A cost-effective diagnostic approach of urothelial carcinomas in 2 mL full voided urine based on novel panels of dual methylated DNA markers

Background: Currently, universal methylated biomarkers for urothelial carcinomas are lacking, and nearly all the diagnostic panels for bladder cancer with accuracy over 90% are multiplex (> 3 markers), and based on large volume urine (> 50 mL), which are not cost-effective for clinical application. The aim of this study was to identify universal marker for urothelial carcinomas based on 2 mL full voided urine. Method: In this study, we discovered the top 25 differential methylation regions with a sliding window method using TCGA cohort, and 8 were validated in 30 healthy blood and 20 normal urine samples by Sanger sequencing, three new regions chr10:101140373-101140735 (hg38), GRASP and AL021918.2 with specificities ≥90% were further validated in 103 tissues (47 bladder cancer and 28 paired carcinoma and normal adjacent tissues of renal pelvis (n=14) and ureter (n=14)), then chr10:101140373-101140735 and AL021918.2 were tested in 2mL urine from 477 participants (199 bladder cancer, 39 renal pelvis and 26 ureter carcinomas, and 213 negative samples including other urological carcinomas and benign diseases), two widely reported bladder cancer biomarkers, TWIST1 and VIM , were also detected in tissue and urine samples for comparison, the technology platform was methylation-specific PCR. Results: : The AUC values of AL021918.2 was highest both in tissue and urine samples in classfying urothelial carcinomas and adjacent normal/negative samples. Overally, In urine, the sensitivities and specificities of AL021918.2 for urothelial carcinomas were 87.12% and 93.90%, specifically, the sensitivities for low-grade bladder cancer, Ta stage bladder cancer, low-grade upper urinary tract urothelial carcinomas were 75.56%, 83.33% and 100%, respectively. When AL021918.2 and VIM were combined, the sensitivity for urothelial carcinomas could reach 93.94%, and the specificity was 92.02%, the sensitivities of the dual-targets panel for Ta stage and low-grade urothelial carcinomas were both higher than 90%. Conclusions: : We found a novel and general urothelial carcinoma biomarker AL021918.2 , outperfoming the existing two bladder cancer markers. The combination of AL021918.2 and VIM had accuracy over 90%, in addition, only 2mL full voided urine was used, greatly improving the simplicity, which had important clinical implications in future applications.