Assessment of photobleaching during endoscopic autofluorescence imaging of the lower GI tract.

Background And Objectives: In autofluorescence endoscopy, the difference in the fluorescence of intrinsic fluorophores is imaged to help visualize pre-malignant lesions, as in the system evaluated here. In this, blue light is used for excitation and the green autofluorescence is normalized by the red diffuse reflectance and presented using a false color scale. The present study was designed to quantify the degree of fluorescence photobleaching induced by the excitation light during use in the colon, since significant photobleaching could lead to false interpretation of the images, particularly false-positive lesions.

Study Design: Measurements were made ex vivo and in vivo, both using the endoscopic imaging system and a separate fiberoptic spectroscopy probe in externalized rat jejunum and in patients undergoing routine colonoscopy, using exposures typical of autofluorescence endoscopic examination.

Results: Photobleaching could be potentially caused at blue light exposure. However, at light intensities and exposure times that are typically used in clinical practice, the average photobleaching (% loss of peak fluorescence intensity) was <1% and <6% in the rat and human tissues, respectively. Nevertheless, the range was large: from -17% to +18% in rats and -33% to +43% in patients, where negative values denote an apparent increase in fluorescence. Both the large positive and negative deviations are believed in part to be due to a measurement artifact caused by uncontrollable tissue motility.

Summary And Conclusions: It is concluded that, using exposures typically encountered in clinical practice, there is minimal photobleaching during fluorescence endoscopy at exposure such as are used in the Onco-LIFE and comparable systems. The small changes in fluorescence intensity and spectral shift that do occur are not likely to be detectable by eye, and so should not impact significantly on the diagnostic accuracy of the technique.