1180 Single institution validation of a sentinel lymph node biopsy risk prediction tool for cutaneous melanoma

Cutaneous melanoma clinical practice guidelines generally recommend a 10% risk threshold to offer sentinel lymph node biopsy (SLNB), with improved stratification desirable reduce unnecessary procedures. A retrospective analysis was conducted among patients undergoing SLNB at an academic center between June 2010 and December 2021 compare rates of SLN-positivity the validated Melanoma Institute Australia (MIA) prediction tool that incorporates patient age, thickness, ulceration, mitotic rate, subtype, lymphovascular invasion (melanomarisk.org.au). The cohort included 978 complete histopathologic data: 578 men, 400 women, mean age 61.5 yrs, 87.2% White. Calculated scores were analyzed as continuous variable. Across histologic subtypes (superficial spreading [51.7%], nodular [25.6%], lentigo maligna [10.2%], acral [7.4%] desmoplastic [2.9%]), calculator significantly underestimated SLN-positivity, most notably for (14% true positive vs 8% calculated positive), superficial (24% 16%), (30% 22%), (11% 2%) per t test (p<0.001 all). When stratified into equal groups (n=326) low-, medium- high-risk scores, highest in younger (mean 56 years) high tumor rate (mean: 7.95/mm2) ulceration 47.5%) analyzing <5% (N=123), 13% SLNB-positive; those 5-10% (N=250), 14% SLNB-positive, >10% (N=605), 33% SLNB-positive. While limited by single institution data, our study suggests despite rigorous attempts improve classification, there is need refinement increase accuracy MIA predictive model.