A machine learning model for predicting lymph node positivity in ovarian cancer: development, validation, and...

Area of Science:

• Oncology
• Medical Informatics
• Machine Learning

Background:

• Ovarian cancer (OC) is a lethal gynecological malignancy with poor prognosis, often diagnosed at advanced stages.
• Lymph node involvement is a critical prognostic factor in OC, influencing treatment planning.
• Predicting lymph node positivity is challenging due to disease heterogeneity and limitations of traditional models with high-dimensional, imbalanced data.

Purpose of the Study:

• To develop and validate a machine learning model for predicting lymph node positivity in ovarian cancer.
• To identify key clinical predictors of lymph node involvement in OC patients.
• To create a user-friendly tool for clinical decision support in OC management.

Main Methods:

• Retrospective analysis of 26,844 OC patients from the SEER database (2000-2021).
• Development of a machine learning model using XGBoost, with SMOTE for class imbalance and LASSO for feature selection.
• External validation using an independent cohort from Fujian Provincial Maternity and Children's Hospital.

Main Results:

• The XGBoost model achieved high performance, with an AUC of 0.98 in the training set and 0.847 in external validation.
• Key predictive features identified include tumor size (≥5 cm), histological subtype, and chemotherapy.
• SHAP analysis indicated tumor size as the most influential factor in predicting lymph node positivity.

Conclusions:

• The study presents the first machine learning model for predicting OC lymph node positivity, validated on diverse cohorts.
• A free online calculator is available to assist clinicians in estimating lymph node positivity risk.
• The tool supports individualized treatment decisions, potentially improving patient outcomes; future research should incorporate molecular data.