Validation of a new ovarian malignancy suspicion index for preoperative evaluation of adnexal masses

Abstract

Background: The currently available ovarian malignancy probability scores incorporate biochemical markers such as CA 125 (Carbohydrate Antigen 125), which is not routinely available in peripheral centers. There is a need for tumour marker independent prediction model to differentiate malignant ovarian masses from their benign counterparts in order to plan appropriate surgery. To formulate and prospectively validate a new Ovarian Malignancy Suspicion Index (OMSI) independent of serum CA 125 level, in preoperative evaluation of adnexal masses admitted for surgery.Methods: This was a combined retrospective and prospective cohort study conducted in a tertiary referral hospital over a period of one and half years. Retrospective sample included 100 subjects who had undergone surgery for adnexal masses and who had definite histopathological report. Detailed data were obtained with respect to age, menopausal status, sonographic findings including solid areas, ascites, mean diameter, bilateralism, and presence of septa. A logistic multivariate regression analysis was carried out to find the best prediction score (OMSI - Ovarian Malignancy Suspicion Index). This model was further evaluated prospectively in 60 subjects for its diagnostic ability to identify benign and malignant ovarian pathology.Results: OMSI at the cut off value of 3.9 differentiated effectively malignant ovarian mass from benign variety with a good diagnostic performance (Sensitivity 100%, Specificity 90.5%, Positive Predictive Value 81.8% and Negative Predictive Value 100%) as good as currently recommended RMI (Risk Malignancy Index) score. It was also found that OMSI > 3.9 was associated with positive ultrasound evidence for ovarian malignancy such as presence of thick septae (90%), solid areas within the tumour (93.8%), papillary projections (100%), bilaterality (90%) and ascites (100%).Conclusions: This study shows that it is possible to derive ovarian malignancy prediction model such as OMSI without including CA 125 with diagnostic ability in par with risk scoring systems such as WHO recommended RMI. Using this model, physicians working in peripheral centers without facilities for estimating serum tumour markers can arrive at the possible diagnosis and plan appropriate management strategies.