Abstract 1275: Treatment response and outcome prediction guided by deep targeted sequencing of urine and plasma samples in patients with muscle-invasive bladder cancer

Abstract

Abstract Background Standard treatment in patients with localized muscle-invasive bladder cancer (MIBC) is neoadjuvant chemotherapy (NAC) followed by radical cystectomy (CX). Importantly, there is currently no clinically validated way to estimate response to NAC before CX, where a pathological assessment of tumor down-staging is used as a proxy for response. Identification and tracking of circulating tumor DNA (ctDNA) has been shown to reflect prognosis and treatment response in several cancers, including bladder cancer. Here, we aim to investigate the capability of a combined plasma, urine supernatant and urine pellet mutation analysis by deep targeted sequencing for detecting residual tumor and treatment response before CX. Methods In total, 92 patients with localized MIBC treated with NAC followed by CX were included in the study. DNA extracted from primary tumors and leukocytes were subjected to whole exome sequencing for identification of approx. 50 mutations per patient that were used to design a custom NGS panel. Urine supernatants (n=281), urine pellets (n=114) and plasma samples (n=167), collected before CX, were subjected to deep targeted sequencing employing unique molecular identifiers (UMI). Mutations were assessed using the Shearwater algorithm and samples were further assessed using Fisher’s Method and bootstrapping of random mutations. In-silico analyses identified the limit of detection to be approx. 0.1%. For 56 patients, plasma mutation data (n=288 samples) were obtained from a previously published study by our group. Results Tumor derived DNA levels in paired urine supernatants and pellets were correlated (rho=0.8), however with higher levels in pellets when adjusting for the extracted amount of DNA (p<0.001). Tumor derived DNA was less frequently observed in plasma (20% of samples) compared to urine supernatants and pellets (63% and 68% of samples, respectively) and at lower levels (p<0.001, both comparisons). Tumor derived DNA levels were associated with tumor stage for plasma samples, but demonstrated variable levels for urine supernatants and pellets. Interestingly, we observed clearance of tumor derived DNA to be significantly associated with NAC response (pathological downstaging to <T1) for urine supernatants (p<0.001) but not for urine pellets (p=0.23) and plasma (p=0.05). Conversely, clearance of tumor derived DNA in plasma was significantly associated with recurrence-free survival (p<0.001) but not for urine supernatants (p=0.06) and urine pellets (p=0.53). Conclusions Tumor derived DNA was more frequently observed in urine samples and at higher levels compared to plasma samples. Interestingly, we show that tumor derived DNA dynamics in urine supernatants might better reflect the local tumor environment and plasma-based tumor derived DNA better reflects outcome following surgery, i.e. disease dissemination. Citation Format: Emil Christensen, Iver Nordentoft, Sara K. Elbæk, Karin Birkenkamp-Demtröder, Ann Taber, Michael Knudsen, Philippe Lamy, Mads Agerbæk, Jørgen B. Jensen, Lars Dyrskjøt. Treatment response and outcome prediction guided by deep targeted sequencing of urine and plasma samples in patients with muscle-invasive bladder cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1275.