Sentinel-lymph-node mapping has been advocated as an alternative staging technique for endometrial cancer. The aim of this study was to measure the sensitivity and negative predictive value of sentinel-lymph-node mapping compared with the gold standard of complete lymphadenectomy in detecting metastatic disease for endometrial cancer.

In the FIRES multicentre, prospective, cohort study patients with clinical stage 1 endometrial cancer of all histologies and grades undergoing robotic staging were eligible for study inclusion. Patients received a standardised cervical injection of indocyanine green and sentinel-lymph-node mapping followed by pelvic lymphadenectomy with or without para-aortic lymphadenectomy. 18 surgeons from ten centres (tertiary academic and community non-academic) in the USA participated in the trial. Negative sentinel lymph nodes (by haematoxylin and eosin staining on sections) were ultra-staged with immunohistochemistry for cytokeratin. The primary endpoint, sensitivity of the sentinel-lymph-node-based detection of metastatic disease, was defined as the proportion of patients with node-positive disease with successful sentinel-lymph-node mapping who had metastatic disease correctly identified in the sentinel lymph node. Patients who had mapping of at least one sentinel lymph node were included in the primary analysis (per protocol). All patients who received study intervention (injection of dye), regardless of mapping result, were included as part of the assessment of mapping and in the safety analysis in an intention-to-treat manner. The trial was registered with ClinicalTrials.gov, number NCT01673022 and is completed and closed.

Between Aug 1, 2012, and Oct 20, 2015, 385 patients were enrolled. Sentinel-lymph-node mapping with complete pelvic lymphadenectomy was done in 340 patients and para-aortic lymphadenectomy was done in 196 (58%) of these patients. 293 (86%) patients had successful mapping of at least one sentinel lymph node. 41 (12%) patients had positive nodes, 36 of whom had at least one mapped sentinel lymph node. Nodal metastases were identified in the sentinel lymph nodes of 35 (97%) of these 36 patients, yielding a sensitivity to detect node-positive disease of 97·2% (95% CI 85·0–100), and a negative predictive value of 99·6% (97·9–100). The most common grade 3–4 adverse events or serious adverse events were postoperative neurological disorders (4 patients) and postoperative respiratory distress or failure (4 patients). 22 patients had serious adverse events, with one related to the study intervention: a ureteral injury incurred during sentinel-lymph-node dissection.

Sentinel lymph nodes identified with indocyanine green have a high degree of diagnostic accuracy in detecting endometrial cancer metastases and can safely replace lymphadenectomy in the staging of endometrial cancer. Sentinel lymph node biopsy will not identify metastases in 3% of patients with node-positive disease, but has the potential to expose fewer patients to the morbidity of a complete lymphadenectomy.

Indiana University Health, Indiana University Health Simon Cancer Center, and the Indiana University Department of Obstetrics and Gynecology.