Abstract Objective Only 2.6 - 28% of the patients that present with haematuria will be diagnosed with bladder cancer. Unnecessary cystoscopies and upper tract imaging lead to high diagnostic costs and patient burden. In order to better select patients for diagnostic cystoscopy and thereby reduce cost and over-testing, we aimed to develop a diagnostic urine assay by combining mutation status of FGFR3, TERT and PIK3CA with methylation markers OTX1 and ONECUT2. Material and Methods A total of 258 patients undergoing cystoscopy for haematuria were included, 175 were diagnosed with bladder cancer and 83 had a non-malignant source of the haematuria. Voided urine samples were collected prior to cystoscopy. DNA was extracted and analyzed for mutations in FGFR3, TERT and PIK3CA. Methylation assays were done using a single-nucleotide probe extension assay with multiple probes for OTX1 (4 probes) and ONECUT2 (5 probes). Methylation percentages were calculated and cut-off values (sensitivity ≥ 85%) were determined for dichotomization. All results were combined and the optimal model was selected through logistic regression analysis. The Bootstrap procedure was used for internal validation. Results Combining the results of the dichotomized methylation markers OTX1 (probe 2), ONECUT2 (probe 1+4) with the mutation markers FGFR3, TERT, PIK3CA and the clinical parameters Age and Gender led to a prediction model with an AUC of 0.95 (95% CI 0.92-0.98) and an optimism-corrected AUC of 0.94. To further improve the model, analyses were repeated for continuous methylation marker variables. Now, adding ONECUT2 probe 2 enhanced the predictive capacity of the model even further (AUC 0.97, 95% CI 0.94 - 0.99 and an optimism-corrected AUC of 0.96). Overall material cost was estimated to be 20 euro per patient. A mutation-analysis first strategy would lead to a potentially larger reduction in cost. This strategy identified 120 patients with bladder cancer based on mutation assays alone. For the remaining 125 mutation-naive patients the prediction model containing the previously selected continuous methylation markers combined with Age and Gender was used. The overall combined strategy sensitivity was 92,1% and specificity was 89% (AUC of 0.96). Conclusion Analyzing haematuria patients for their risk of bladder cancer may lead to a reduction of unnecessary diagnostic cystoscopies and upper urinary tract imaging. This will present a major cost reduction in the urological practice. Combining continuous marker variables for OTX1 and three probes for ONECUT2 with mutation markers FGFR3, TERT, PIK3CA and Gender and Age results in an accurate prediction model with an AUC of 0.97 and an optimism-corrected AUC of 0.96. A mutation-analysis first strategy seems most practical and feasible (Sens 92%, Spec 89%). The assay must be further improved to reach a more satisfactory overall sensitivity (>95%), for instance by adding RAS mutation analysis. Citation Format: Kim E.M. van Kessel, Willemien Beukers, Irene Lurkin, Kirstin A. van der Keur, Lars Dyrskjot, Ulrika Segersten, Torben F. Ørntoft, Nuria Malats, Per-Uno Malmström, Francisco X. Real, Chris H. Bangma, Ellen C. Zwarthoff. A urine based assay to select patients for initial cystoscopy. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 549. doi:10.1158/1538-7445.AM2015-549
Read full abstract