A Novel Prediction Model for Monorchidism in Boys With Unilateral Nonpalpable Testes: Diagnostic Utility of Preoperative Ultrasonography

INTRODUCTION: In boys with non-palpable testis (NPT), the affected side may present as a viable testis (VT) or a nonviable testis (NVT), the latter representing true monorchidism. This study evaluated the utility of ultrasonography (US) for predicting monorchidism and testicular location in unilateral NPT. METHODS: We retrospectively reviewed boys with unilateral NPT who underwent preoperative US between 2009 and 2025. Testicular volumes were converted to age-adjusted standard deviation (SD) scores. A rule-based model classified NVT when the affected testis was visualized with an SD score ≤ cutoff A and the unaffected testis had an SD score ≥ cutoff B. If the affected testis was not visualized, an SD score ≥ cutoff C for the unaffected testis indicated NVT; remaining cases were classified as VT. Predictions were compared with intraoperative findings. Receiver operating characteristic analysis determined optimal cutoffs. Internal validation using repeated stratified five-fold cross-validation and bootstrap resampling assessed model performance and threshold stability. US-based localization was also compared with surgical findings. RESULTS: Sixty-nine boys were included; surgery identified 51 NVT and 18 VT. Optimal cutoffs were A = -2.60, B = +0.45, and C = +0.60. The model achieved an AUC of 96.1%, with 100% sensitivity, 92.2% specificity, 81.8% positive predictive value, and 100% negative predictive value. Internal validation showed median AUCs of 90.0% and 89.4%, with stable thresholds. Localization accuracy was 66.6%. CONCLUSION: A rulebased US model may predict monorchidism in unilateral NPT with a very low risk of misclassifying VT as NVT, despite its limited ability to determine testicular location.