Yang R. et al., 2025: Predicting ESWL success for ureteral stones: a radiomics-based machine learning approach.
Ran Yang 1, Dan Zhao 1, Chunxue Ye 1, Ming Hu 2, Xiao Qi 2, Zhichao Li 3
1Department of Radiology, Chongqing Western Hospital, No. 301, Huafu Avenue North, Jiulongpo District, Chongqing, 400050, China.
2Department of Radiology, Second People's Hospital of Jiu Long Po District, No. 318 Huayu Road, Jiulongpo District, Chongqing, 400052, China.
3Department of Radiology, Chongqing Western Hospital, No. 301, Huafu Avenue North, Jiulongpo District, Chongqing, 400050, China.
Abstract
Objectives: This study aimed to develop and validate a machine learning (ML) model that integrates radiomics and conventional radiological features to predict the success of single-session extracorporeal shock wave lithotripsy (ESWL) for ureteral stones.
Methods: This retrospective study included 329 patients with ureteral stones who underwent ESWL between October 2022 and June 2024. Patients were randomly divided into a training set (n = 230) and a test set (n = 99) in a 7:3 ratio. Preoperative clinical data and noncontrast CT images were collected, and radiomic features were extracted by outlining the stone's region of interest (ROI). Univariate analysis was used to identify clinical and conventional radiological features related to the success of single-session ESWL. Radiomic features were selected using the least absolute shrinkage and selection operator (LASSO) algorithm to calculate a radiomic score (Rad-score). Five machine learning models (RF, KNN, LR, SVM, AdaBoost) were developed using 10-fold cross-validation. Model performance was assessed using AUC, accuracy, sensitivity, specificity, and F1 score. Calibration and decision curve analyses were used to evaluate model calibration and clinical value. SHAP analysis was conducted to interpret feature importance, and a nomogram was built to improve model interpretability.
Results: Ureteral diameter proximal to the stone (UDPS), stone-to-skin distance (SSD), and renal pelvic width (RPW) were identified as significant predictors. Six radiomic features were selected from 1,595 to calculate the Rad-score. The LR model showed the best performance on the test set, with an accuracy of 83.8%, sensitivity of 84.9%, specificity of 82.6%, F1 score of 84.9%, and AUC of 0.888 (95% CI: 0.822-0.949). SHAP analysis indicated that the Rad-score and UDPS were the most influential features. Calibration and decision curve analyses confirmed the model's good calibration and clinical utility.
Conclusion: The LR model, integrating radiomics and conventional radiological features, demonstrated strong performance in predicting the success of single-session ESWL for ureteral stones. This approach may assist clinicians in making more accurate treatment decisions.
BMC Med Imaging. 2025 Jul 4;25(1):268. doi: 10.1186/s12880-025-01817-8. PMID: 40615969; PMCID: PMC12228301
Comments 1
Authors of this paper are two radiologists. Rumours are that radiologist will be among the firstto lose their jobs because machines can interpret images more reliable and can better deal with simple numbers on which the output of images is based. “In this paper 1,595 radiomic features were extracted from the images”, “enabling the identification of lesion characteristics that are not visually discernible” by radiologists. Seventeen lesion characteristics were finally extracted to serve as material in the statistics and generated graphs. Understanding the meaning of the numbers - like with Hounsfield units - is probably impossible and unnecessary. The translation into clinical practice is pure mathematics just as the correlation to clinical results which are not causal, which can probably never be proven. In this and all similar papers we have to wait if the future shows if they are just forgotten or - until their results and promises - are confirmed in prospective studies - for every lithotripter.
I was surprised to see in Supplementary Fig. 4 that “the transverse diameter of the renal pelvis (RPW) at its widest level, measured on axial images perpendicular to its longitudinal axis” was selected and confirmed as a significant predictor. The transverse diameter of unobstructed renal pelvis has a natural large variety of up to several cm while the anterior-posterior diameter of the renal pelvis is typically used to determine obstruction (1,2) and varies only by a few mm in unobstructed kidneys.
1 Hodhod A, Eid H, Capolicchio JP, Petrella F, Sadri I, El-Sherbiny M, Jednak R, Lacroix C. How can we measure the renal pelvic anteroposterior diameter in postnatal isolated hydronephrosis? J Pediatr Urol. 2023 Feb;19(1):75-82. doi: 10.1016/j.jpurol.2022.08.007. Epub 2022 Aug 20. PMID: 36100553.
2 Popiolek M, Jendeberg J, Olin M, Wagenius M, Sundqvist P, Lidén M. Advancing decision-making in shock wave lithotripsy for upper ureteral stones: the role of radiological stone impaction markers. Urolithiasis. 2025 Jul 17;53(1):139. doi: 10.1007/s00240-025-01797-y. PMID: 40676249; PMCID: PMC12271247.
Peter Alken