Abstract
Accurate and noninvasive diagnosis is critical in the management of urothelial carcinoma (UC). This study develops and prospectively validates a multitarget urine RNA (mt-uRNA) test for noninvasive UC detection. The RNA panel, built using quantitative RT-PCR and support vector machine algorithms, achieved an area under the curve of 0.956 in the upgraded training set of 688 samples. In the prospective validation phase, urine samples from 752 patients in three Chinese medical centers were analyzed, and the test achieved an overall accuracy of 93.4%, with a sensitivity of 92.3% and specificity of 94.1%. This robustness was maintained across subgroups, demonstrating high sensitivity for carcinoma in situ, low-grade, and upper tract UC (83.3%, 89.9%, and 97.3%) and high accuracy in hematuria, residual and recurrent subpopulations (93.6%, 86.9%, and 92.7%). These findings underscore the mt-uRNA test's high diagnostic utility for noninvasive UC detection, offering a promising alternative/adjunct to endoscopy for hematuria investigations and surveillance.