My recent research has focused on Monte Carlo simulations of light transport in layered tissues. Light transport is currently used clinically both as a therapeutic tool (e.g., photodynamic therapy) and a diagnostic tool (reflectance and/or fluorescence measurements of tissues). A concern in all these cases is the difficulty of knowing which regions of the tissues are sufficiently illuminated for therapeutic results, or from which regions the collected fluorescence was emitted.
A Monte Carlo computer program models the process of light transport by simulating the motion of a large number (typically hundreds of millions) of photons as they interact with the tissue. The program models the absorption of photons (which may be followed by fluorescence or photodynamic therapy), the scattering of the photons within the sample and at interfaces, and the possible detection of the photons. These models must rely on accurate experimental data for a variety of optical properties of the tissues under study, and must accurately model all the photo-physical processes in the tissues.
From the detailed results of simulations, one can determine, e.g., which regions of the tissues are illuminated or from which areas the detected light was produced. Often many simulations are needed to explore the expected range of some parameters and find an optimum set of values. One can also determine the sensitivity of a particular result to a poorly-known parameter by varying that parameter between many simulations.
My current Monte Carlo program is a single-computer Fortran-95 program that handles the absorption and scattering of light in rectangular samples. It has been validated against measurements on tissue phantoms and against other Monte Carlo codes. I am currently collaborating with Dr. Linda Jones at the College of Charleston, Dr. Mike Wallace at the Mayo Clinic in Jacksonville, Florida, and with Dr. Laura Marcu and her colleagues at the Cedars-Sinai Medical Center in Los Angeles.
The photon path in a highly scattering medium (like skin) can be very complex! The two images below show paths from two simulations that differed by the value of the anisotropy factor "g". This is the average of the cosine of the angle through which the photon is scattered.
If g=0, the light is uniformly scattered, while if g=1 the light is all scattered forward. g=-1 corresponds to strong backscatter (like your headlights in fog). The left-hand image has g=0.20 (fairly isotropic), while the right-hand image has g=0.90 (strong forward scattering). We see that the "g" value influences how deep in the tissue the photon will travel before re-emerging from the skin.
Click on each image for a larger version.
|
|
|
 |
 |
 |
|
These show the volume of tissue that are sampled by an "integrating sphere" detector which records all the light that reflects back from the sample. Other images on the individual pages (click the pictures) show how this volume changes if only portions of this total reflected light are recorded. The samples differ in their optical properties: the left-most has approximately the characteristics of human skin, the second contains particles that scatter light more in the forward direction, and the third shows a sample with very little absorptions. |
||