Radiomics capabilities in the interpretation of ultrasound and CT data in patients with chronic kidney disease: A review

The purpose of this review is to explore the possibilities of radiomics in interpreting ultrasound and multislice spiral computed tomography data in patients with chronic kidney disease (CKD). Radiomics is a promising area of medical image analysis based on the extraction of quantitative features not available in standard visual analysis and the subsequent use of artificial intelligence methods for their processing and interpretation. The article discusses the basics of radiomic methods, including texture analysis of images and the creation of diagnostic models using machine learning algorithms. The advantages of radiomic characteristics, in particular statistical features of order II and higher orders, in assessing interstitial fibrosis and other abnormal changes in the renal parenchyma are discussed in detail. The results of studies demonstrating a strong correlation of radiomic signs with histological changes detected during kidney biopsy are presented. The prospects of radiomics as a non-invasive approach for assessing kidney damage and monitoring CKD progression are emphasized. The conclusion indicates the need for further research to standardize and expand the use of radiomic methods in clinical practice to improve the diagnosis accuracy and prognostic assessment of patients with CKD.