Chemotherapy for recurrent ovarian carcinoma (ROC) produces response rates of 10-80% depending on the prevalence of platinum resistance. Most patients relapse within 1 year and median progression-free survival (PFS) is generally no more than 6 months. Previous pretherapeutic chemosensitlvity assays mostly failed to improve the outcome of patients with ROC. Newly developed ATP assays show promising retrospective correlation with clinical outcome. We report here the first results of ATP assay-directed chemotherapy in patients with ROC. Therapy was selected by the ATP tumor chemosensitivity assay (ATP-TCA) in a prospective open-label pilot trial for ROC. Objective response rate (ORR), PFS and overall survival (OAS) of the first 25 evaluable patients were retrospectively compared with those of 30 others having similar characteristics who were treated empirically within the same period. The actuarial median observation times were 80 weeks for the ATP-TCA group and 83.5 weeks for the control group, respectively. In the control group, a 37% ORR [two complete responses (CR) and nine partial responses (PR)] was followed by a median PFS of 20 weeks and a median OAS of 69 weeks, mainly related to the use of single-agent chemotherapy. The ORR in the ATP-TCA group was 64% (eight CR and eight PR) (p=0.04) with the majority of responses (11 of 16) achieved with novel combinations. The median PFS in this group was 50 weeks (p=0.003) and the median OAS was 97 weeks (p=0.145). Survival of responding patients was similar in both groups. Chemotherapy guided by the ATP-TCA produced a greater benefit with regard to both ORR and PFS in platinum-refractory patients. ATP-TCA-directed chemotherapy for ROC compares favorably with chemotherapy chosen by a clinician and often leads to the choice of novel drug combinations. These promising results now warrant confirmation by prospective randomized trials.