Improving sentinel lymph node detection rates in endometrial cancer: how many cases are needed?

OBJECTIVE: To describe sentinel lymph node (SLN) detection rates in endometrial cancer and estimate how many cases are needed to achieve >90% SLN detection.
METHODS: We conducted a prospective study of patients undergoing primary surgery for endometrial cancer between September 2005 and March 2009. Lymph node mapping was performed using blue dye injection into the cervix in all cases. Additional injection methods included blue dye injection in the uterine fundus, and cervical injection of Tc99m. SLNs were identified and removed, followed by regional lymph node dissection. The results were analyzed according to two study periods: an "early" phase (September 2005-December 2007) and a "late" phase (January 2008-March 2009).
RESULTS: One hundred and fifteen patients with endometrial cancer were included. The cervix was the only site of injection in 82 cases (71%), while a combined cervical and fundal injection was performed in 33 cases (29%). Overall, SLN detection was achieved in 98 (85%) cases. In the initial 27 months of the study, a SLN was identified in 50 of 64 cases (78%), with 2 false negative cases. In the subsequent 15 months, successful mapping was achieved in 48 of 51 cases (94%) with no false negative cases. When examining an individual provider's performance, after the first 30 cases, the rate of successful mapping significantly increased from 77% to 94% (P=0.033).
CONCLUSION: Sentinel node mapping in uterine cancer requires a dedicated effort to achieve high detection rates. Surgeons should determine their individual detection rates and false negative rates. Our data demonstrate that high SLN detection rates can be achieved in women with uterine cancer and increasing surgical volume (30 cases) is associated with significantly increased detection rates.