Photo Credit: Nasekom

The following is a summary of “Application of interpretable machine learning algorithm to predict lymph node metastasis in cutaneous malignant melanoma,” published in the April 2025 issue of Dermatology by Wang et al. 

Cutaneous malignant melanoma (CMM) was recognized as the deadliest skin cancer, where accurate prediction of lymph node metastasis played a key role in guiding personalized treatment and improving outcomes.  

Researchers conducted a retrospective study using interpretable machine learning to develop predictive models for lymph node metastasis in CMM by analyzing clinical, pathological, and biomarker data from the Surveillance, Epidemiology, and End Results (SEER) database.  

They analyzed data from 2,448 individuals with CMM in the SEER database to develop 6 machine learning models—support vector machine, random forest (RF), XGBoost, LightGBM, adaptive boosting, and gradient boosting decision tree—to predict lymph node metastasis. The key predictors were identified using Gaussian Naive Bayes and gradient boosting and used Shapley additive explanations (SHAP) for individualized model interpretation. Model performance was evaluated by accuracy, specificity, sensitivity, Brier score, and area under the receiver operating characteristic curve (AUC).  

The results showed that the RF algorithm executed the best performance, with AUC of 0.897, accuracy of 0.821, sensitivity of 0.876, specificity of 0.765, and a Brier score of 0.086. Key factors influencing predictions were chemotherapy, T stage, ulceration, pretreatment lactate dehydrogenase (LDH) levels, and radiation therapy. SHAP analysis revealed a strong association, emphasizing the role of LDH as a significant predictive biomarker.  

Investigators concluded that an accurate predictive model was established for lymph node metastasis in patients with CMM using machine-learning techniques. 

Source: karger.com/drm/article-abstract/doi/10.1159/000545959/925943/Application-of-interpretable-machine-learning