Key drug development features in the design of early-phase oncology trials: An FDA perspective.

2597 Background: FDA drug approval requires demonstrated efficacy in adequate, well-controlled trials. In oncology, two key questions should be asked prior to initiation of “pivotal” efficacy trials: 1) which patient subsets should be enrolled into efficacy trials; and 2) what dose(s) should be tested? Improved phase 1 (P1) trials may help answer these questions and maximize likelihood for success in later development. We developed a knowledge management (KM) tool to aid FDA scientists in amassing a database of P1 trial characteristics, with the intention of developing standardized regulatory advice for future oncology drug development. Methods: We evaluated medical literature and investigational new drug (IND) P1 protocols submitted to FDA to determine key elements for database development. We then convened a panel of FDA clinical oncologists, pharmacologists, and pharmacogeneticists to determine what data should be captured to 1) allow comparisons of early drug development features for successful and unsuccessful development programs; and 2) inform future regulatory policy. Results: The KM tool was devised to capture information that could signal whether enrichment strategies should be considered or differential drug responses could be expected in later trials. In addition, the searchable database captures key elements from early phase oncology trials such as: a) molecule type and mechanism of action for targeted or non-targeted agents; b) molecular alterations known to associate with disease or drug target; c) PD assessment in tumor and surrogate tissues; d) molecularly- or clinically-enriched dose escalation and expansion cohorts; e) availability of a molecular test to select patients most likely to benefit from therapy; f) dosing schedule; g) dose escalation method and protocol-defined escalation endpoint; h) dose-limiting toxicity; i) observed responses; and j) recommended vs. actual phase 2 doses. Conclusions: P1 trials are important in helping guide drug development. We developed the first regulatory database of P1 trial design features to serve as a source for comprehensive assessment of early oncology drug development programs, which may be used for future policy development.