Imaging lung tissue absorption in thoracoscopy surgery from coherent multiple scattering of near-infrared light

Abstract

The thoracoscope is the standard tool for lung tumor segmentectomy. Traditional thoracoscopy uses white light illumination and camera recording to provide real-time and high-resolution video. Surgeons inflate the lung with pure oxygen and wait 20 ~ 25 minutes to identify the resection region. This determination procedure is based on the color and structure changes in tissue. Tissue physiological information, especially the absorption underneath the superficial layer, can assist surgeons in more precise differentiation of the healthy and diseased regions, thus shortening this waiting period. To extract the deep tissue absorption during thoracoscopy surgery, we utilized the single wavelength coherent light in the near-infrared (NIR) window (842 nm diode laser). The coherent light propagation facilitates the statistical separation of multiple scattering component based on the random matrix description. The deep tissue absorption centered at the depth of transport mean free path (l_t) can be calculated based on multiple scattering component. Both Monte Carlo simulation and in-vitro phantom experiment validate the proposed method. Together with the white light video, the pseudo-colored absorption map assists the surgeons in shortening the waiting time by 7 minutes in a clinical trial (n=5, Video-assisted thoracic surgery (VATS) segmentectomy). The proposed method opens a new avenue in extracting deep tissue physiological information based on the random matrix description of coherent light propagation.