The successful development of new oncology drugs requires phase I trials to accurately define the safety, tolerability, and most importantly maximum tolerated dose (MTD) of the study drug(s). Mischaracterization of the MTD can result in the failure of an otherwise promising drug due to perceived inactivity or excessive toxicity. Despite the importance of accurate MTD estimation, current trials designs frequently do not achieve this goal. One major reason is that current phase I eligibility criteria do not sufficiently control for potential confounders of dose-limiting toxicity (DLT) risk.
To address this significant limitation, we have previously leveraged a very large database of National Cancer Institute (NCI)–sponsored phase I studies and developed a nomogram to predict cycle 1 drug-related DLT risk. Our final nomogram includes both baseline patient characteristics (AST, albumin, WBC, estimated GFR, ECOG PS, and oncologic diagnosis [lymphoma yes/no]) and protocol characteristics (number of therapeutic agents, agent type [biologic yes/no]) to predict DLTs (Figure 1). Because all these factors are known at the time of enrollment, our nomogram can be used a priori to identify patients at significantly increased risk for DLT. This tool has the potential to redefine patient selection for phase I trials by improving patient safety while simultaneously increasing the accuracy of MTD estimation. By avoiding DLTs that require cohort expansion to demonstrate safety, this nomogram will allow fewer patients to be enrolled and therefore improve the speed and reduce the cost of phase I studies.
Figure 1: Nomogram to Predict Cycle 1 Drug-Related Dose-Limiting Toxicity on Phase I Trials
Although potentially practice changing, this nomogram must be validated in an independent dataset before it can be used routinely in practice. The purpose of the current research proposal is to retrospectively validate this DLT prediction nomogram in an independent population of MSKCC patients previously enrolled to phase I trials in the Developmental Therapeutics Service.
Preferred Project Dates
Students are expected to be in attendance all eight weeks, start and end dates inclusive.