Selection of active drugs for ovarian cancer based on CA-125 and standard response rates in phase II trials.

Abstract

To determine whether precise definitions of response based on serial CA-125 levels can predict the activity of drugs in phase II trials for ovarian cancer as accurately as standard criteria. Fourteen different drugs for relapsed ovarian cancer were analyzed in 25 treatment groups in 19 clinical trials. Response rates were estimated in 1,457 assessable patients according to standard criteria and in 1,092 assessable patients according to CA-125. For each drug trial, the observed response rates acted as the input data for an evaluation of how the two criteria would perform in a hypothetical Gehan two-stage phase II trial, accepting a target drug efficacy rate of 20% and a rejection error of 5%. CA-125 and clinical response criteria were concordant in 20 of the 25 groups, with less than 5% chance of rejecting the drug in nine groups and greater than 5% in 11 groups. In four groups, the drug had less than 5% chance of being rejected by CA-125 but greater than 5% chance of being rejected by standard criteria. The difference in the classification of drugs by the standard and CA-125 response criteria was not statistically significant (P =.38, McNemar's test). CA-125 response rates were slightly higher than standard response rates by a factor of 1.11. Definitions based on a 50% or 75% decrease of CA-125 levels accurately predicted which drugs in phase II trials for relapsed ovarian cancer were active and justified further investigation.