Chen Z. et al., 2021: Automated Generation of Personalized Shock Wave Lithotripsy Protocols: Treatment Planning Using Deep Learning.
Chen Z, Zeng DD, Seltzer RGN, Hamilton BD.
Shenzhen Artificial Intelligence and Data Science Institute (Longhua), Longhua, Shenzhen, China.
The State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China.
Translational Analytics and Statistics, Tucson, AZ, United States.
School of Medicine, University of Utah, Salt Lake City, UT, United States.
Background: Though shock wave lithotripsy (SWL) has developed to be one of the most common treatment approaches for nephrolithiasis in recent decades, its treatment planning is often a trial-and-error process based on physicians' subjective judgement. Physicians' inexperience with this modality can lead to low-quality treatment and unnecessary risks to patients.
Objective: To improve the quality and consistency of shock wave lithotripsy treatment, we aimed to develop a deep learning model for generating the next treatment step by previous steps and preoperative patient characteristics and to produce personalized SWL treatment plans in a step-by-step protocol based on the deep learning model.
Methods: We developed a deep learning model to generate the optimal power level, shock rate, and number of shocks in the next step, given previous treatment steps encoded by long short-term memory neural networks and preoperative patient characteristics. We constructed a next-step data set (N=8583) from top practices of renal SWL treatments recorded in the International Stone Registry. Then, we trained the deep learning model and baseline models (linear regression, logistic regression, random forest, and support vector machine) with 90% of the samples and validated them with the remaining samples.
Results: The deep learning models for generating the next treatment steps outperformed the baseline models (accuracy = 98.8%, F1 = 98.0% for power levels; accuracy = 98.1%, F1 = 96.0% for shock rates; root mean squared error = 207, mean absolute error = 121 for numbers of shocks). The hypothesis testing showed no significant difference between steps generated by our model and the top practices (P=.480 for power levels; P=.782 for shock rates; P=.727 for numbers of shocks).
Conclusions: The high performance of our deep learning approach shows its treatment planning capability on par with top physicians. To the best of our knowledge, our framework is the first effort to implement automated planning of SWL treatment via deep learning. It is a promising technique in assisting treatment planning and physician training at low cost.
JMIR Med Inform. 2021 May 11;9(5):e24721. doi: 10.2196/24721. PMID: 33973862. FREE ARTICLE
This paper is published in an open access journal: JMIR Medical Informatics. (https://medinform.jmir.org/about-journal/about-the-publisher)
The impact factor is 2,96, rising (https://medinform.jmir.org/announcements/278). The article processing charge is $1900.
The data used stem from a not referenced source “of 171 physicians who recorded outcomes in the International Stone Registry, a database of accumulated treatment records for all patients treated within a national network of SWL services provided by Translational Analytics and Statistics, a lithotripsy service provider.” I could not find this “International Stone Registry” in the internet or any other publication.
The author quote from the literature preoperative patient characteristics (PPC) which influence the ESWL success like BMI, skin-to-stone distance, composition, density and location of stones or overall stone burden. However the factors they used for their calculations were only the usual treatment steps “ie, ternaries of a power level, a shock rate, and number of shocks”, and rather simple patient characteristics like “gender, age, stone location, stone size, mean arterial pressure, anticoagulant use, sedation use, whether multiple stones existed, and whether strapping was applied.” Like this an impressive first-glance number of data n=8583 from practices of renal SWL treatments from 54 physicians were used for their calculations. These were the 1216 cases with high success rates. It is all mathematics and statistics and at the end, there is no recommendation what to do with a patient. A conclusion is “clinical studies are warranted to confirm the effectiveness and efficiency of this framework.”
I assume that the data stem from a mobile lithotripter service in the USA. They were used to retrospectively assemble and mix the data of the successful cases in order to predict the success of future cases. If this really pays off, there will be a publication on the real life effects. However I'm in doubt about it.