OBJECTIVE The p53 gene plays a critical role in cellular response to DNA damage and has been implicated in the response to platinum compounds in ovarian carcinoma patients. Because taxanes could induce p53-independent apoptosis, we assessed the relevance of p53 gene status to response in ovarian carcinoma patients receiving paclitaxel and platinum-containing chemotherapy. METHODS Forty-eight previously untreated patients with advanced disease received standard paclitaxel/platinum-based chemotherapy. In tumor specimens collected at the time of initial surgery, before therapy, p53 gene status and expression were examined by single-strand conformation polymorphism, sequence analysis, and immunohistochemical analysis. Microsatellite instability analysis was performed on available samples from 30 patients. RESULTS Thirty-four (71%) of the 48 patients had a clinical response. Pathologic complete remission was documented in 13 (27%) of 48 patients. p53 mutations were detected in 29 (60%) of 48 tumors. Among the patients with mutant p53 tumors, 25 patients (86%) responded to chemotherapy. Only nine (47%) of 19 patients with wild-type p53 tumors responded to the same treatment. The overall response rate and the complete remission rate were significantly higher among patients with mutant p53 tumors than among patients with wild-type p53 tumors (P: =.008). Most of the tested tumors not associated with complete remission (10 of 12 tumors) were also characterized by microsatellite instability. The complete remission rate was higher among patients with tumors without microsatellite instability (five of seven patients). CONCLUSIONS In contrast to the limited efficacy of treatment with paclitaxel in combination with standard platinum doses against wild-type p53 ovarian tumors, patients with mutant p53 ovarian tumors were more responsive to paclitaxel-based chemotherapy. The pattern of response to chemotherapy containing paclitaxel is different from that associated with high-dose cisplatin therapy. Determining p53 mutational status can be useful in predicting therapeutic response to drugs effective in ovarian carcinoma.