The prognostic value of nuclear morphometric analysis in serous ovarian carcinoma.

The objective of this study was to determine whether nuclear morphometric data can predict survival, disease progression, and chemotherapeutic response in ovarian serous carcinoma. Nuclear morphometric parameters were determined from archival hematoxylin and eosin sections of 132 serous tumors. Clinicopathologic and morphometric parameters were evaluated as to their individual and independent prognostic value and prediction of chemotherapy response. Nuclear parameters were found to strongly correlate with extent of disease residuum, tumor grade, and FIGO stage. Univariate analysis revealed stage, grade, preoperative CA125, presence of ascites, extent of disease residuum, standard deviation of nuclear area (SDNA), nuclear perimeter (NP), SDNP, nuclear length (NL), nuclear breadth (NB), orthoferet, and equivalent diameter (ED) to be significant predictors of overall survival (OS) and disease-free survival (DFS). Grade, stage, extent of disease residuum, presence of ascites, SDNA, NP, NL, NB, and orthoferet were found to be significant predictors of chemotherapy response. Multivariate analysis revealed extent of disease residuum (P <or= 0.01) and ED (P = 0.002) to be significant predictors for OS. FIGO stage (P <or= 0.01) and ED (P = 0.039) were significant predictors of DFS. NL (P = 0.041) and extent of residual disease (P = 0.003) were the strongest predictors of chemotherapy response with correct classification rates of 68.8% and 70.3%, respectively. In all stages, nuclear morphometry was easy to perform and highly reproducible. Independent prognostic significance was achieved for OS and DFS analysis. Results also suggest that nuclear morphometry can provide significant information to predict chemotherapy response in platinum-treated serous ovarian cancer.