Contemporary capabilities of artificial intelligence technologies in the diagnosis of kidney and urinary tract tumors: a review
Digital Diagnostics · 2026-08-28 · 7 (3) · 351-365
DOI: 10.17816/dd703079 ↗Abstract
The rising incidence of renal and urinary tract tumors highlights the critical need for more accurate, noninvasive diagnostic methods and unbiased assessment of tumor aggressiveness. Conventional imaging techniques and histopathological verification often fail to reliably predict biological tumor behavior, leading to both overtreatment and delayed detection of high-risk disease. Consequently, research focus has increasingly shifted toward the diagnostic capabilities of artificial intelligence. This review synthesizes findings from recent studies evaluating artificial intelligence algorithms in the diagnosis of renal and urinary tract tumors. Cumulative evidence indicates that deep learning models trained on computed tomography, magnetic resonance imaging, ultrasonography, endoscopy, and histology images achieve high accuracy in automated tumor segmentation, differentiation between benign and malignant neoplasms, and identification of histological subtypes and tumor grades. We also examine studies applying artificial intelligence to predict recurrence risk, overall survival, therapeutic response, and postoperative functional preservation. Deep learning-derived radiomic and radiogenomic signatures demonstrate substantial prognostic value and likely correlate with underlying biological features. In conclusion, artificial intelligence expands noninvasive phenotypic characterization of urinary tract tumors; however, translation of these frameworks into routine clinical practice requires further validation.
EUVIMED