Introduction: The detection of asymptomatic bladder carcinoma remains a challenge. There is no accepted non-invasive, low-cost, and reliable screening modality. Normal human urine contains over 2000 separate chemical entities, reflective of systemic physiology and metabolism. Of existing bladder cancer detection tests, the vast majority rely on a single or a few molecular markers for diagnosis. We have been developing and validating a novel approach to molecular urinalysis, using a combination of Raman spectroscopic, computational, and physicochemical analytical methods. The method is termed Raman chemometric urinalysis and it assesses the whole composition of the evaluated fluid.

Methods: Urine specimens from patients with bladder cancer, other genitourinary (GU) diseases, and healthy volunteers were analyzed using a Raman chemometric urinalysis method. Computational and statistical analysis of spectra from urine specimen was conducted. Based on obtained data, we developed a chemometric urinalysis to detect complex molecular signatures associated with bladder cancer and other GU diseases.

Results: A total of 133 patients were included in this study. Computational analysis was able to discern molecular signatures, indicative of the presence of bladder cancer in this mixed population, with 82.4% sensitivity and 79.5% specificity. Based on the methods to construct these signatures, screen could be fine-tuned for either high sensitivity or specificity.

Conclusions: The molecular composition of urine from patients with bladder cancer differs from that of urine from healthy volunteers and patients with GU disorders, as seen in the unique Raman spectral fingerprints identified in this study. This method is based on analyzing the whole spectrum of the metabolites in the urine. While further testing and validation are needed, applying Raman spectroscopy to molecular urinalysis has the potential to improve the cost and efficacy of bladder cancer screening and surveillance.