An individual patient data (IPD) meta-analysis of the impact of thrombocytosis (?plts) on overall survival (OS) whilst using an intermittent chemotherapy (iCTx) strategy in advanced colorectal cancer (aCRC).

e15044 Background: iCTx in pts with aCRC offers potential for improvement in QoL. The COIN trial is the largest study to compare iCTx v. continuous strategies in aCRC, a pre-specified subgroup analysis of 16 baseline factors was undertaken among pts with stable or responding disease after 3 mths of first-line therapy to see if the relative treatment effect differed by subgroup. Baseline ⇡plts alone identified a group of pts with significantly worse OS when an iCTx strategy was applied. Here we seek to validate this finding in other intermittent strategy trials. Methods: Published RCTs of iCTx in aCRC were identified via literature review. Eligible trials could allow one or more re-introductions of “full” initial regimen either upon progression or after a set period of time. Outcome and platelet data were requested and collated into a central database. The COIN trial was declared the discovery dataset and other eligible trials the validating datasets. Two co-primary hypotheses were agreed based upon the COIN trial results: Hypothesis 1: In pts with baseline ⇡plts, any planned complete stop of all therapyis detrimental to OS when compared to any maintenance strategy. Hypothesis 2: In pts with baseline ⇡plts, any planned stop of oxaliplatin(Ox) therapy is detrimental to OS when compared to any equivalent strategy where Ox is maintained. Unadjusted IPD meta-analysis was performed according to a pre-specified statistical plan. Results: All trials had broadly similar inclusion criteria . Incidence of ⇡plts range 17-32%. ⇡plts was a poor prognostic marker. Combining IPD from all trials, iCTx was not detrimental to OS. Hypothesis 1 included AIO-0207, CAIRO3, COIN B, OPTIMOX 2 and GISCAD with 1622 pts, HR for interaction of ⇡plts with treatment strategy 0.97 (0.66-1.40), p = 0.78. Hypothesis 2 included TTD MACRO, NORDIC VII and OPTIMOX I, with 1268 pts, HR for interaction 1.36 (0.71-2.62), p = 0.18. Conclusions: These IPD meta-analyses do not validate COIN trial findings that showed reduced OS in pts with baseline ⇡plts who are given a planned treatment break. Sensitivity analyses will be presented, including impact of RAS mut status.