Advisor(s)

Dana H. Brooks

Contributor(s)

Eric Miller, David Boas, Stefan Carp

Date of Award

2008

Date Accepted

6-2008

Degree Grantor

Northeastern University

Degree Level

M.S.

Degree Name

Master of Science

Department or Academic Unit

College of Engineering. Department of Electrical and Computer Engineering.

Keywords

Electrical engineering. Diffuse optical tomography

Subject Categories

Tomography, Cancer--Diagnosis

Disciplines

Engineering

Abstract

Diffuse optical tomography (DOT) is an imaging modality with potential use in breast cancer detection, diagnosis and monitoring, brain imaging, and small animal imaging. In this work we focus on breast cancer imaging. It has been observed that malignant and normal tissues differ in blood flow (F) and oxygen consumption (OC). We are interested in using DOT to image those parameters since they are functionally related to the main absorbers of near infrared light in tissue, oxygenated and deoxygenated hemoglobin. In particular we look at dynamic changes in F and OC when the breast is put under mammo graphic compression, as is done using an experimental combined X-ray/DOT system under development by our collaborators at Mass General Hospital. It is believed that these changes provide a potential source of contrast between tumor and other tissue. In this work we analyze, in a simulation scenario, the ability of the imposition of a hemodynamically constrained model of temporal changes in oxygen saturation (SO2) to improve our algorithm for extracting OC and F from reconstruction of hemoglobin concentrations. Using a three dimensional transmission measurement geometry we investigate how the recovery of metabolic parameters is influenced by the observation time interval, the SNR of the measurements, the regularization parameter choices, and a assumptions about the dynamic behavior of total hemoglobin concentration. We conclude that we can reconstruct a 1.5x OC and 2.5x F contrast for a spherical inclusion with less than 20% error as long as we use at least 90 seconds of 60dB SNR data even in the presence of variations in total hemoglobin concentration. Further more we have noticed that by introducing a dynamic model we are able to somewhat compensate for the errors in the reconstructed tumor OC.

Document Type

Master's Thesis

Rights Holder

Eleonora Vidolva


View Full Text

Click button above to open, or right-click to save.

Included in

Engineering Commons

Share

COinS