Starting dose selection and dose escalation for oncology small molecule first-in-patient trials: Learnings from a survey of approved drugs.

Abstract

e15648 Background: The ideal starting dose (SD) for an oncology first-in-patient (FIP) trial should be low enough to be safe but not too far removed from therapeutically relevant doses. A low SD combined with small dose increments could lead to a lengthy dose escalation (DE) and could expose patients unnecessarily to sub-therapeutic dosing. Over the past decade, the SD selection for oncology FIP trials was predominantly based on ICH S9 Guidance. This work attempts to assess the performance of current approaches in SD selection and DE methods for small molecule oncology drugs (SMOD), with an aim to identify potential opportunities for more efficient FIP DE methods and to minimize patients receiving sub-therapeutic dosing. Methods: A systematic review of the FIP trials of USFDA approved SMODs from 2010 to early 2019 was performed. The data sources included the FDA “USPI”, relevant “Summary basis of approval” documents, and published articles. The review focused on FIP trial data including the SD, DE design, the number of cohorts to the maximum tolerated dose (MTD) or the recommended phase 2 dose (RP2D), and the number of subjects in each cohort. These data were analyzed to evaluate the efficiency and appropriateness of current approaches for FIP SD selection and DE. Results: Of 53 SMODs approved between 2010 and early 2019, ~ 65% were kinase inhibitors and ~74% were approved for solid tumor indications. All SMODs appeared to follow ICH S9 Guidance for SD selection and none of SMODs had more than doubling for dose increments in the escalation. The median MTD to SD ratio was 8 (1 to 125). A majority (57%) of the trials used a 3+3 design for DE and had a median of 4 cohorts (0 to 11) to reach MTD or RP2D from the SD. About 30% of the FIP trials had ≥ 6 DE to reach the MTD or RP2D, and in this subset, greater than 55% of the trial participants received subtherapeutic dosing. Conclusions: The ICH S9 Guidance-based SD and the DE methods for SMODs have generally performed well in most cases; the SD was well tolerated and the MTD or RP2D were achieved with reasonable efficiency. However, for a portion of SMODs the SD appeared to be too low or the DE was too conservative, suggesting opportunities for improvement to minimize the number of patients dosed at sub-therapeutic doses.