1. A multicenter study to develop a non-invasive radiomic model to identify urinary infection stone in vivo using machine-learning
- Author
-
Xiaoguang Zou, Ya Qiu, Jianqiu Kong, Junjiong Zheng, Aierken Tuerxun, Tianxin Lin, Jesur Batur, Abudukeyoumu Abulajiang, Zhuo Wu, Sihong Lu, Lifang Huang, Shaoxu Wu, Hao Yu, and Zhenfeng Shi
- Subjects
0301 basic medicine ,medicine.medical_specialty ,Urinary infection ,business.industry ,Non invasive ,030232 urology & nephrology ,Logistic regression ,Confidence interval ,03 medical and health sciences ,030104 developmental biology ,0302 clinical medicine ,Decision curve analysis ,Multicenter study ,Nephrology ,Medicine ,Radiology ,Internal validation ,business ,Selection operator - Abstract
Urolithiasis is a common urological disease, and treatment strategy options vary between different stone types. However, accurate detection of stone composition can only be performed in vitro. The management of infection stones is particularly challenging with yet no effective approach to pre-operatively identify infection stones from non-infection stones. Therefore, we aimed to develop a radiomic model for preoperatively identifying infection stones with multicenter validation. In total, 1198 eligible patients with urolithiasis from three centers were divided into a training set, an internal validation set, and two external validation sets. Stone composition was determined by Fourier transform infrared spectroscopy. A total of 1316 radiomic features were extracted from the pre-treatment Computer Tomography images of each patient. Using the least absolute shrinkage and selection operator algorithm, we identified a radiomic signature that achieved favorable discrimination in the training set, which was confirmed in the validation sets. Moreover, we then developed a radiomic model incorporating the radiomic signature, urease-producing bacteria in urine, and urine pH based on multivariate logistic regression analysis. The nomogram showed favorable calibration and discrimination in the training and three validation sets (area under the curve [95% confidence interval], 0.898 [0.840-0.956], 0.832 [0.742-0.923], 0.825 [0.783-0.866], and 0.812 [0.710-0.914], respectively). Decision curve analysis demonstrated the clinical utility of the radiomic model. Thus, our proposed radiomic model can serve as a non-invasive tool to identify urinary infection stones in vivo, which may optimize disease management in urolithiasis and improve patient prognosis.
- Published
- 2021
- Full Text
- View/download PDF