[Computer-based morphometric image analysis of endometrial hyperplasia]

Former studies have shown that only 20% of endometrial hyperplasias progress to carcinoma. Objective criteria for predicting the outcome of hyperplasias have been lacking. Because hysterectomy has been the therapy of choice to avoid malignant development, overtreatment of these patients is still a clinical problem. A computer-based morphometric image analysis system has proved to be a considerable improvement for the pathologist in assessing which patients will develop cancer. The system is based on ten nuclear and twelve glandular architectural features. The three parameters with the best statistical power of discriminating between patients developing malignancy and those remaining healthy were combined into a morphometric classification rule (D-score or discriminating score). We give a brief description of this method and its background.