Critical view of safety assessment in sentinel node dissection for endometrial and cervical cancer: artificial intelligence to enhance surgical safety and lymph node detection (LYSE study)
This study aims to evaluate the feasibility of video-based assessment rate of Critical Views of Safety criteria for sentinel lymph node dissection in endometrial and cervical cancer. Goal of these Critical Views of Safety is to help standardize the evaluation of surgical quality, improve the precision of sentinel lymph node identification, and lead to better patient outcomes. This international multi-center observational prospective study was conducted from April to September 2024. Surgical videos from patients with cervical and endometrial carcinoma undergoing minimally invasive sentinel lymph node dissection were collected. A total of 3 Critical Views of Safety criteria (lateral pararectal space, lateral paravesical space, internal iliac artery) were proposed based on the anatomical structures defined as mandatory to be identified before sentinel node dissection, according to previously published expert consensus. A total of 3 independent surgeons, blinded to each other's assessments, evaluated whether the proposed criteria were identifiable in the endoscopic surgical video to establish applicability (content validity) and inter-rater agreement (reliability). A total of 80 patients were enrolled, of these, 71 cases (88.8%) had videos suitable for annotation, 64 (90.1%) underwent sentinel lymph node dissection for endometrial cancer, and 7 (9.9%) for cervical cancer; the median age was 52 years (IQR 34-71) and median body mass index was 28.8 kg/m The low assessment rate of the internal iliac artery criteria should raise concerns about missing sentinel lymph nodes in the internal iliac and pre-sacral area. The assessment of such standardized safety criteria could potentially standardize the procedures, thereby improving adherence to guidelines. The introduction of the video assessment of these criteria lays the foundation for exploring the feasibility of artificial intelligence algorithms to automatically assess and document the Critical Views of Safety in surgical videos.