Abstract
BACKGROUNDS: The preoperative differentiation of borderline ovarian epithelial tumors (BOETs) remains challenging. While systemic inflammatory markers such as the platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), and monocyte-to-lymphocyte ratio (MLR) have shown prognostic value in ovarian cancer, their diagnostic potential in BOETs and the efficacy of a multimodal panel incorporating them have not been fully elucidated. This study aimed to evaluate the predictive value of combining these inflammatory markers with tumor markers (CA125, HE4) and specific ultrasonographic features for the diagnosis of BOETs. RESULTS: The study enrolled 301 patients (101 benign, 99 BOETs, 101 malignant). The NLR and MLR in the BOETs group were significantly higher than in the benign group but lower than in the malignant group (P < 0.001). CA125 and HE4 levels in the BOETs group were higher than in the benign group but substantially lower than in the malignant group (P < 0.05). The microcystic pattern was a highly specific feature for BOETs, with its prevalence being significantly higher in the borderline group (53.54%) compared to the benign (0.99%) and malignant (6.93%) groups (P < 0.001). The prevalence of a predominantly solid component, rich blood supply, irregular tumor morphology, and ascites showed a progressive increase from the benign to the borderline and malignant groups, with significant differences among all three (P < 0.001). The combination of inflammatory and tumor markers yielded AUCs of 0.714 for differentiating benign from borderline tumors and 0.861 for malignant from borderline tumors, outperforming individual markers (P < 0.05). Combining the microcystic pattern with the count of other positive sonographic features significantly improved diagnostic performance over either feature alone for both benign vs. borderline (AUC = 0.886) and malignant vs. borderline (AUC = 0.914) differentiations (P < 0.05). The comprehensive multimodal approach (inflammatory markers, tumor markers, microcystic pattern, and count of positive sonographic features) demonstrated robust diagnostic efficacy, with an AUC of 0.914 for benign vs. borderline differentiation and an AUC of 0.933 for malignant vs. borderline differentiation (P < 0.001). CONCLUSIONS: A multimodal approach integrating PLR/NLR/MLR, CA125, HE4, and the ultrasonographic microcystic pattern significantly enhances the preoperative differentiation of BOETs, and reveal that BOETs have biological characteristics of "inflammatory activation" to a certain extent.