95 clinical trials of anticancer drugs were done in 40 patients with advanced ovarian cancer who had undergone an in-vitro soft-agar human tumour stem-cell assay for drug sensitivity. In the 95 clinical trials, 13 of 21 patients responded clinically to drugs predicted to be effective in vitro, while in 73 of 74 instances in which drug resistance had been predicted in vitro there was no response to treatment. 2 patients achieved a complete remission of 1 and 12+ months and 11 had partial remissions lasting a median of four months. Prediction of in-vitro resistance was 99% accurate. The 62% complete or partial remission rate to agents with proven in-vitro efficacy was achieved in extensively pretreated ovarian cancer patients who were unlikely to respond to empirically chosen secondary agents. The human tumour stem-cell assay, in which ovarian tumour colony-forming units are grown in a two-layer agar-culture system incorporating tests for drug sensitivity or resistance, can be effectively used to plan a suitable chemotherapy regimen for ovarian cancer patients.