Objectives Sentinel lymph node (SLN) removal may be a midterm between no and full pelvic dissection in early endometrial cancer. Whereas the use of blue dye alone in SLN detection has a poor accuracy, its integration in an SLN algorithm may yield better results and overcome hurdles such as the requirement of nuclear medicine facility.

Methods Sixty-six patients with clinical stage I endometrial cancer were prospectively enrolled in a multicentre study between May 2003 and June 2009. Patent blue was injected intraoperatively into the cervix. We retrospectively assessed the accuracy of a previously described SLN algorithm consisting of the following sequence: (1) pelvic node area is inspected for removal of all mapped SLN and (2) excision of every suspicious non-SLN, (3) in the absence of mapping in a hemipelvis, a standard ipsilateral lymphadenectomy is then performed.

Results Sentinel nodes were identified in 41 patients (62.1%), mostly in interiliac and obturator areas. None was detected in the para-aortic area. Detection was bilateral in 23 cases (56.1%). Seven patients (10.6%) had positive nodes. The false-negative rate was 40% using SLN detection alone. When the algorithm was applied, the false-negative rate was 14.3%. The use of a SLN algorithm would have avoided 53% of lymphadenectomies

Conclusion Our multicentric evaluation validates the use of a SLN algorithm based on blue-only sentinel node mapping in early-stage endometrial cancer. The application of such SLN algorithm should be evaluated in a prospective context and might lead to decrease unnecessary lymphadenectomies.