Abstract
A novel optical approach is proposed for cancerous tumor detection using transient radiation signals. In this method, target tissue is illuminated by near-infrared ultrashort laser pulses from various surface source points, and backscattered time-resolved light signals are collected at the same surface points. By analyzing the log-slopes of decaying signals over all points on the source-detection grid, a log-slope distribution on the surface is obtained. After administration of absorption contrast agents, the presence of cancerous tumors increases the decaying steepness of the transient signals. The mapping of log-slope difference between native tissue and absorption-enhanced cancerous tissue indicates the location and projection of tumors on the detection surface. In this paper, we examine this method through the detection of a small tumor in a model tissue phantom via computer simulation. The model has a spherical tumor of 6 mm in diameter embedded at the tissue center. Monte Carlo methods were employed to simulate the light transport and signal measurement. It is shown that the tumor in the tissue model can be accurately projected onto the detection surface by the proposed log-slope difference mapping method. The image processing is very fast and does not require any inverse optimization in image reconstruction.