A multimodal approach for the diagnosis of bladder tumours

Abstract

Urothelial carcinoma (UC) is the most common type of bladder cancer, and its treatment depends from both tumour invasiveness (stage) and aggressiveness (grade). The gold standard for detecting UC is white-light cystoscopy, followed by tissue biopsy and histopathological examination; however, such process is invasive, time-consuming, operatordependent and prone to sampling errors. In this framework, optical spectroscopy techniques could be a promising solution for fast and label-free diagnosis of bladder tissues and for early detection of UC. Thus, we combined autofluorescence, diffuse reflectance and Raman spectroscopy in a compact and transportable setup based on an optical fibrebundle probe. This experimental setup was used for studying fresh biopsies of urothelial tumour (140 samples) and healthy bladder (50 samples) collected from 90 patients undergoing Transurethral Resection of Bladder Tumours (TURBT). The aim of this study was to develop an automated classification of the examined tissues based on the intrinsic spectral information provided by all three techniques. We found that healthy and diseased tissues showed significant spectral differences for each technique, resulting in high accuracy (up to 90%) from a Linear Discriminant Analysis (LDA) routine. In particular, fluorescence spectroscopy – excited either with blue or UV light – provided very good results in detecting UC. However, tumour grading and staging proved to be more challenging tasks, for which no single spectroscopic technique could provide sufficient sensitivity and specificity. Therefore, we found that a multimodal approach can improve significantly the diagnosis of UC stages and grades.