.60, 0.91 p value 0.0046 Time (months) No. at risk T, trastuzumab 294 290 277 266 246 223 209 185 173 143 147 117 113 90 90 64 71 47 56 32 43 24 30 16 21 14 13 7 12 6 6 5 4 0 1 0 Van Cutsem E. ASCO 2009 Bang YJ, et al. Lancet 2010 0 0
Crizotinib in locally advanced or metastatic ALK-positive lung cancer Shaw, et al. NEJM 2013
Epidermal growth factor receptor (EGFR) and KRAS Khambata-Ford S, et al. J Clin Onc 2007;25:3230 7
Relating KRAS status to efficacy: PFS Cetuximab + FOLFIRI HR=0.63; p=0.007 mpfs wild-type (n=172): 9.9 months mpfs mutant (n=105): 7.6 months FOLFIRI HR=0.97; p=0.87 mpfs wild-type (n=176): 8.7 months mpfs mutant (n=87): 8.1 months Progression-free survival estimate 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Cetuximab + FOLFIRI mutant Cetuximab + FOLFIRI wild-type 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 FOLFIRI mutant FOLFIRI wild-type 0 2 4 6 8 10 12 14 16 Months 0 2 4 6 8 10 12 14 16 Months Van Cutsem E, et al. ASCO 2008
Enrichment strategy (PRIME study) KRAS 1 RAS 2 RAS/RAF 2 KRAS MT KRAS WT RAS MT RAS WT RAS/RAF MT RAS/RAF WT 40% 60% 51.7% 48.3% 57.4% 42.6% (KRAS exon 2) KRAS WT: n=656 KRAS MT: n=440 (KRAS Ascertainment rate 93%) (KRAS/NRAS exon 2,3,4) RAS WT: n=512 RAS MT: n=548 (RAS Ascertainment rate 90%) (KRAS/NRAS exon 2,3,4; BRAF exon 15) RAS/RAF WT: n=446 RAS/RAF MT: n=601 1 Douillard JY, et al. J Clin Oncol 2010; 28:4697-705; 2 Oliner K, et al. ASCO 2013 #3511.
DFS in dmmr patients, pooled data Stage II (N=102) Stage III (N=63) % Disease Free 100 90 80 70 60 50 40 30 20 10 0 0 1 2 3 4 5 Years Untreated 87% Treated 72% HR: 2.80 (0.98-8.97) p=0.05 % Disease Free 100 90 80 70 60 50 40 30 20 10 0 0 1 2 3 4 5 Years 5 yr DFS 5 yr DFS Untreated 62% Treated 67% HR: 1.08 (0.44-2.68) p=0.86 Sargent, JCO 2009
CRC biomarkers classification Metastatic KRAS Adjuvant MMR WT Mutant dmmr pmmr BRAF NRAS. codon gene signature
Molecular Classification of CRC Molecular Classification of Colon Cancer Stage II-III Salazar, et al. ASCO 2013
Surrogate marker Effects of treatment Surrogate on surrogate and true Trt and on true endpoint S endpoint must be T must be correlated correlated Buyse et al, Biostatistics 2000;1:49; Gail, Pfeiffer and van Houwelingen, Biostatistics 2000;1:231.
PFS as a surrogate marker for OS Correlation between Experimental 5FU PFS hazard ratio and OS hazard PFS Association between PFS and OS Overall survival ratio Buyse et al, JCO 2007; 25: 5218.
ACCENT: 3yr DFS vs 5yr OS 1,3 1,2 1,1 HR for 5 Year OS 1 0,9 0,8 R 2 = 0.80 0,7 0,6 May 2004: ODAC recommends 3-yr DFS as new regulatory endpoint for FULL approval in adjuvant colon cancer 0,5 0,5 0,6 0,7 5-6 0,8 July 2013, Barcelona 0,9 1 1,1 1,2 1,3 HR for 3 Year DFS
Median number of patients in trials 300 250 200 150 100 50 0 5FU infusion Mayo Regimen 5FU HD infusion
Median number of patients in trials 1000 800 600 400 200 0 Raltitrexed Irinotecan Irinotecan Oxaliplatin Capecit. Capecit.
1600 1400 1200 1000 800 600 400 200 0 Median number of patients in trials Bevacizu. Bevacizu. PTK Cetux Panitumu. Panitumu.
Median number of patients in trials 1400 1200 1000 Targeted Therapies 800 600 400 200 5FU New Chemotherapy 0 2 trials are needed for first registration 1992 2000 2008
Sample size and hazard-ratio (HR) 1200 Sample size and HR α 0.05 2-sided β 0.80 Acrual 24m Follow-up 36m Median PFS 9m Sample Size 1000 800 600 400 200 0 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 HR http://hedwig.mgh.harvard.edu/sample_size/time_to_event/para_time.html
The requirement that new agents be demonstrated to be safe and effective before marketing approval is granted has led to a traditional paradigm for drug development that typically requires thousands of patients, hundreds of millions of dollars, and >1 decades to fulfill. Unless this paradigm is changed, there will not be enough patients, money, or time available to evaluate all of the promising new anticancer agents in development. New clinical trial designs and end points are necessary to permit more efficient evaluation of putative cancer treatments so that the most promising agents can move forward quickly, while disappointing agents are rapidly identified and discarded. Schilsky RL. Clin Cancer Res 2002, 8:935-8
Studies Received to Date Study Group P I Number of Patients E3200 ECOG Giantonio 820 BICC-C Pfizer Fuchs / Rothenberg 547 AVF2107g Hurwitz 923 NO16966 Cassidy / Saltz 2035 Roche/GNE N016967 Rothenberg 627 AVF2192 Kabbinavar 209 N9741 NCCTG / Sanofi Goldberg 795 NCIC CO.17 NCIC-CTG Jonker 572 OPTIMOX 1&2 Tournigand 823 C97-3 GERCOR Tournigand 226 C181 Peeters/ 1186 Van Cutsem PRIME (C203) Amgen Douillard 1183 PACCE (C249) Hecht 1054 C408 5-6 July 2013, Van Barcelona Cutsem 463 26
Studies Received to Date Study Group P I Patients CAIRO 1&2 DCCG Punt 1575 COIN Maughan 2445 FOCUS1 & FOCUS2 MRC (UK) Seymour 2577 MACRO Diaz Rubio / Aranda 480 03-TTD-01 TTD Diaz Rubio / Aranda 342 MAX AGITG Niall Tebbutt 471 HORG 99.30 HORG Souglakos 283 GONO GONO Falcone 244 N9841 NCCTG Pitot 491 AIO 22 Porschen 474 FIRE II (CIOX) AIO 177 HORIZON II 1076 HORIZON III AZ Hoff/Schmoll 1601 BOND MERCK Cunningham 329 Total 24,028 Studies recently added shown in boldface 27
Results Trial Level PFS vs. OS 28 28
Early objective response 29
30
Conclusions Type of marker: Identification Validation Goal Prognostic Easy, but often flawed Frequent, but often disappointing Helpful for therapeutic strategy Predictive Hard, needs randomized trial Very rare, needs large randomized trial Avoids ineffective therapy Decreases cost Surrogate Hard, needs metaanalysis or large randomized trial Very rare, needs large randomized trial Faster results, smaller population Decreases cost