We studied the feasibility of Raman spectroscopy for the diagnosis of bladder cancer in vivo. Since the invasion stage is crucial for the treatment choice, a high-volume based Raman probe was used to investigate the potential of determining the invasiveness of bladder cancer. High quality spectra were obtained from suspicious and nonsuspicious bladder locations during the procedure of transurethral resection of bladder tumors (TURBT) with collection times of 1-5 s. Multivariate analysis was used to generate the classification models. The algorithm was able to distinguish bladder cancer from normal bladder locations with a sensitivity of 85% and a specificity of 79%. The Raman spectra of bladder cancer stages showed a gradual increase in the intensity of specific amino acid peaks and, most likely, an increase in the intensity of DNA peaks.