Predictors of sentinel lymph node mapping failure in endometrial cancer: a retrospective multicenter study

Objective: This study evaluated risk factors for indocyanine green (ICG) sentinel lymph node (SLN) mapping failure in endometrial cancer patients and assessed their value in predicting SLN mapping failure at the individual patient level. Methods: We conducted a retrospective study of endometrial cancer patients who underwent ICG-guided SLN biopsy. Risk factors for SLN mapping failure were identified using univariate and multivariate logistic regression analyses. Additionally, an imbalanced random forest (RF) algorithm was trained to predict SLN mapping success or failure. Predictive modeling was used to complement traditional statistical inference by translating population-level associations into patient-specific risk estimates. Results: A total of 190 patients were included, of whom 35 experienced SLN mapping failure. Univariate analyses showed that age 65 years and over (OR = 2.87, 95% CI 1.31-6.26; p = 0.007) and serous histological subtype (OR = 4.12, 95% CI 1.41-12.0; p = 0.006) were significantly associated with mapping failure. Multivariate analysis revealed that only age remained statistically significant (OR = 1.07 per year; 95% CI 1.02-1.13; p = 0.006). No significant associations were observed for obesity, morbid obesity, tumor grade, or FIGO stage. The discriminative ability of the RF model was poor (cross-validated AUC = 0.58, 95% CI 0.47-0.69). Conclusion: We suggest-on average-age 65 years and over and a serous histological subtype are risk factors for SLN mapping failure. However, the presence of these risk factors in an individual patient does not reliably predict SLN mapping failure. Additionally, more specific predictors may be needed for accurate individualized prediction.