Proteomic Validation of MUC4 in the Differential Diagnosis of Endometrial and Ovarian Serous Carcinomas.

Endometrial serous carcinoma (ESC) and high-grade ovarian serous carcinoma (HG-OSC) share morphologic, immunohistochemical, and molecular features, and the determination of the primary tumor site is challenging, particularly in cases with both uterine and ovarian involvement. This study aimed to evaluate the diagnostic utility of MUC4 protein expression for distinguishing ESC from HG-OSC. We retrospectively analyzed 79 resection specimens, comprising 36 ESC and 43 HG-OSC cases, including an additional 39 samples from extraprimary sites obtained from 6 ESC and 18 HG-OSC cases, which were also incorporated into the study. Immunohistochemical expression of MUC4 and WT1 was quantified as the percentage of positive tumor cells. Statistical analyses included the Mann-Whitney U test, independent samples t-test, χ2 test, Phi contingency coefficient, and multivariable binary logistic regression to assess the independent predictive value of MUC4 and WT1 for tumor origin. MUC4 expression was significantly more frequent in ESC than in HG-OSC (n=29/36, 81% patients in ESC vs. n=11/43, 26% in HG-OSC, P <0.001), whereas WT1 expression was predominantly observed in HG-OSC (n=37/41, 90% in HG-OSC vs. n=10/33, 30% in ESC, P<0.001). MUC4 expression was detected in all cases with an in-situ component associated with invasive ESC (n=5, 100%). Combined immunophenotypic analysis showed that the WT1-/MUC4+ profile was strongly associated with ESC (n=18/33, 55%), whereas the WT1+/MUC4- profile was characteristic of HG-OSC (n=32/41, 78%) (P<0.001). In multivariable logistic regression, MUC4 positivity was independently associated with endometrial origin (OR=20.5, 95% CI: 4.99-83.68, P<0.001), whereas WT1 positivity was associated with a decreased likelihood of endometrial origin (OR=0.12, 95% CI: 0.03-0.45, P=0.002). The optimal cut-off value was determined as ≤10% for MUC4, yielding a sensitivity of 86.0% and specificity of 77.8% (Youden index: 0.638), and ≥5% for WT1, with a sensitivity of 90.2% and specificity of 68.6%. Notably, the combined model showed superior diagnostic performance compared with either marker alone, with an AUC of 0.900 (95% CI: 0.820-0.964). At the optimal cut-off (≥0.513), the combined model achieved a sensitivity of 95.1% and specificity of 72.7%, indicating improved overall accuracy in differentiating between ESC and HG-OSC. This study demonstrates that MUC4, previously identified as differentially expressed in TCGA transcriptomic data, may have significant diagnostic value for distinguishing ESC from HG-OSC, both in the primary tumor site and in areas of multifocal involvement. MUC4, particularly when combined with WT1, is a promising immunohistochemical marker that may support accurate determination of tumor origin and may aid in diagnosis in challenging cases. Further validation in larger cohorts is warranted.