Hirsch B. et al., 2020: Outcome groups and a practical tool to predict success of shock wave lithotripsy in daily clinical routine

Hirsch B, Abt D, Güsewell S, Langenauer J, Betschart P, Pratsinis M, Vetterlein MW, Schmid HP, Wildermuth S, Zumstein V.
Department of Urology, Kantonsspital St. Gallen, Rorschacherstrasse 95, 9007, St. Gallen, Switzerland.
Biostatistics, Clinical Trials Unit, Rorschacherstrasse 95, 9007, St. Gallen, Switzerland.
Department of Urology, University Medical Centre Hamburg-Eppendorf, Martinistrasse 52, 20247, Hamburg, Germany.
Department of Radiology and Nuclear Medicine, Kantonsspital St. Gallen, 9007, St. Gallen, Switzerland.
Department of Urology, Kantonsspital St. Gallen, Rorschacherstrasse 95, 9007, St. Gallen, Switzerland.

Abstract

Purpose: To improve outcome prediction of extracorporeal shock wave lithotripsy (SWL) by development of a model based on easily available clinical and radiographical predictors and suitable for daily clinical use.

Materials and methods: We evaluated predictive factors for SWL success in 517 consecutive patients suffering from urinary calculi who underwent SWL between 2010 and 2018. Analyses included descriptive statistics, receiver operating characteristic statistics and logistic regression. Predictive value was improved by combining parameters using model selection and recursive partitioning.

Results: Of the 517 patients, 310 (60.0%) had a successful SWL. Best individual predictor of SWL success was mean attenuation (MAV), with an area under the curve (AUC) of 0.668, and an optimal cutpoint (OC) of 987.5 HU. The best multivariable model, including MAV, stone size, skin to stone distance (SSD), presence of an indwelling stent, and four interaction effects, yielded an AUC of 0.736. Recursive partitioning would categorize patients into three outcome groups with high (76.9%), intermediate (41%) and low (10%) success probability. High probability of SWL success (76.9%) was found for patients with a stone with MAV ≤ 987 HU or with MAV > 987 HU but stone size ≤ 11 mm and SSD (45°) ≤ 88 mm.

Conclusion: A model based on four established predictors, and provided as an Excel®-Tool, can clearly improve prediction of SWL success. In addition, patients can be classified into three defined outcome groups based on simple cutpoint combinations. Both tools improve informed decision-making in daily clinical practice and might reduce failure rates.
World J Urol. 2020 May 20. doi: 10.1007/s00345-020-03253-5. PMID: 32436072. Online ahead of print.