Abstract 965: Urine cell-free DNA multi-omics combining copy number analysis with fragmentomics to detect minimal residual disease in bladder cancer patients

Abstract Background: Standard-of-care for patients with localized muscle-invasive bladder cancer (BC) is neoadjuvant chemotherapy followed by radical cystectomy (RC). Bladder-sparing treatments are limited by our current inability to sensitively detect minimal residual disease (MRD). Here, we analyzed urine, a noninvasive and proximal biofluid, from patients with bladder cancer and utilized an approach that incorporates copy number-derived tumor fraction as well as fragmentomics to sensitively detect MRD and predict pathologic complete response (pCR). Methods: We acquired urine preoperatively from 51 BC patients (80% muscle-invasive) on the day of standard-of-care RC, and after neoadjuvant chemotherapy in 57% of patients. We performed whole genome sequencing (WGS) of urine cfDNA (median depth: 10X) from all 51 BC patients and 13 healthy adults. Tumor fraction level based on genome-wide copy number alterations was estimated using ichorCNA. We inferred fragment size and genomic coverage of urine cfDNA using Picard Tools. Then, we computed the short-to-long (S/L) fragment ratio for genomic bins in each sample. The S/L ratio per bin was derived by partitioning the genome into 100 kb bins and evaluating the ratio of GC-corrected short fragments (50-150 bp) to long fragments (151-250 bp) within each bin. A normalized genome-wide S/L score was calculated by averaging across all bins. Results: In our cohort of 51 BC patients, 45% of patients achieved pCR (n = 23) while 55% had residual disease detected in their surgical sample (no pCR; n = 28). Patients with no pCR had significantly higher copy number-derived tumor fractions in urine compared to patients with pCR (median 10.6% vs 1.9%, p = 0.0002) and healthy adults (n = 12) (median 10.6% vs 1.8%, p = 0.003). Tumor fraction achieved an AUC of 0.89 for classification of No pCR from pCR with sensitivity of 82% and specificity of 92%. Presumed cancer-free patients (healthy and pCR) had significantly longer urine cfDNA median fragment sizes than patients with no pCR (median 177 bp vs 156 bp, p = 0.002). Presumed cancer-free patients (healthy and pCR) also had significantly lower S/L ratio scores than no pCR patients (mean .98 vs 1.5, p = 0.007). The top 100 bins with maximally differential S/L ratio between these two patient groups were enriched for genes involved in KEGG pathways implicated in cancer and immune signaling (autophagy, transcriptional misregulation in cancer, Th1 differentiation) indicating biological relevance. Conclusion: Copy number alteration-derived tumor fraction inference combined with fragmentomic analysis of urine cell-free DNA can enable sensitive detection of MRD to predict pCR and reveal potentially onco-relevant pathways. Citation Format: Arpit Panda, Irfan Alahi, Pradeep S. Chauhan, John Sheng, Nathan Colon, Cayce Nawaf, Alexander Shiang, Peter K. Harris, Faridi Qaium, Eric H. Kim, Melissa A. Reimers, Woodson Smelser, Zachary L. Smith, Aadel A. Chaudhuri. Urine cell-free DNA multi-omics combining copy number analysis with fragmentomics to detect minimal residual disease in bladder cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 965.