- American Physical Society Sites
- Meetings & Events
- Policy & Advocacy
- Careers In Physics
- About APS
- Become a Member
By Dr. Irene Georgakoudi
Our work at the MIT Spectroscopy Laboratory (http://web.mit.edu/spectroscopy/www/), directed by Dr. Michael Feld, focuses on the development of novel spectroscopic techniques for the biophysical characterization of tissue. Such techniques have the potential to transform the field of medical diagnosis, offering powerful new means for quantitative tissue analysis, wide area surveillance, and biopsy guidance in a non-invasive way. Here, we describe in more detail a recently developed method for analyzing tissue spectra, called Tri-Modal Spectroscopy or TMS. TMS is the combination of three spectroscopic techniques, which characterize different aspects of tissue biochemistry, structure and morphology. Our aim is to use TMS to detect precancerous (dysplastic) changes, not easily detected with currently available technologies.
Fig. 1: (a) Measured tissue fluorescence spectra from non-dysplastic (solid line) and dysplastic (dashed line) Barrett's esophagus tissue sites. (b) Corresponding intrinsic fluorescence spectra. (c) Ratio of NADH to collagen fluorescence to the intrinsic fluorescence spectra of non-dysplastic (squares) and dysplastic (triangles) tissues.
During standard endoscopic procedures, we acquire fluorescence spectra at eleven laser excitation wavelengths between 337 and 620 nm and one white light (350-750 nm) reflectance spectrum in less than one second. Light delivery and collection is mediated through an optical fiber probe. The acquired spectra contain information about the uppermost tissue layers, where almost 90% of cancers begin.
From the recorded fluorescence and reflectance spectra, we extract three types of spectroscopic information: intrinsic fluorescence, diffuse reflectance and light scattering. Intrinsic fluorescence spectroscopy (IFS) refers to the recovery of tissue fluorescence spectra that are free of distortions introduced by tissue scattering and absorption. To remove these distortions, we combine measured fluorescence and reflectance spectra using a photon-migration-based picture1. The extracted intrinsic fluorescence spectra are decomposed to provide quantitative information on the biochemical tissue composition and the changes that take place in pre-cancerous tissues (Fig 1).
The measured reflectance spectra consist mainly of photons that are scattered many times before being detected. We use a model that is based on diffusion theory to describe the diffusely reflected light and, thus, to extract information about the absorption and the reduced scattering coefficients of tissue (diffuse reflectance spectroscopy or DRS2). The reduced scattering coefficient depends mainly on the morphology of the connective tissue, which provides structural support for the epithelium, the most superficial tissue layer. We observe consistent changes in the reduced scattering coefficient of dysplastic tissues (Fig. 2b).
Fig. 2: (a) Measured reflectance spectra from non-dyspalstic (solid line) and dysplastic (dashed line) Barrett's esophagus tissue sites. (b) Slope and intercept of a line describing the wavelength dependence of the reduced scattering coefficient. (c) Nuclear number density plotted vs. nuclear enlargement. Non-dysplastic: squares; dysplastic: triangles.
A small fraction (2-5%) of the reflected photons are detected after undergoing single back-scattering events. The major target particles for this type of scattering are the nuclei of epithelial cells. Changes in the shape, size and number density of cell nuclei are histopathological hallmarks of dysplasia. Analysis of the singly-backscattered light spectrum using light scattering theory (light scattering spectroscopy or LSS), provides information about the size, and number density of cell nuclei (Fig. 2c) 3,4 without tissue removal or processing.
Since IFS, DRS and LSS provide complementary information about tissue biochemistry and morphology, their combined use, i.e. TMS, can serve as an excellent tool for biophysical tissue characterization and the detection of pre-cancerous lesions. Indeed, TMS is a superior tool for the detection of dysplastic changes in Barrett's esophagus 5 and the cervix.
Presently, we are testing software that is designed to perform TMS analysis in 4-8 s at the time of data collection. Thus, we can test directly the potential of this tool as a real-time guide to biopsy, and, ultimately, as a tool that could, in some cases, replace biopsies.
1. Zhang, Q. et al. Optics Letters 25, 1451-1453 (2000).
2. Zonios, G. et al. Applied Optics 38, 6628-6637 (1999).
3. Perelman, L. et al. Phys Rev Let 80, 627-630 (1998).
4. Backman, V. et al. Nature 406, 35-36 (2000).
5. Georgakoudi, I. et al. Gastroenterology 120, 1620-1629 (2001).
Irene Georgakoudi, Ph.D.
Phone: (617) 258-9487.