Random Forest Model Outperforms T.O.HO. Score for Predicting Stone-Free Rate After RIRS

A random forest model predicted stone-free status after retrograde intrarenal surgery with AUC 0.843, sensitivity 98.1%, vs T.O.HO. score AUC 0.615 (95% CI 0.556–0.674). Internal validation only; stone thickness and nephrostomy key predictors.

• RF model outperformed T.O.HO. score (AUC 0.843 vs 0.615; accuracy 87.5%, sensitivity 98.1%, specificity 53.1%).
• Key predictors included stone thickness, preoperative nephrostomy, stone length, stone width, and age.
• Limitation: Internal validation only; external prospective validation required before clinical use.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.