Comparison of nomogram with machine learning techniques for prediction of overall survival in patients with tongue cancer

The prediction of overall survival in tongue cancer is important for planning personalized care and patient counselling. This study compares the performance a nomogram with machine learning model to predict cancer. were built using large data set from Surveillance, Epidemiology, End Results (SEER) program database. comparison necessary provide clinicians comprehensive, practical, most accurate assistive system this population. used included records 7596 patients. considered algorithms logistic regression, support vector machine, Bayes point boosted decision tree, forest, jungle. These mainly evaluated terms areas under receiver-operating characteristic (ROC) curve (AUC) accuracy values. algorithm that produced best result was compared decision-tree outperformed other algorithms. When external validation data, tree an 88.7% while showed 60.4%. In addition, it found age patient, T stage, radiotherapy, surgical resection prominent features significant influence on model’s survival. provides more reliable prognostic information than nomogram. However, level transparency offered by estimating patients’ outcomes seems confident strengthened principle shared making between clinician. Therefore, combination – (NomoML) predictive may help improve care, patients, facilitates management-related decisions.