Comprehensive Analysis and Prognostic Modeling of Epithelioid Trophoblastic Tumor.

Abstract

To evaluate clinicopathologic factors associated with progression-free survival in patients with epithelioid trophoblastic tumor and to construct validated prognostic models. This study combined a retrospective institutional case series (57 patients) with systematically identified literature cases (150 patients), totaling 207 individuals with epithelioid trophoblastic tumor diagnosed between 1998 and 2024. The primary end point was progression-free survival. Prognostic variables were identified with Cox and least absolute shrinkage and selection operator regression. Two nomograms were developed: a clinical model and a clinicopathologic model. Internal validation was performed with 1,000 bootstrap resamples. Model performance was assessed with concordance index, time-dependent receiver operating characteristic curves, calibration plots, and decision-curve analysis. Patients with stage IV disease had the shortest median progression-free survival (9.0 months); those with stage II-III disease and solitary extrauterine lesions had outcomes similar to those of patients with stage I disease (P=.219). Independent prognostic factors included high mitotic count (5 or more per 10 high-power fields; hazard ratio [HR] 17.91, 95% CI, 4.13-77.70, P<.001), interval from last pregnancy of 20 months or more (HR 2.59, 95% CI, 1.32-5.08, P=.006), tumor size 4 cm or larger (HR 1.81, 95% CI, 1.05-3.15, P=.034), and International Federation of Gynecology and Obstetrics stage IV (HR 4.46, 95% CI, 1.67-11.93, P=.003). Model discrimination (C index) was 0.72 (95% CI, 0.69-0.75) for the clinical model and 0.82 (95% CI, 0.78-0.86) for the clinicopathologic model. Clinicopathologic characteristics are significantly associated with prognosis in epithelioid trophoblastic tumor. The validated nomogram-based model may support individualized clinical risk assessment.