Abstract:529 Background: Radical cystectomy is recommended for patients with BCG-unresponsive (BU) non-muscle invasive bladder cancer (NMIBC) due to high risk of progression. Improved methods for assessing these risks would greatly facilitate the potential for bladder preservation. Here, we tested the capacity of the urinary comprehensive genomic profile (uCGP) to predict event-free survival (EFS) in patients with BU-NMIBC treated with atezolizumab in the single arm phase 2 trial SWOG S1605. Methods: Urine was collected from patients with BU NMIBC (CIS, Ta, T1) treated with at least one dose of intravenous atezolizumab at baseline and before the 5 th cycle of therapy (3 months). The uCGP was assessed using UroAmp (Convergent Genomics). Risk scores for recurrence at baseline and at 3 months were calculated by a prespecified machine learning algorithm incorporating alterations in 60 genes and low pass whole genome sequencing, and categorized as high versus low. Molecular response was classified based on change in uCGP between the two time points. Risk scores and molecular response were submitted to SWOG for clinical correlation with EFS using a Cox model, adjusting for CIS status. Time to event was calculated from date of study entry (baseline risk) or from the 2 nd collection time (3-month risk) to first high grade (HG) recurrence or persistent CIS at 3 months. Death unrelated to bladder cancer and patients last known to be alive without HG recurrence were censored at date of last visit. Results: Samples were provided at baseline in 89 patients and before the 5 th cycle in 77; 68 had both samples available for paired analysis. The risk score at baseline was classified as high in 69% of samples (73% for CIS ±Ta/T1 and 62% for Ta/T1). At 12 and 18 months the EFS probabilities were 26% and 23% for high-risk and 67% and 51% for low-risk patients, respectively, with a HR of 2.82 (95% CI: 1.58, 5.03; p<0.001). The risk score at the 3-month timepoint was classified as high in 81% of samples, and the EFS probabilities at 12 and 18 months after urine collection were 26% and 22% for high-risk and 80% and 72% for low-risk patients, respectively, with a HR of 3.39 (95% CI: 1.41, 8.13; p<0.006). Molecular response to treatment was classified as complete (CR) in 8%, partial (PR) in 14%, stable (SD) in 25% and progression (PD) in 46%, with 7% having no detectable genomic abnormalities at both time points. Clinical recurrence was observed by 18 months in 0/6 patients with CR, 7/9 (78%) with PR, 11/14 (79%) with SD and 25/33 (76%) with PD by genomic profile. Conclusions: This study suggests that uCGP at baseline and after 4 cycles of treatment can identify genomic patterns associated with an increased risk of HG persistence, recurrence or progression in BU NMIBC treated with immune checkpoint inhibition. Future studies will determine if this can be used to guide early treatment intensification.

A Novel Gene Expression Scoring System Predicts Recurrence in Non‐Muscle‐Invasive Bladder Cancer Patients

Comprehensive Analysis of Multi-Omics Data of Recurrent Gliomas Identifies a Recurrence-Related Signature As a Novel Prognostic Marker.

Gene-Expression Signature Predicts Postoperative Recurrence in Stage I Non-Small Cell Lung Cancer Patients

Identification of grade-related genes and construction of a robust genomic-clinicopathologic nomogram for predicting recurrence of bladder cancer

Transcriptome analysis of low-risk and high-risk non-muscular invasive bladder cancer patients to reveal disease progression related genes.

Feasibility Study on Predicting Recurrence Risk of Bladder Cancer Based on Radiomics Features of Multiphase CT Images

Prognostic factors of non-muscle invasive bladder cancer: a study based on next-generation sequencing

Eight-gene Risk Score Predicts Progression and Prognosis in Bladder Cancer

Long noncoding RNAs to predict postoperative recurrence in bladder cancer and to develop a new molecular classification system

Development and validation of a recurrence risk assessment model for high-grade bladder cancer based on TCGA and GEO

Combination of a Novel Gene Expression Signature with a Clinical Nomogram Improves the Prediction of Survival in High-Risk Bladder Cancer

Development and external validation of a novel nomogram model for predicting postoperative recurrence-free survival in non-muscle-invasive bladder cancer

Construction and Validation of a Novel Eight-Gene Risk Score&nbsp;to Predict the Malignant Progression and Prognoses in Bladder Cancer Patients

A novel gene signature combination improves the prediction of overall survival in urinary bladder cancer

Identification of Signature Genes Associated With Invasiveness and the Construction of a Prognostic Model That Predicts the Overall Survival of Bladder Cancer

Abstract 3827: Predictions for Recurrence of Non-Muscle Invasive Bladder Cancers by Circulating Mirnas

Correlation of urinary comprehensive genomic profile with risk of recurrence of BCG-unresponsive non-muscle invasive bladder cancer treated with atezolizumab in SWOG S1605.

Deep Learning-based Recurrence Prediction in Patients with Non-muscle-invasive Bladder Cancer

Construction and validation of a novel eight-gene risk score to predict the progression and prognosis of bladder cancer

Multimodal recurrence scoring system for prediction of clear cell renal cell carcinoma outcome: a discovery and validation study

A predictive nomogram for individualized recurrence stratification of bladder cancer using multiparametric MRI and clinical risk factors