Primary healthcare-friendly prostate cancer prediction model using routine clinical parameters: a multicenter study

Objective: This study aimed to develop and validate a nomogram model integrating routinely available parameters, making it applicable to primary healthcare settings for optimizing prostate cancer (PCa) risk stratification in patients with elevated prostate-specific antigen (PSA) levels, aiming to reduce unnecessary biopsies. Methods: = 281; 93 malignant and 188 benign cases) to assess the model's generalizability. Results: The nomogram model incorporated prostate volume, total PSA (tPSA), free-to-total PSA ratio (f/t PSA), and age, demonstrating discriminative performance with AUC values of 0.816 (95% CI: 0.796-0.836; training set), 0.833 (95% CI: 0.802-0.862; internal validation set), and 0.776 (95% CI: 0.720-0.832; external validation set). Its clinically acceptable performance outperformed that of individual parameters in the training set (all p-values <0.001 by Delong's test). The ROC curve and PR curve both demonstrated the robust predictive performance of the prediction model across all three study cohorts. The calibration curve showed strong agreement between the predicted probability and actual risk. DCA confirmed clinical net benefit across a wide range of risk thresholds. Conclusion: This nomogram provides a non-invasive, cost-effective, individualized PCa risk assessment tool for Chinese patients with elevated PSA levels. Its generalizability was confirmed through external validation, demonstrating effectiveness in optimizing biopsy decisions and suitability for primary healthcare settings, thereby reducing healthcare burden.