Our aim is to apply image analysis on photosensitizer fluorescence and compare the relationship between histopathology and endoscopic fluorescence imaging. The correlation between hypericin fluorescence and histopathology of diseased tissue was explored in a clinical study involving 58 fluorescence cystoscopic images from 23 patients. Based on quantification of fluorescence colorimetric parameters extracted from the image analysis, diagnostic functions were developed to pathologically classify the bladder cancer. Our preliminary results show that the differences in fluorescence intensity ratios among the three different grades of bladder cancer are statistically significant. The results also show a decrease in macroscopic fluorescence intensity that correlated with higher cancer grades. By combining both the red-to-green and red-to-blue fluorescence intensity ratios into a 2-D scatter plot and defining diagnostic linear discrimination functions on the data points, this technique is able to yield an average sensitivity and specificity of around 68.6% and 86.1%, respectively, for pathological cancer grading of the three different grades of bladder cancer in our study. We conclude that our proposed approach in applying colorimetric intensity ratio analysis on hypericin fluorescence shows potential to optically grade bladder cancer in situ.