IBCN 2017: Biomarkers

Lisbon, Portugal (UroToday.com) Lambertus A.L.M. Kiemeney from Radboud university medical center, Radboud Institute for Health Sciences, Nijmegen, The Netherlands discussed ‘The role of germline genetic variants in the prognosis of non-muscle invasive bladder cancer: a meta-GWAS’. The study aimed to identify common germline genetic determinants of NMIBC recurrence and progression by performing the first meta-analysis of genome-wide association studies (meta-GWAS). Three cohorts were included: the Nijmegen Bladder Cancer Study (NBCS, Nijmegen, The Netherlands), the Toronto Biobank (TB, Toronto, Canada), and the Bladder Cancer Prognosis Program (BCPP, Birmingham, UK). They genotyped common single nucleotide variants (SNVs) using GWAS arrays and performed imputation using the 1000 Genomes phase1 v3 and GoNL4 reference panels 1,451, 777, and 684 samples from the NBCS, TB and BCPP, respectively. Three SNVs showed genome-wide significant association with RFS (p<5x10-8), and one with PFS (smallest p-value 2.6x10-8). Bio-informatic follow-up analyses for prioritization of top findings are currently ongoing, including in silico annotation, eQTL analysis using data from The Cancer Genome Atlas (TCGA), and study of SNV function with regard to e.g. transcription binding sites and chromatin interactions. Prioritized results will be tested for association in additional replication cohorts and these results will increase insight into the mechanisms of prognosis in NMIBC and may point the way to biomarkers that add to the currently used prediction models in the clinic.

Kate E. Williamson from the Centre for Cancer Research and Cell Biology, Queen's University Belfast, Northern Ireland ‘Biomarker classifiers for detection of bladder cancer in patients with hematuria are confounded by smoking and increased age’. They calculated a classification performance profile for each of 156 hematuria patients recruited to a case control study using a repeated random sub-sampling cross-validation of 1000 iterations. Each iteration generated a unique random forest UC diagnostic classifier using the biomarker data from a randomly sampled training dataset (n = 109 patients). This classifier was then applied to a test dataset representing the remaining patients (n = 47 patients). Each patient was assigned as correctly, incorrectly or inconsistently classified on the basis of the frequency of their assignations as control or UC across the 1000 test datasets. Out of 46 controls, 53 UCs had high probabilities for correct classification; 23 controls and 24 UCs had high probabilities for incorrect classification; seven controls and three UCs were inconsistently classified. The nonsmoker subpopulation achieved AUC = 0.95 (sd = 0.1). Twenty-two of the 23 UCs with a high probability for incorrect classification were either smokers or aged > 65 years. The authors concluded non-smokers, aged ≤65 years with microhematuria could be accurately identified using bespoke UC diagnostic classifiers. Further validation in larger numbers of patients are needed to hopefully use this classifier as a means to reduce the cystoscopy burden allowing those at greatest risk to be stratified for urgent cystoscopy.

Speaker(s): Lambertus A.L.M. Kiemeney, Radboud Institute for Health Sciences, Nijmegen; Kate E. Williamson from the Centre for Cancer Research and Cell Biology, Queen's University Belfast, Northern Ireland

Written by: Stephen B. Williams, M.D., Associate Professor, Division of Urology, The University of Texas Medical Branch, Galveston, TX. and Ashish M. Kamat, M.D. Professor, Department of Urology, Division of Surgery, The University of Texas MD Anderson Cancer Center, Houston, TX., at the International Bladder Cancer Network - October 21, 2017- Lisbon, Portugal

Login