Mouse and large-animal toxicology studies of twelve antitumor agents: relevance to starting dose for phase I clinical trials.

Abstract

Large-animal toxicology is presently used to establish a starting dose for clinical trials with new cancer chemotherapeutic agents. The relevance of dog, monkey, and mouse data for Phase I clinical trials has been retrospectively analyzed with twelve diverse agents (chlorozotocin, maytansine, anguidine, tritylcysteine, piperazinedione, Baker's antifol, thalicarpine, 3-deazauridine, gallium nitrate, cis-dichlorodiammineplatinum(II) (DDP), 4′-(9-acridinylamino)methanesulfon-manisidide (AMSA), and N-phosphonacetyl-l-aspartic acid (PALA). Schedules studied clinically included a daily x 5 schedule and a single dose schedule (three drugs), a daily x 5 schedule only (three drugs), and a single dose schedule only (six drugs). One-thrid of the toxic dose low (TDL) in the most sensitive large-animal species (dog or monkey), expressed in mg/m2, was a tolerable starting dose in humans in all instances for the schedules employed. The number of dose escalation steps to reach the human maximum tolerated dose (MTD), according to the commonly used Fibonacci dose escalation scheme, varied from 2 to more than 12. Had one-third the LD10 in mice, expressed in mg/m2 been applied, this would also have yielded safe starting dose levels, and would actually have required a lesser number of dose escalations to reach the human MTD. This analysis confirms that mouse data may be quite useful in determining safe starting doses for Phase I trials with anticancer chemotherapeutic agents.