Urinary Comprehensive Genomic Profiling Predicts Urothelial Carcinoma Recurrence and Risk of Disease Progression: A Multi-Institutional Longitudinal Surveillance Cohort
Andrew E. Amini, BS1, Hiten D. Patel, MD2, Vincent M. Caruso, MS3, Jason J. Lee, MD1, Goran Rac, MD4, Shalin C. Desai, MD4, Kevin G. Philips, PhD3, Shulin Wu, MD1, Chin-Lee Wu, MD, PhD1, Peter S. Lentz, MS3, Mahdi Goudarzi, PhD3, Daniel S. Fischer, MS3, Brian C. Mazzarella, MD3, Vincent T. Bicocca, PhD3, Trevor G. Levin, PhD3, Gopal N. Gupta, MD4, Adam S. Feldman, MD, MPH1, Keyan Salari, MD, PhD1.
1Massachusetts General Hospital, Boston, MA, USA, 2Northwestern University, Chicago, IL, USA, 3Convergent Genomics, Inc., San Francisco, CA, USA, 4Loyola University Medical Center, Chicago, IL, USA.
BACKGROUND: Risk stratification of patients with urothelial carcinoma (UC) remains a clinical challenge, with high rates of recurrence and disease progression. Urinary comprehensive genomic profiling (uCGP) has potential to aid in treatment selection.
METHODS: This is a blinded, multicenter case-control study of banked urine specimens collected from UC surveillance patients with long-term followup. Using UroAmp™, uCGP was performed on 120 subjects with a history of UC and negative surveillance cystoscopy at time of collection. Subjects were predicted high- or low-risk of recurrence using a machine learning algorithm. Kaplan-Meier curves were computed to estimate recurrence-free survival in recurrence risk groups. Univariable and multivariable Cox proportional hazards models estimated the association between risk predictions and recurrence-free survival (RFS).
RESULTS: High recurrence risk was predicted in 37% (45/120) of subjects; of these, 19 had a clinical recurrence, 12 of which were high-grade. Of the 75 predicted low-risk patients, 8 (11%) experienced recurrences, only one of these occurred within the first 12 months of surveillance. RFS was significantly worse in high-risk versus low-risk patients (hazard ratio 4.4, p=0.0017). In a multivariable Cox proportional hazards model, UroAmp risk category remained significantly associated with recurrence after adjusting for other clinical and pathologic features (Fig 1).
CONCLUSIONS: uCGP can predict future recurrence of high-risk UC with substantial lead time. If validated with further studies, uCGP may provide a powerful way to risk-stratify patients and enable personalized strategies of UC management based on genomic risk.
Figure 1. uCGP Predicted Recurrence Risk. UC surveillance patients with negative cystoscopy and long-term follow-up with outcomes were analyzed for recurrence risk (n = 120, validation cohort). (A) Kaplan-Meier curves for recurrence-free survival by UroAmp predicted risk. Significance: P = 0.00026 (Log-rank test). (B) Univariable and (C) multivariable Cox proportional-hazard regression analysis of UroAmp recurrence risk groups and clinical risk factors. For stage, T1+ indicates the grouping of patients with T1, T2, and T3 disease. 
Back to 2023 Abstracts