Noninvasive detection of inhomogeneities in turbid media with time-resolved log-slope analysis

Abstract

Detecting foreign objects embedded in turbid media using noninvasive optical tomography techniques is of great importance in many practical applications, such as in biomedical imaging and diagnosis, safety inspection on aircrafts and submarines, and LIDAR techniques. In this paper we develop a novel optical tomography approach based on slope analysis of time-resolved back-scattered signals collected at the medium boundaries where the light source is an ultrafast, short-pulse laser. As the optical field induced by the laser-pulse propagates, the detected temporal signals are influenced by the optical properties of the medium traversed. The detected temporal signatures therefore contain information that can indicate the presence of an inhomogeneity as well as its size and location relative to the laser source and detection systems. The log-slope analysis of the time-resolved back-scattered intensity is shown to be an effective method for extracting the information contained in the signal. The technique is validated by experimental results and by Monte Carlo simulations.

Introduction

Many applications of near-infrared (NIR) ultrashort laser pulses in optical tomography have been proposed and studied extensively by various researchers in the past few years. These investigations are motivated by the potential for noninvasively assessing the optical properties of absorbing–scattering media such as biological tissues. Unlike X-ray-based computed tomography (CT) that relies on high-energy rays, optical tomography utilizes low powered NIR light that is benign to human skin and tissue. There are several advantages associated with time-resolved optical imaging. For example, the instrumentation system is small, less complex and that the signal detection can be easily obtained with less manipulation and restriction. However, the detection of foreign objects inside a turbid medium proves to be a challenging task because the incident laser light experiences energy attenuation and multiple scattering events within the turbid medium before reaching the detectors located at boundary surfaces. Rapid advances in laser technology have made it possible to obtain pulses in sub-picosecond and femtosecond time scales and this has opened new and promising avenues for time-resolved optical imaging. Using this technology, either diffused, ballistic or snake components of light signals can be utilized for optical imaging of optically thin to thick media as long as the detected signal is strong enough to be filtered from background noise and experimental uncertainty. Several recent studies [1], [2], [3] have reported on the feasibility of determining optical properties of thick turbid media from time-resolved light scattering measurements via simply applying the diffusion theory.

In this paper, we have investigated the possibility of detecting an inhomogeneity in turbid media by analyzing the temporal signals at the decaying tail of the backscattered laser pulse. We propose a new log-slope analysis method that is simple, innovative and efficient. The verification of this method in detecting an inhomogeneity in a turbid medium is conducted experimentally by embedding a tiny graphite inclusion into a block of tissue phantom. A Monte Carlo (MC) simulation program has also been implemented to model the 3D transient laser radiative transfer for further parametric investigations and other relevant analyses. Brewster and Yamada [3] have shown that the asymptotic log-slope of a homogeneous semi-infinite slab is steeper as the scattering albedo decreases. A study conducted by Zaccanti et al. [4] also shows that there is a significant dependence between the scattering coefficient of a homogeneous medium and the broadening of laser pulse, thus affecting the steepness of asymptotic log-slope. Recently, Guo and Kim [5] further showed that the asymptotic log-slopes of reflected signals in a finite 3D geometry are proportional to the absorption coefficient of the embedded inhomogeneity. Our concept of detecting an inhomogeneity of relatively high absorption compared to the surrounding tissue is to evaluate the log-slope from back-scattered laser signals at various positions on the surface of the medium. The influence of absorptivity and size of the embedded inhomogeneity on the log-slope is studied. The effects of laser and detector positions for locating the graphite inclusion are also examined.