Dose-response relationship in phase I clinical trials: A European Drug Development Network (EDDN) collaboration study.

3084 Background: Alkylating agents have a strong dose effect relationship and this concept is used for development of novel cytotoxic (CTX) drugs. However, the need to reach MTD to obtain optimal benefit with molecularly targeted agents (MTA) is controversial. In this study we aimed to evaluate the relationship between dose and efficacy in a large cohort of phase I patients (pts) with refractory solid tumors. METHODS We retrospectively collected data on 2232 consecutive pts treated in phase I trials in 14 European institutions in 2005 - 2007. Inclusion criteria for this analysis were: a) pts treated within single agent studies where MTD was defined; b) RECIST response available. We used logistic and Cox regression models to study the relationship of dose (as continuous variable of % of MTD) with objective response rate (ORR), disease control rate (DC) and overall survival (OS). The clinical tumor growth rate (CTGR) - a log ratio of time between advanced disease diagnosis and phase I entry to number of treatment lines- was used to identify the impact of slow growing tumors on DC. RESULTS 1,015 pts were eligible (MTA 72%; CTX 28%). The ORR was 3% and DC at 3 and 6 months was 25% and 11%. Median CTGR was 22 weeks. There was no difference between ORR in pts receiving MTA or CTX but the DC rate at 3 and 6 months was higher for MTA (28% and 12%) than CTX (17% and 6%; p<0.01). This difference was attributed to the higher number of pts with slow growing tumors (CTGR > 20 weeks) treated with MTA (59%) vs. CTX (48%; p=0.001). Higher doses were associated with ORR to CTX agents (median dose: responders 92% vs. non-responders 62%; OR=6.4, p=0.06) but not to MTA. There was a relationship between dose and OS for CTX agents (univariate Cox regression HR=0.63 [95%CI 0.43-0.93], p=0.02) but not for MTA (HR=1.06 [95%CI 0.87-1.27], p=0.55). However in the multivariate model adjusted for clinical and demographic variables no relationship was found between dose and OS for either MTA or CTX. CONCLUSIONS These results support current development of novel CTX agents based on MTD. We found no direct relationship between higher doses and ORR or OS with MTA in unselected pts. There is an urgent need to identify better biomarkers to guide dose escalation trials of MTA.