Chi-Wei Chen et al., 2024: Using ensemble learning and hierarchical strategy to predict the outcomes of ESWL for upper ureteral stone treatment
Chi-Wei Chen , Wayne-Young Liu, Lan-Ying Huang, Yen-Wei Chu
1Graduate Degree Program of Smart Healthcare & Bioinformatics, I-Shou University, Kaohsiung City, Taiwan; Department of Biomedical Engineering, I-Shou University, Kaohsiung City, Taiwan.
2Doctoral Program in Medical Biotechnology, National Chung Hsing University, Taichung City, Taiwan; Department of Urology, Jen-Ai Hospital, Taichung City, Taiwan.
3Doctoral Program in Medical Biotechnology, National Chung Hsing University, Taichung City, Taiwan.
4Doctoral Program in Medical Biotechnology, National Chung Hsing University, Taichung City, Taiwan; Graduate Institute of Genomics and Bioinformatics, National Chung Hsing University, Taichung City, Taiwan; Institute of Molecular Biology, National Chung Hsing University, Taichung City, Taiwan; Agricultural Biotechnology Center, National Chung Hsing University, Taichung City, Taiwan; Rong Hsing Research Center for Translational Medicine, Taichung City, Taiwan; Ph. D Program in Medical Biotechnology, National Chung Hsing University, Taichung City, Taiwan; Smart Sustainable New Agriculture Research Center (SMARTer), Taichung, 402, Taiwan.
Abstract
Urinary tract stones are a common and frequently recurring medical issue. Accurately predicting the success rate after surgery can help avoid ineffective medical procedures and reduce unnecessary healthcare costs. This study collected data from patients with upper ureter stones who underwent extracorporeal shock wave lithotripsy, including cases of successful as well as unsuccessful stone removal after the first and second lithotripsy procedures, and constructed prediction systems for the outcomes of the first and second lithotripsy procedures. Features were extracted from three categories of information: patient characteristics, stone characteristics, and extracorporeal shock wave lithotripsy machine data, and additional features were created using Feature Creation. Finally, the impact of features on the models was analyzed using six methods to calculate feature importance. Our prediction model for the first lithotripsy, selected from among 43 methods and seven ensemble learning techniques, achieves an AUC of 0.91. For the second lithotripsy, the AUC reaches 0.76. The results indicate that the detailed and binary information provided by patients regarding their history of stone experiences contributes differently to the predictive accuracy of the first and second lithotripsy procedures. The prediction tool is available at https://predictor.isu.edu.tw/ks.
Comput Biol Med. 2024 Jul 23:179:108904. doi: 10.1016/j.compbiomed.2024.108904. Online ahead of print. PMID: 39047504 FREE ARTICLE
Comments 1
The authors have used several learning methods to predict the outcome of SWL for proximal ureteral stones. Patient and stone data were used in the different mathematical algorithms. The details of SWL apparently were limited to kV and pulses, but without any other information on how the treatments were carried out.
Has the development of new lithotripters made SWL more complicated? If not. it seems difficult to understand the prediction tools discussed in this report. Basic education and training on how SWL should be used, and the need of lithotripter knowledge is most important in the reviewer’s opinion.
It is of note that the authors chose proximal ureteral stones for their calculations. From a clinical point of view, it had indeed been more interesting to see how the algorithms might have predicted the outcome of SWL of lower pole stones with different geometrical properties of the collecting system.
I must confess that the mathematical methods used in this report are beyond my own knowledge. But the basic question is if predictive methods of this kind are better and superior to clinical experience? It can be assumed that the purpose of the SWL treatments in this report mainly was used to confirm the value of the mathematical process.
Hans-Göran Tiselius