Integrative analysis of urine cell-free DNA for the detection of residual disease in localized bladder cancer patients.

Abstract

559 Background: We previously developed a liquid biopsy assay to measure urine tumor DNA (utDNA) levels based on detection of single nucleotide variants (SNVs) in urine cell-free DNA (cfDNA). Nonsilent SNV detection in urine from muscle-invasive bladder cancer (MIBC) patients prior to radical cystectomy (RC) was associated with pathologic residual disease and worse progression-free survival (Chauhan et al., PLOS Medicine, 2021). Given the multiple types of genomic alterations present in bladder cancer (BC), here we developed a multi-omics approach for estimating utDNA levels without restricting our approach to SNVs. We performed ultra-low pass whole genome sequencing (ULP-WGS) based copy number analysis and urine Cancer Personalized Profiling by deep Sequencing (uCAPP-Seq) of urine cell-free DNA to predict pathologic complete response (pCR) in localized BC patients. Methods: We acquired urine preoperatively from 65 BC patients (69% muscle-invasive) on the day of standard-of-care RC, and after neoadjuvant chemotherapy in 48% of patients. We performed ULP-WGS of urine cfDNA from all 65 BC patients and 11 healthy adults. utDNA levels based on genome-wide copy number alterations (CNAs) in urine cfDNA was estimated using ichorCNA. In order to derive a SNV-based utDNA level as well, uCAPP-Seq was applied to urine cfDNA samples derived from 42 patients using a 145 kb panel of 49 consensus driver genes commonly mutated in MIBC. Results: In our cohort of 65 BC patients, 55% of patients achieved pCR ( n = 36) while 45% had residual disease detected in their surgical sample (no pCR; n = 29). Comparing ULP-WGS-derived utDNA levels between the groups, patients with no pCR had significantly higher CNA-derived tumor fractions in urine compared to patients with pCR (median 8.9% vs 1.8%, p = 0.01) and healthy adults ( n = 11) (median 8.9% vs 0%, p = 0.006). Further analysis with uCAPP-Seq in 42 patients revealed that nonsilent SNV-based utDNA detection correlated significantly with the absence of pCR ( p < 0.001) with a sensitivity of 81% and specificity of 81%. Moreover, utDNA-positive patients exhibited significantly worse progression-free survival compared to utDNA-negative patients (HR = 7.4; 95% CI: 1.4–38.9; p = 0.02). Conclusions: Bladder cancer patients who did not attain a pCR at the time of RC had greater genome-wide copy number alterations and nonsilent single nucleotide variants in their urine cfDNA compared to patients with pCR. These results suggest that integrative multi-omics of urine derived from MIBC patients has potential real-world clinical impact for bladder-sparing approaches in select patients.