p53 Mutations in epithelial ovarian cancers: Possible role in predicting chemoresistance

PURPOSE The investigators undertook a retrospective study to determine (1) whether p53 mutations are predictors of survival in patients with advanced epitheli al ovarian cancer, (2) whether p53 status by sequencing is associated with established prognostic indicators, and (3) the agreement of results between direct sequencing of p53 mutations and immunohistochemistry. MATERIAL AND METHODS This study was retrospective review of 43 patients with advanced epithelial ovarian cancer treated with surgery and by paclitaxel-based chemotherapy. Clinical data were extracted from the charts. By use of paraffin-embedded b locks, p53 analysis was carried out by (1) direct sequencing and (2) immuno histochemistry. Kaplan-Meier estimates were used for overall and disease-fr ee survivals. To determine whether p53 mutation (sequencing) was related to prognosis and to determine the agreement between p53 abnormalities by sequ encing and immunohistochemistry, risk ratios were calculated. RESULTS Mean age at diagnosis was 57.4 years. Surgical stages were as follows: 5% w ere stage IIC, 79% were stage III, and 16% were stage IV. Seventy-seven per cent of tumors were serous, and 56% of tumors were grade 3. All patients re ceived paclitaxel-based chemotherapy. Mean disease-free and overall surviva ls were 16.4 and 22.6 months, respectively, p53 abnormalities were detected by sequencing in 53% of cases and by immunohistochemistry in 70%. Agreemen t between both techniques was 68%. Patients with stages IIIC/IV had a risk of 1.7 of having a p53 mutation by sequencing; grade, histology, disease-fr ee survival and overall survival were not predictive of p53 mutation status . DISCUSSION The 54% mutation rate may he underestimated by limiting our analysis to exo ns 5 to 9, p53 mutation status was not predictive of survival (disease free and overall) or of chemoresistance; this suggests that paclitaxel-based ap optosis is independent of the p53 gene. The concomitant use of cisplatin. a n inducer of apoptosis that is dependent on the normal function of p53, mak es the interpretation of these results difficult. Histopathologic factors w ere not statistically associated with p53 status, but more advanced surgica l stage and tumor grade were suggestive of higher rates of p53 mutations, i mplying more aggressive behavior. Finally, the lack of agreement between re sults obtained by sequencing and immunohistochemistry highlights the limita tion of the latter technique and the possibility of underestimating abnorma l p53 function. In conclusion, because discrepancies exist between the two techniques, we recommend direct sequencing of exons 4 through 10 to determi ne the true prevalence of p53 mutation. Furthermore, larger randomized stud ies are required to elucidate the role of p53 in predicting chemoresponse i n advanced epithelial ovarian cancer (Cancer J 2000;6:302-308).