Comprehensive analysis of cell-free tumour DNA in plasma and urine tracks disease burden and reveals clonal evolution in muscle invasive bladder cancer


Session type:

Christopher Smith1,Kristan van der Vos2,Keval Patel1,Charles Massie1,Francesco Marass1,James Morris1,Dana Tsui1,Florent Mouliere1,Vincent Gnanapragasam1,Tim Forshew3,Bas van Rhijn2,Nitzan Rosenfeld1,Michiel van der Heijden2
1University of Cambridge,2Netherlands Cancer Institute,3University College London



Muscle Invasive Bladder Cancer (MIBC) is an aggressive disease with a significant risk of metastatic progression. Neo-adjuvant chemotherapy after radical cystectomy can improve overall survival but benefit is limited to ~50% of eligible patients. Presently there is no accurate and prompt predictor of treatment response. Cell-free tumour DNA (cftDNA) present in plasma (PLS), urine-supernatant (USN) or -cell-pellet (UCP) may represent a non-invasive means to rapidly predict treatment response and track disease progression.


We used a combination of Tagged Amplicon Sequencing (TAm-Seq) and shallow Whole Genome Sequencing (sWGS) to detect and track somatic point-mutations and copy-number abnormalities (SCNA) in 345 longitudinal DNA samples. These were extracted from matched initial endoscopic tumour resection (TUR), PLS, USN or UCP of 20 MIBC patients. Presence and levels of cftDNA were correlated with treatment response and disease course.


TAm-Seq identified 37 mutations (including TP53, KRAS) in TUR, PLS, USN or UCP in 15/20 patients. 81% of these were detected in TUR whilst the remaining 19% were exclusive to peripheral fluids. In all patient TUR, including those where no point-mutations were identified, sWGS revealed genome wide and focal SCNAs, and these were represented to different degrees in matched PLS (27.3%), USN (58.9%) and UCP (58.9%). Indeed, private somatic events were observed in all fluid compartments, suggesting that all are valuable targets for testing. Whilst presence and levels of cftDNA at the first time-point did not significantly correlate with treatment response, longitudinal analysis revealed a general correlation between cftDNA dynamics and ultimate disease course, as well as evidence of clonal evolution.


We demonstrate combined urine and plasma cftDNA analysis as a means of tracking disease course in MIBC. Point-mutation and SCNA levels dynamically changed over time and revealed evidence of clonal evolution. Larger scale analyses are required to confirm predictive utility of cftDNA for treatment response.