MP16-04: Minimal residual disease detection supports high-grade bladder cancer risk-stratification during recommended repeat transurethral resection
Friday, May 3, 2024 1:00 PM to 3:00 PM · 2 hr. (US/Central)
221B
Abstract
Information
Full Abstract and Figures
Author Block
Eugene Shkolyar*, Kathleen Mach, Palo Alto, CA, Daniel S Fischer, Barbara Hamlington, Vincent Bicocca, Trevor Levin, South San Francisco, CA, Joseph C Liao, Palo Alto, CA
Introduction
High early recurrence and upstaging rates for patients with high-risk, localized urothelial carcinoma of the bladder (UCB) implicate the presence of minimal residual disease (MRD) after transurethral resection of bladder tumor (TURBT). A non-invasive urine test providing prognostic MRD detection could augment clinical decision-making. We evaluated the performance of urinary comprehensive genomic profiling (uCGP) to detect MRD in a cohort of patients undergoing repeat TURBT for high-grade (HG) UCB.
Methods
In this retrospective study uCGP was performed on 22 urine samples from 21 patients undergoing re-TURBT for pathologically confirmed HG UCB. Urine was collected immediately prior to surgery. uCGP was performed using the UroAmp assay (Convergent Genomics), which quantifies somatic tumor mutations from urine-derived DNA. Previously validated algorithms for MRD recurrence risk and grade prediction were tested. The primary analysis was disease classification and grade predictions compared to restaging pathology.
Results
Stage distribution at re-TURBT was 27% pTa, 32% pT1, 23% pT2, and 18% pT0. MRD-detection returned definitive results on all samples and correctly classified 89% of UCB-positive specimens (Figure 1A). The HG prediction algorithm performed with 100% PPV (95% CI 71-100). Among T0 samples, MRD-detection had a specificity of 50%, with 2 false-positives. Long-term follow-up of these samples revealed that one patient had an immediate recurrence (cystoscopy at 3 month) with HG T2 disease. The second patient had a history of carcinoma in situ (CIS) but no documented recurrence; their MRD mutation profile included TERT, ARID1A, TP53, and RB1, all mutations enriched in CIS compared to HG papillary UCB, strongly suggesting the presence of occult CIS. The lone patient with longitudinal collections demonstrated a large MRD mutation burden at the time of repeat resection that was consistent with HG T1 restaging. MRD detection at a follow-up resection (“re-re-TURBT”) showed no evidence of disease and was consistent with benign restaging (Figure 1B).
Conclusions
These results support uCGP-guided risk-stratification of HG UCB patients undergoing re-TURBT. MRD detection provides prognostic information for multifocal and visually occult disease.
Source Of Funding
This work was supported by the Department of Veterans Affairs BLR&D Merit Review I01BX004962
Author Block
Eugene Shkolyar*, Kathleen Mach, Palo Alto, CA, Daniel S Fischer, Barbara Hamlington, Vincent Bicocca, Trevor Levin, South San Francisco, CA, Joseph C Liao, Palo Alto, CA
Introduction
High early recurrence and upstaging rates for patients with high-risk, localized urothelial carcinoma of the bladder (UCB) implicate the presence of minimal residual disease (MRD) after transurethral resection of bladder tumor (TURBT). A non-invasive urine test providing prognostic MRD detection could augment clinical decision-making. We evaluated the performance of urinary comprehensive genomic profiling (uCGP) to detect MRD in a cohort of patients undergoing repeat TURBT for high-grade (HG) UCB.
Methods
In this retrospective study uCGP was performed on 22 urine samples from 21 patients undergoing re-TURBT for pathologically confirmed HG UCB. Urine was collected immediately prior to surgery. uCGP was performed using the UroAmp assay (Convergent Genomics), which quantifies somatic tumor mutations from urine-derived DNA. Previously validated algorithms for MRD recurrence risk and grade prediction were tested. The primary analysis was disease classification and grade predictions compared to restaging pathology.
Results
Stage distribution at re-TURBT was 27% pTa, 32% pT1, 23% pT2, and 18% pT0. MRD-detection returned definitive results on all samples and correctly classified 89% of UCB-positive specimens (Figure 1A). The HG prediction algorithm performed with 100% PPV (95% CI 71-100). Among T0 samples, MRD-detection had a specificity of 50%, with 2 false-positives. Long-term follow-up of these samples revealed that one patient had an immediate recurrence (cystoscopy at 3 month) with HG T2 disease. The second patient had a history of carcinoma in situ (CIS) but no documented recurrence; their MRD mutation profile included TERT, ARID1A, TP53, and RB1, all mutations enriched in CIS compared to HG papillary UCB, strongly suggesting the presence of occult CIS. The lone patient with longitudinal collections demonstrated a large MRD mutation burden at the time of repeat resection that was consistent with HG T1 restaging. MRD detection at a follow-up resection (“re-re-TURBT”) showed no evidence of disease and was consistent with benign restaging (Figure 1B).
Conclusions
These results support uCGP-guided risk-stratification of HG UCB patients undergoing re-TURBT. MRD detection provides prognostic information for multifocal and visually occult disease.
Source Of Funding
This work was supported by the Department of Veterans Affairs BLR&D Merit Review I01BX004962