Development and assessment of machine learning algorithms for predicting postoperative outcomes in biliary tract cancers

1

The study developed a machine learning model to predict postoperative outcomes in patients with biliary tract malignancies using multiple risk factors.
2

Data from 6,951 patients in the SEER database and 245 patients from a medical center were utilized for model training and validation.
3

Independent prognostic factors for overall survival included age, sex, marital status, tumor site, differentiation, AJCC stage, and lymph node status.
4

The random forest model demonstrated strong performance in predicting survival rates, validated by AUC values and decision curve analysis.
5

Chemotherapy was associated with improved overall survival among patients with positive lymph nodes in both training and test cohorts.

1