GSBS Dissertations and Theses


Design and Performance of a Localized Fiber Optic, Near-Infrared Spectroscopic Prototype Device for the Detection of the Metabolic Status of "Vulnerable Plaque": in-vitro Investigation of Human Carotid Plaque: a Dissertation

Approval Date

January 2003

Document Type

Doctoral Dissertation


Graduate School of Biomedical Sciences


Carotid Stenosis; Fiber Optics; Spectroscopy, Near-Infrared; Academic Dissertations


INTRODUCTION: The "vulnerable plaque" is defined as the "precursor lesion" that ultimately ends in acute coronary thrombi (clots) that create a heart attack. Macrophages and inflammatory cells, found preferentially in vulnerable plaque, sustain their activity in the plaque through anaerobic metabolism and lactate production. The ultimate goal is to assess anaerobic metabolism in-vivo by measuring tissue pH and lactate concentration in atherosclerotic plaques using optical spectroscopy. The proposed in-vitro optical probe design, experimental method, and spectroscopic data analysis methodology are established in this research.

METHODS: A fiber optic probe was designed and built based on both Monte Carlo simulations and bench testing with the goal to collect light from a small volume of tissue. A simulation of the depth penetration of the proposed probe was performed on normal and atherosclerotic aortic tissue, and the final probe was bench tested using normal aorta. A method was developed to preserve plaque metabolic status of tissue harvested from patients. Human atherosclerotic tissue obtained immediately after carotid endarterectomy was placed in Minimum Essential Medium (MEM) with non-essential amino acids supplement, bubbled with 75%O2/20%N2/5%CO2 at 37°C. Tissue pH, pCO2, pO2 and temperature with (n=7) and without (n=2) the media preparation over time were reviewed to assess plaque viability and maintenance of physiological conditions. Additional plaques placed in media were used for development of chemometric methods to measure pH and lactate. Areas of each plaque were randomly chosen for analysis. Reflectance spectra were collected with a dispersive spectrometer (400-1100 nm) and a Fourier-transform near-infrared spectrometer (1100-2400 nm) using the fiber optic probe. Reference measurements for tissue pH and lactate were made with glass microelectrodes and micro-enzymatic assay, respectively. Partial least-squares (PLS) data analysis was used to develop multivariate calibration models on an initial set of 5-6 plaques relating the optical spectra to the reference tissue pH (n=20) or the lactate concentration (n=21) to assess data quality. The coefficient of multiple determination (R2), the standard error of cross-validation (SECV), and the number of factors were used to assess the model performance. Additional points were collected from ~14 plaques and added to preliminary data. Pre-processing techniques were then used to see if preliminary data results could be improved by reducing different sources of variability with the introduction of more points.

RESULTS: Monte Carlo simulations and depth penetration tests with the final probe design showed light is collected from ~1 mm3 volume of tissue using a 50 micron source-receiver separation. Tissue pH, pCO2, pO2 and temperature values demonstrated that the plaques were viable and stable in the media preparation for a maximum of 4 hours. Data from the first six plaques collected for lactate analysis showed that for seventeen points, a six-factor model produced adequate results (R2=0.83 SECV=1.4 micromoles lactate/gram tissue). Data from the first five plaques collected for tissue pH analysis, showed for seventeen different points, a three-factor model produced adequate results (R2=0.75 SECV=0.09 pH units). When additional points were added to either data set, model results were degraded.

CONCLUSIONS: The in-vitro optical probe design and experimental procedures was established and the feasibility of the optical method demonstrated with preliminary data. However, with the addition of more data points, different sources of tissue and spectral variability were observed to affect calibration. The gross pathology type and mismatched optical volume to reference measurement volume limited the tissue pH determination. The reference measurement precision, the spatial resolution of the reference lactate measurement, and unmodeled tissue variability (water and proteins) limited the lactate determination. Large variability in all optical measurements was observed. Additional in-vitro data collection would be required such that the variability due to the tissue is reduced and any spectrometer variability adequately compensated to be able to use the optical calibration in-vivo.


In the process of seeking author's permission to provide full text.

Rights and Permissions

Copyright is held by the author, with all rights reserved.

This document is currently not available here.