The prognostic value of image analysis in ovarian cancer.

Abstract

Histologic grading is very important for treatment decisions in ovarian cancer. All grading systems contain a significant subjective component, which could be reduced by including objective measurements into the diagnostic decision. Image analysis was used to determine nuclear area and ploidy distributions in 42 patients with epithelial ovarian cancer, and the results were related to tumor grade and clinical outcome. The mean nuclear area, mean optical density, number of hyperploid cells, and the standard deviation between measurements were significantly higher in Grade 2 and 3 tumors compared with Grade 1 tumors, in rapidly progressive tumors compared with less aggressive malignancies, and in recurrent tumors compared with primary lesions. The number of nuclei with very high DNA content was found to be of prognostic importance. Image analysis thus provides additional prognostic information in epithelial ovarian cancer.