Objective Sentinel lymph node (SLN) mapping with indocyanine green (ICG) detected by robotic near infrared (NIR) imaging is a feasible technique. The optimal site of injection (cervical or endometrial) for endometrial cancer has yet to be determined. We prospectively evaluated SLN mapping after cervical and endometrial injections of ICG to compare the detection rates and patterns of nodal distribution.

Methods Twenty-nine subjects with endometrial cancer undergoing robotic hysterectomy with lymphadenectomy by a single surgeon received SLN mapping with robotic fluorescence imaging. Seventeen patients received cervical injections of 1 mg of ICG and 12 patients received hysteroscopic endometrial injections of 0.5-mg ICG. Detection rates between the 2 groups were compared using Fisher exact tests. Continuous variables such as operating room times and body mass index were compared using t tests.

Results The SLN detection rate was 82% (14/17) for cervical and 33% (4/12) for hysteroscopic injection (P = 0.027). Sentinel lymph nodes were seen bilaterally in 57% (8/14) of the cervical injection group and 50% (2/4) of the hysteroscopic group. Para-aortic SLNs were seen in 71% (10/14) of patients who mapped after cervical injection and 75% (3/4) patients who mapped after hysteroscopic injection. There was 1 false-negative SLN in the cervical injection group.

Conclusions Cervical ICG injection achieves a higher SLN detection rate and a similar anatomic nodal distribution as hysteroscopic endometrial injection for SLN mapping in patients with endometrial cancer.