Abstract / Summary
Abstract Background To differentiate ovarian growing teratoma syndrome (OGTS) from immature teratoma (IMT) based on semantic CT imaging features and provide an adjunct imaging tool to support preoperative assessment and treatment planning. Materials and methods Patients with pathologically proven OGTS or IMT who underwent preoperative abdominal and pelvic CT were enrolled. Quantitative and qualitative CT features were compared between the two groups. Multivariable logistic regression was used to identify independent differentiating factors between OGTS and IMT, and a generalized linear mixed model was fitted to account for intra-patient clustering. Diagnostic performance was assessed using the area under the receiver operating characteristic curve (AUC) and compared using DeLong’s test. Internal validation of the combined model was performed using 1,000 bootstrap resamples within the same lesion-level development cohort to estimate optimism-corrected performance. Results A total of 121 lesions were included in this study, comprising 66 OGTS lesions and 55 IMT lesions. Pelvic effusion, ground-glass calcification, and quasi-circular calcification were identified as independent CT features for differentiating OGTS from IMT. The combined model incorporating these three predictors achieved an AUC of 0.818 (95% CI: 0.743–0.893), with a sensitivity of 71.2%, specificity of 81.8%, and overall accuracy of 76.0% at the optimal Youden index-derived cut-off value of 0.664. Conclusion The combined model based on CT imaging features showed promising ability for the preoperative discrimination of OGTS and IMT and may serve as an adjunct tool for individualized treatment planning.