ADJUVANT FOLFOX4+/- CETUXIMAB IN KRAS WILD-TYPE PATIENTS WITH RESECTED STAGE III COLON CANCER RESULTS FROM THE PETACC8 INTERGROUP TRIAL

ABSTRACT Introduction Conventional imaging may underestimate response in APC, as it incorporates tumour desmoplasia which is prevalent in its microenvironment. The carbohydrate antigen 19-9 (CA 19-9) has been studied widely in APC. However, there has been no prospective phase 3 trial confirming CA 19-9 kinetics as a surrogate endpoint of OS. Methods CA 19-9 was measured at baseline and if elevated every 8 weeks in patients (pts) with APC in a randomised trial of gemcitabine (G) vs. G plus axitinib (ClinicalTrials.gov number NCT00471146). Pts were excluded if this measurement was missing. To assess the relationship between CA 19-9 kinetics (time-course) and OS, at least 3 measurements had to be available. Pre-treatment log-10 (CA 19-9) levels on OS (prognostic role) was modelled by Cox proportional hazards regression. The kinetics of log-10 (CA 19-9) (predictive role) was modelled by Cox regression where CA 19-9 was used as a time-varying covariate. These analyses used pre-treatment CA 19-9 and CA 19-9 kinetics as a continuous covariates and not as categorized variables. For both analyses, univariate (primary analysis) and multivariate Cox models were utilized to compare the effect of other covariates on the CA 19-9 results. Results 432 pts had at least one evaluable CA19-9 measurement (50% baseline ECOG = 1, ECOG = 2 was a study exclusion criterion, 75% metastatic disease). The hazard ratio (HR) for baseline CA19-9 was 1.16 (95% CI, 1.025-1.314) indicating there is a 16% increase in the risk of OS when there is a 10-fold increase in baseline CA19-9 between pts. The effect of baseline CA19-9 was 1.10 (0.978-1.24) when significant covariates were included (baseline ECOG, disease stage). There were 211 pts for the CA19-9 kinetic analysis (1036 measurements over time). The HR for CA19-9 when used as a time-varying covariate was 1.46 (1.292-1.637). This indicates that just prior to a given time, there is 46% increase in the hazard (or risk) of OS when one pt has a 10-fold higher CA19-9 value than another pt just prior to that time. In the multivariate analysis, the HR was 1.39 (1.232-1.566) with disease stage as the only significant covariate. Conclusion CA 19-9 kinetics is a useful surrogate tool in assessing systemic treatment efficacy in APC. This is the first phase 3 trial analysis in APC co01nfirming CA 19-9 kinetics as a surrogate endpoint of OS.