LAUSR

OBJECTIVES To assess the performance sentinel lymph node (SLN) biopsy and effect of ultrastaging in clinically early stage endometrial cancer. METHODS Patients with endometrial cancer prospectively enrolled after informed consent was obtained. The cervix was injected superficially with 1 mL of ISB and 1 mL of ICG (diluted 1:25) at 3 and 9 o'clock each. SLN biopsy was followed by complete pelvic lymphadenectomy (aortic lymphadenectomy at the discretion of the surgeon). Lymph nodes (LNs) were analyzed by standard sectioning with H&E; ultrastaging of SLN was done retrospectively and blinded to treating physicians. RESULTS 204 patients received dye injections. In 184 (90.2%) patients at least one SLN was identified. Of all patients, 138 (68%) had bilateral mapping. In the patients with successful mapping of a hemipelvis, ICG detected SLNs in 83% and ISB in 64% of cases (p < 0.0001). Median BMI (kg/m2) for patients with successful mapping was 35.7 compared to 40.1 for those who did not map (p = 0.01). Twenty-three (11.3%) patients had positive LNs. Applying the SLN algorithm, positive nodes were detected in 21/23 (91.3%). The negative predictive value (NPV) was 98.9% (95% CI: 96.01% to 99.71%). Eleven patients had positive SLN with isolated tumor cells (ITCs) or micrometastases detected on ultrastaging. Including these patients, 34 (17%) had positive LNs, increasing the NPV to 99% and sensitivity to 94%. There were no recurrences in patients with ITCs only. CONCLUSIONS SLN assessment in endometrial cancer is feasible and safe with high NPV (99%). ICG was more effective in detecting SLN compared to ISB. Although ultrastaging detected additional positive LNs, treatment based on standard sectioning appears reasonable but further research is needed.