[Phase I trials--clinical studies of antineoplastic agents]

The establishment of the starting dose and the dose escalation are the principal issues of the Phase I trials of anticancer agents. We report the procedures and results of the Phase I studies we participated in Japan in the 1980's concerning 17 intravenous anticancer agents. The drugs indicated a correlation between the mouse LD10 and the man MTD (maximum tolerated dose) in mg/m2. Median mouse LD10 (135 mg/m2) approximated to median man MTD (137 mg/m2) in mg/m2. One fifth of the mouse LD10 was lower than the man MTD. Therefore, as recognized at the 23rd Annual Congress, Japan Society for Cancer Therapy, the lower one of either 1/5 the mouse LD10 or 1/3 the dog TDL (toxic dose low) in mg/m2 has to be determined as safe starting dose. The modified Fibonacci search scheme has been generally adopted for the dose escalation. 14 applicable drugs were examined including 7 drugs in the early 1980's and 7 drugs in the late 1980's. The real number of steps that reached the man MTD was compared to the number of the steps taken in the dose escalation by the Fibonacci's method. The real steps were more than the Fibonacci's ones in the late 1980's. It showed the tendency of a more careful and safer dose escalation, however, to put it critically, the dose escalation was not efficient enough. It is expected that the contradictory problem between safety and efficacy in the Phase I studies will be solved by developing methods like pharmacokinetic study in animals and man.