Infrared video imaging of subsurface vessels: a feasibility study for the endoscopic management of gastrointestinal bleeding.

The feasibility of infrared video imaging of subsurface vessels in the stomach was investigated both experimentally and in more detail using computer simulations of light propagation. Infrared video imaging was first attempted in several experimental situations. Images of a human arm illuminated with infrared light (wavelength > 700 nm) revealed subcutaneous venous structures not revealed by visible light (wavelength of 500 to 600 nm). An infrared-sensitive video endoscope was used to view both a human arm and normal stomach wall. Infrared illumination within the stomach enhanced only the larger subsurface vessels. Infrared transillumination of a rat skin flap window chamber allowed video recording of images during injection of an absorbing dye, indocyanine green, into the blood volume and showed that indocyanine green can enhance the contrast in infrared images of small vessels. Computer simulations of vessels of varying depths and sizes indicated successful detection was possible by infrared imaging. Computer simulations demonstrated that the shadow caused by an imaged subsurface vessel has two characteristics: (1) the central loss of reflectance, which indicates the size of the vessel, and (2) the full-width half-maximum of the reflectance loss, which indicates the depth of the vessel. The simulations further suggested that images of small vessels can be dramatically enhanced (68-fold) by indocyanine green, which attenuates the transmittance of scattered light from behind the vessels to the surface for observation. On the other hand, indocyanine green enhances the contrast of large vessels to a lesser degree (2.6-fold). The ultimate goal is to develop an endoscopic video imaging system capable of capturing reflected light from the stomach wall.(ABSTRACT TRUNCATED AT 250 WORDS)