Phase I trials of molecularly targeted agents: should we pay more attention to late toxicities?

Phase I trials traditionally aim at determining the recommended phase II dose (RP2D) using grade ≥ 3 toxicity data from cycle 1 (C1) only. This design dates from the era of conventional chemotherapy and may not be relevant for new molecularly targeted agents (MTAs) usually administered in a chronic fashion and for which late or moderate toxicities may deserve particular attention.

All consecutive patients treated in phase I trials of MTAs at the Royal Marsden Hospital and Institut Gustave Roussy between January 2005 and July 2008 were included. Gastrointestinal, skin, and clinical renal toxicities of any grade and grades 3 to 4 toxic events of any type occurring at any cycle on treatment were recorded. Doses administered, treatment interruptions, dose modifications, and prescription of comedications were analyzed.

A total of 445 patients (1,566 cycles; median treatment duration, 55 days) were included in 36 eligible trials; 790 toxicities (590, grade 1; 176, grade, 2; 24 grades, 3 to 4) and 1,819 toxicities (1,521, grade 1; 265, grade 2; 33, grades 3 to 4) were recorded during and after C1, respectively; 57% of the grades 3 to 4 toxicities occurred after C1; 50% of patients presented their worst-grade toxicity after C1. The risk of grades 3 to 4 toxicity was 3% in cycles 1 to 6 and was almost null afterwards. No cumulative toxicities were observed. Median toxicity duration was 15 days, with comedication administered in 68% of events.

Moderate and severe toxicities occur regularly after the first cycle in phase I trials of MTAs and may deserve increased attention in the RP2D process for these agents.