Sentinel lymph node assessment in endometrial cancer: a systematic review and meta-analysis.

BACKGROUND In the staging of endometrial cancer, controversy remains regarding the role of sentinel lymph node mapping compared with other nodal assessment strategies. OBJECTIVE We conducted a systematic review to evaluate the diagnostic accuracy and clinical impact of sentinel lymph node mapping in the management of endometrial cancer. DATA SOURCES We searched Medline, Embase, and the Cochrane Central Registry of Controlled trials for studies published in English before March 25, 2016 (PROSPERO CRD42016036503). STUDY ELIGIBILITY CRITERIA Studies were included if they contained 10 or more women with endometrial cancer and reported on the detection rate, sensitivity, and/or impact on treatment or survival of sentinel lymph node mapping. STUDY APPRAISAL AND SYNTHESIS METHODS Two authors independently reviewed abstracts and full-text articles for inclusion and assessed study quality. The detection rate, sensitivity, and factors associated with successful mapping (study size, body mass index, tumor histology and grade, injection site, dye type) were synthesized through random-effects meta-analyses and meta-regression. RESULTS We identified 55 eligible studies, which included 4915 women. The overall detection rate of sentinel lymph node mapping was 81% (95% confidence interval, 77-84) with a 50% (95% confidence interval, 44-56) bilateral pelvic node detection rate and 17% (95% confidence interval, 11-23) paraaortic detection rate. There was no difference in detection rates by patient body mass index or tumor histology and grade. Use of indocyanine green increased the bilateral detection rate compared with blue dye. Additionally, cervical injection increased the bilateral sentinel lymph node detection rate but decreased the paraaortic detection rate compared with alternative injection techniques. Intraoperative sentinel lymph node frozen section increased the overall and bilateral detection rates. The sensitivity of sentinel node mapping to detect metastases was 96% (95% confidence interval, 91-98); ultrastaging did not improve sensitivity. Compared with women staged with complete lymphadenectomy, women staged with sentinel lymph node mapping were more likely to receive adjuvant treatment. CONCLUSION Sentinel lymph node mapping is feasible and accurately predicts nodal status in women with endometrial cancer. The current data favors the use of cervical injection techniques with indocyanine green. Sentinel lymph mapping may be considered an alternative standard of care in the staging of women with endometrial cancer.