Preoperative Imaging for Nodal Assessment in Endometrial Carcinoma: A Comparative Study of MRI and 18F-FDG PET/CT Across Risk Groups

Background and Objectives: Accurate preoperative assessment of lymph node metastasis is essential for surgical staging and treatment planning in endometrial carcinoma. Magnetic resonance imaging (MRI) is widely used for local staging; however, its nodal performance is limited by reliance on morphologic criteria. 18F-FDG PET/CT provides whole-body functional assessment and may support decision-making when sentinel lymph node (SLN) mapping is unavailable. This study compared MRI and 18F-FDG PET/CT for preoperative detection of pelvic and paraaortic nodal metastases across pragmatic histologic risk groups. Materials and Methods: This retrospective diagnostic-accuracy study included 255 consecutive surgically treated patients with histologically confirmed endometrial carcinoma who underwent both preoperative pelvic MRI and whole-body 18F-FDG PET/CT at a tertiary gynecologic oncology center between December 2023 and December 2025. All patients underwent pelvic lymphadenectomy, 111 patients (43.5%) also underwent paraaortic lymphadenectomy, and SLN mapping was not performed. Patients were categorized as Group 1 (Grade 1–2 endometrioid adenocarcinoma) or Group 2 (Grade 3 endometrioid or non-endometrioid/aggressive histology). Diagnostic estimates were reported with 95% confidence intervals, paired comparisons were performed using exact McNemar tests, and multivariable bias-reduced logistic regression was used to assess independent imaging associations with nodal metastasis. Results: Any nodal metastasis was documented in 24 patients (9.4%). 18F-FDG PET/CT showed higher sensitivity than MRI for patient-level nodal detection (91.7%, 95% CI 74.2–97.7 vs. 37.5%, 95% CI 21.2–57.3), with similarly high specificity (97.0%, 95% CI 93.9–98.5 vs. 97.8%, 95% CI 95.0–99.1). The NPV of 18F-FDG PET/CT was 99.1% (95% CI 96.8–99.8), compared with 93.8% (95% CI 90.0–96.2) for MRI. MRI sensitivity was particularly low in Group 1 (18.2%, 95% CI 5.1–47.7), whereas PET/CT retained high sensitivity in both groups, though the perfect performance observed in Group 2 had wide confidence limits. Conclusions: 18F-FDG PET/CT demonstrated superior sensitivity and NPV compared with MRI for preoperative nodal assessment in endometrial carcinoma. These results support the selective use of PET/CT as an adjunctive staging tool in settings without SLN mapping, but they should not be interpreted as evidence that PET/CT can replace histologic nodal staging.