Towards the Point of Care and noninvasive classification of bladder cancer from urine sediment infrared spectroscopy. Spectral differentiation of normal, abnormal and cancer patients

Abstract

Bladder cancer (BC) is the 9th cancer cause of death and one of most cost-intensive in the world. The diagnostic tools are still not at all satisfactory. Herein we evaluated the potential of infrared spectroscopy to detect molecular changes that precede and accompany the carcinogenesis in voided urine sediment. We collected 165 samples from patients being diagnosed for BC and measured them with attenuated total reflectance Fourier transformed infrared spectroscopy (ATR FTIR). Samples were primarily divided into three groups according to cytology that indicated the presence of normal, abnormal and cancer cells. ATR FTIR spectra of sediments were analyzed with the use of partial least square discriminant analysis (PLSDA). The 1800–750 cm−1 region discriminated the three groups with selectivity and sensitivity values around 68% using cytology as a reference method. These cross-validation values (which were found significant according to a permutation test) were comparable to the sensitivity and specificity values of cytology versus the gold standard (histology). The average spectra of each class and the regression vectors of the PLS-DA indicated that an increased content of carbohydrates and nucleic acids as well as transformations of protein secondary structures were the main discriminators of healthy patients from abnormal and cancer groups. Additionally, we revised the obtained classification according to diagnosis made on histopathological assessment of bladder sections. We finally discuss the potential of the technique to be used as a Point of Care (PoC) testing tool.