21Oct 2023
Machine learning-based prediction model and visual interpretation for prostate cancer

Machine learning-based prediction model and visual interpretation for prostate cancer

Prostate cancer diagnosis relies on PSA testing, but its accuracy needs improvement. A model was built using XGBoost and patient data from the Chinese National Clinical Medical Science Data Center. Variables included age, BMI, PSA-related parameters, and serum biochemical parameters. The XGBoost model (AUC 0.82) outperformed f/tPSA (AUC 0.75), tPSA (AUC 0.68), and fPSA (AUC 0.61). SHAP analysis ranked f/tPSA as the most crucial variable, while inorganic phosphorus, potassium, CKMB, LDL-C, and creatinine were also significant. PCa risk thresholds were identified for f/tPSA, inorganic phosphorus, potassium, CKMB, LDL-C, and creatinine. The model's wide availability and high net benefit make it advantageous for underdeveloped areas, aiding in prostate cancer diagnosis and screening.

  • #urology

Like

Save

Share