A scoring method for differentiating incidental ovarian endometriotic cysts from ovarian cystadenomas based on CT and clinical features

Detecting and distinguishing between ovarian endometriotic cysts and ovarian cystadenomas using computed tomography (CT) scans is a clinical challenge due to their similar CT characteristics. This study aims to identify key CT and clinical features that can effectively differentiate these two diseases and to develop a simple and practical scoring system. We conducted a retrospective analysis of 202 patients who underwent pre-operative contrast-enhanced CT at two medical centers. The subjects were divided into training cohort (n=151) and validation cohort (n=51). Utilizing univariate analyses and binary logistic regression, predictive factors for ovarian endometriotic cysts and ovarian cystadenomas were identified to construct an initial model. Subsequently, a scoring system was developed by assigning weights to each feature based on the initial model. The accuracy of both models (the initial model and the scoring system) was quantified by calculating the area under the receiver operating characteristic curve (AUC). We further determined the optimal cutoff point for the scoring system (0 – 15 points) and established three predictive ranges based on the probabilities of ovarian endometriotic cyst occurrence (8 points). The initial model, which incorporated two clinical factors (age and CA125level) and three CT characteristics (wall thickness, density heterogeneity, and edgeadhesion), demonstrated an AUC of 0.993. In the training cohort, the proportionsof patients correctly diagnosed with ovarian endometriotic cysts within these threescore ranges were 0%, 38.89%, and 96.83%, respectively, while in the validationcohort, they were 0%, 33.33%, and 95.00%. Thus, having a refined scoring range canhelp improve diagnostic accuracy, with a higher score indicating a diagnosis withovarian endometriosis cysts, and assist with differentiating patients with differentrisk levels.