Introduction:

Cell-free DNA (cfDNA) has emerged as a promising liquid biopsy biomarker for various malignancies. In particular, voided urine cfDNA has shown potential as a biomarker within urothelial bladder cancer. With a specificity of 99.3% and sensitivity of 86.7% for detecting bladder cancer, urine cfDNA has proven superior to plasma cfDNA, making it the preferred biomarker in this context.

Despite these advancements, studies of cfDNA within upper tract urothelial carcinoma (UTUC) are limited. The existing literature on UTUC cfDNA predominantly focuses on plasma-derived cfDNA, leaving voided urine cfDNA relatively unexplored. While it seems logical to extrapolate the benefits of urine cfDNA from bladder cancer to UTUC, this area has yet to be thoroughly investigated.

Therefore, we utilized low-pass whole genome sequencing (LP-WGS) to detect genomic alterations in cfDNA from both urine and plasma samples of UTUC patients, with the goal of establishing urine cfDNA as a reliable biomarker for UTUC.

Methods:

We prospectively enrolled four patients with known or suspected UTUC undergoing radical nephroureterectomy. We collected peripheral blood and voided urine samples from all patients preoperatively.  CfDNA was isolated from urine and plasma using standard protocols. Genomic DNA was extracted from radical nephroureterectomy specimens and established as tissue DNA.

Next-generation sequencing using LP-WGS was performed on all samples (urine, plasma, tissue) to a minimum depth of 0.1x. Sequenced reads were matched against the human reference genome (hg19). Via a custom-developed clinically validated bioinformatics pipeline, copy number alterations (CNAs) were identified – which are genomic regions of provided samples that deviated from the reference genome. CNAs of urine and plasma were matched against tissue samples, and differences in CNAs at the chromosomal and sub-chromosomal level were identified amongst all three samples. Statistical z-scores for CNAs were utilized to determine the magnitude of detected CNAs in urine and plasma cfDNA.

Results:

Four patients underwent radical nephroureterectomy; final pathology was pT1N0 (n=2) and pT2Nx (n=2). All patients had preoperative voided urine cfDNA, preoperative plasma cfDNA and tumor tissue DNA available for analysis.

At the chromosomal level, we observed that CNAs in urine cfDNA closely mirrored that of tissue. In contrast, there was low concordance between CNAs in plasma cfDNA and tissue. This discrepancy is illustrated in Figure 1. Pearson correlation coefficients of CNAs between urine cfDNA and tissue ranged from 0.873 to 0.997, indicating strong concordance. Conversely, the correlation coefficients for CNAs between plasma cfDNA and tissue were much lower, ranging from -0.242 to 0.213.

On a sub-chromosomal level, a similar pattern of concordance was observed between urine cfDNA and tissue, with comparable aberrations detected in plasma cfDNA and tissue. Amplifications, deletions, and arm level gains/losses identified in tissue specimens were duplicated to a high level in matched urine cfDNA.

Conclusion:

In this pilot study, we demonstrated that for upper tract urothelial carcinoma, copy number alterations (CNAs) of urine cfDNA closely mirrored that of tissue, on both the chromosomal and sub-chromosomal level. Conversely, this relationship between plasma cfDNA and tissue was not seen. Our findings also serve as proof of concept that UTUC can be detected in urine supernatant using LP-WGS. Taken together, these data suggest that urine cfDNA has the potential to serve as an accurate biomarker for UTUC, offering a less invasive yet equally informative approach to UTUC diagnosis and surveillance.

Funding: N/A