Urinary DNA-methylation and protein biomarkers identify urothelial carcinoma among other genitourinary diseases and cancer-free individuals

Background: For more than 80 years, cystoscopy has been the gold standard for identification of urothelial carcinoma (UCa). Because of many factors, such as pain of the patients during this procedure or the costs involved, non-invasive detection of UCa remains a challenge. Herein, we verify our previously identified urinary biomarkers C-X-C Motif Chemokine Ligand 16 (CXCL16) and transforming growth-factor beta induced protein (TGFBI) on the protein level as well as the CpG sites ALOX5, TRPS1 and an intergenic region on Chromosome 16 on DNA methylation level in an independent cross-sectional study.

Methods: We collected N = 1119 urines from individuals coming to urological and gynecological check-ups, follow-up care or patients suspicious for UCa or already diagnosed for different urologic or gynecologic cancer entities. We performed methylation analysis of various CpG sites with DNA isolated from urine sediment and quantified the concentration of the protein markers CXCL16 and TGFBI in the corresponding urine supernatant using ELISA. We tested for patient-group differences with two-sided Wilcoxon rank sum tests and examined the performance with receiver operating characteristic curves. For verification, we analyzed the marker performance with previously set cutoff-values and marker combinations with established and experimental algorithms (with logical OR-conjunction, iterative threshold-based biomarker and score combining algorithm "PanelomiX").

Results: Evaluation confirmed that our previously identified protein and DNA methylation biomarkers can distinguish UCa from frequent urological and gynecological cancers. CXCL16 and TGFBI discriminated UCa cases with a sensitivity of 31% and 56% and a specificity of 94% and 85%, respectively. Combining methylation markers resulted in UCa detection in men with a sensitivity of 54% and a specificity of 94%. Extending analysis by combining all methylation and protein markers (up to five markers in total) yielded a convincingly high specificity of 97% at a sensitivity of 72% for the identification of UCa patients within a heterogeneous collective of cancer-free individuals and patients suffering from urological or gynecological cancers.

Conclusion: Combining various biomarkers at protein and DNA level demonstrates a new option of non-invasive UCa diagnosis in urine, and thus might help to reduce the number of unnecessary cystoscopies, especially in patients without a history of UCa.