Prediction model of ocular metastasis from primary liver cancer: Machine learning-based development and...

Area of Science:

• Oncology
• Medical Informatics
• Ophthalmology

Background:

• Ocular metastasis (OM) is a rare complication of primary liver cancer (PLC).
• Early detection of OM is crucial for improving patient prognosis.
• Predictive models can aid in identifying at-risk PLC patients.

Purpose of the Study:

• To develop a clinical predictive model for OM in PLC patients using machine learning (ML).
• To identify key clinical variables associated with OM development in PLC.

Main Methods:

• Retrospective analysis of 1540 PLC patients.
• Development and validation of six ML models, including extreme gradient boost (XGB).
• Evaluation using receiver operating characteristic (ROC) curves and construction of a web calculator.

Main Results:

• The XGB ML model demonstrated high diagnostic performance (AUC=0.993).
• Six key variables (CA125, ALP, AFP, TG, CA199, CEA) were identified as significant predictors.
• An online web calculator was created for personalized OM risk assessment.

Conclusions:

• The XGB-based model effectively predicts OM in PLC patients.
• Early identification of high-risk patients facilitates timely intervention.
• This tool can improve diagnosis, treatment, and quality of life for PLC patients with OM.