Discovery and Development of a TGFβ3/IL-13 Plasma Risk Score for Preoperative Discrimination of Benign and Malignant...

Background/Objectives: Differentiating between benign and malignant ovarian adnexal masses before surgery is a significant diagnostic challenge. This study examined a wide range of circulating plasma molecules (immune and inflammatory) to identify potential markers to classify tumors as benign or malignant. The goal was to develop a proof-of-concept tool for preoperative risk stratification. Methods: Twenty-three plasma molecules were evaluated in women with benign lesions (n = 68) and women with malignant ovarian tumors (n = 25). To evaluate the behavior of each marker in a normal population, women without adnexal pathology were also included (n = 11). The results were compared using the Kruskal-Wallis test with Dunn post hoc pairwise comparisons, the Benjamini-Hochberg correction for multiple comparisons, and ROC analysis. TGF-β isoforms and CA125 were measured in the complete cohort (n = 104), while the remaining molecules were assessed in a reduced cohort (n = 51). Two classification approaches were developed using TGF-β3 and IL-13: a sequential two-step algorithm and an ordinal logistic regression model. Apparent performance was corrected for optimism using Harrell's bootstrap procedure. Results: Ten molecules reached statistical significance. Among the significant molecules, TGF-β3 identified ovarian lesions regardless of malignant status (AUC = 0.983), while IL-13 showed perfect malignancy-specific separation from benign cases (AUC = 1.000). Their combination in an ordinal logistic regression model produced three non-overlapping diagnostic strata across all three groups, significantly outperforming CA125, HE4, and ROMA at the malignancy-separation boundary (DeLong p < 0.05) in our exploratory cohort. Conclusions: TGF-β3 and IL-13 represent a biologically complementary marker pair whose combined ordinal regression model shows promising performance and warrants prospective validation in larger, multicenter independent cohorts as a preoperative risk-stratification tool for ovarian adnexal masses.