Fig. 1
Construction and verification of the nomogram prediction model. (A)The nomogram prediction model for PUI in patients with bladder cancer. (B) ROC curve demonstrates the model’s strong ability to discriminate between patients with and without PUI. The AUC of 0.848 (0.8 after LOOCV) indicates excellent predictive accuracy. X-axis: Specificity (False Positive Rate), Y-axis: Sensitivity (True Positive Rate), Curve: The ROC curve, representing the sensitivity and specificity of the model at different thresholds. LOOCV, leave-one-out cross-validation. (C) Calibration curve demonstrated that the actual value closely aligns with the prediction model, suggesting good calibration. (D) Decision Curve Analysis (DCA) evaluates the clinical utility of the nomogram across a range of threshold probabilities. X-axis: Threshold Probability, Y-axis: Net Benefit. Nonadherence prediction nomogram: the net benefit curve of the model. All: the net benefit assuming all patients receive TUR biopsy. None: the net benefit assuming no patients receive TUR biopsy. LOOCV: the net benefit curve of the model after LOOCV. The higher the curve, the greater the clinical net benefit of using the model within that threshold range.
