Sentinel Lymph Node Mapping Without Intraoperative Frozen Section in Preoperative Endometrial Intraepithelial...

Endometrial intraepithelial neoplasia/atypical endometrial hyperplasia (EIN/AEH) carries a substantial risk of concurrent endometrial carcinoma at hysterectomy. Sentinel lymph node (SLN) mapping requires cervical injection into an intact uterus, so the decision to map must be made before uterine pathology is known. A strategy that defers that decision to intraoperative frozen section therefore does not choose between mapping and no mapping; it forecloses mapping altogether. We evaluated the technical feasibility and diagnostic yield of SLN mapping performed without frozen section in patients with preoperative EIN/AEH and estimated how much of the nodal disease identified might not have been detected under a frozen-section-guided pathway.

In this retrospective single-center study, 86 patients with a preoperative diagnosis of EIN/AEH underwent hysterectomy, bilateral salpingo-oophorectomy, and SLN mapping with methylene blue between November 2021 and May 2026. No patient underwent intraoperative frozen section, and all SLNs underwent ultrastaging. Patients were grouped by final histopathology, and carcinomas were staged according to FIGO 2023. A counterfactual analysis applied the Mayo intraoperative triage criteria to the final uterine pathology and considered two mechanisms separately: patients who would not have undergone nodal assessment at all, and patients who would have undergone lymphadenectomy but whose disease volume might not have been detected by conventional nodal pathology.

Final pathology showed persistent EIN/AEH in 37 patients (43.0%) and endometrial carcinoma in 49 (57.0%). Patients upgraded to carcinoma were older (60.3 ± 9.5 vs. 53.1 ± 10.4 years, p = 0.001), more often postmenopausal (87.8% vs. 54.1%, p = 0.001), and had greater endometrial thickness (18.6 ± 8.7 vs. 13.5 ± 5.4 mm, p = 0.004). At least one SLN was retrieved in 82 of 86 patients (95.3%) and bilaterally in 76 (88.4%); bilateral detection did not differ significantly between groups (94.6% vs. 83.7%, p = 0.177). SLNs lay in the obturator region in 62.2% of hemipelves and in the external iliac region in 26.3%, together accounting for 88.5%. Nodal involvement was identified in 11 patients (12.8% of the cohort; 22.4% of carcinomas): isolated tumor cells in two, micrometastasis in six, and macrometastasis in three. Among the patients who underwent additional non-sentinel nodal assessment, no metastatic non-SLN was identified in the presence of negative SLNs; no reference standard was applied to the remaining SLN-negative patients, so a false-negative rate cannot be estimated. No patient with EIN/AEH-only pathology had nodal involvement. Under simulated frozen-section triage, 2 of the 11 node-positive patients (18.2%) would not have undergone nodal assessment at all; a further 6 had low-volume disease that might not have been identified by conventional nodal pathology. Taken together, up to 8 of 11 (72.7%), and up to 6 of the 9 with micro- or macrometastasis (66.7%), might have remained undetected under the specified hypothetical pathway.

SLN mapping performed before hysterectomy and without intraoperative frozen section was technically feasible with methylene blue alone and identified occult nodal disease, a substantial part of which might not have been detected under a frozen-section-guided pathway. These findings establish feasibility and diagnostic yield only: the study has no comparison group and supports no conclusion about survival, morbidity, or cost-effectiveness. Because 43.0% of patients had EIN/AEH-only pathology, mapping should form part of individualized preoperative counseling rather than be applied universally.