|
Tara Yates is now working with the Robotics
Research Group, Dept. of Engineering Science, University of Oxford. |
AbstractThis thesis describes the evaluation of time-resolved optical tomography for detecting and specifying breast disease. Optical tomography involves transilluminating the breast using near infrared (NIR) light. Characteristic absorption by oxy- and deoxy- haemoglobin at NIR wavelengths can be exploited to yield oxygen saturation and blood volume information. This information may provide a distinction between the high vascularisation often associated with malignant lesions and benign or normal breast tissue. A 32 channel time-correlated single photon counting system
is utilised to perform a series of investigations on both healthy
volunteers and patients with pre-diagnosed lesions. Specific
datatypes, extracted from a histogram of the times of flight
of photons across the breast, are used to reconstruct images
of the optical properties (absorption µa and reduced scattering
coefficient µs'). The reconstruction is performed using
a non-linear, finite element based algorithm. A second system based on a hemisphere filled with a coupling fluid is investigated. Preliminary investigations into the optimal optical properties of the coupling fluid and reconstruction techniques are also performed. The first investigations on healthy volunteers are presented. Initial findings suggest that this method provides superior information to the ring system due to its three dimensional capability and its ability to provide consistent coupling. Download thesis: Title [52 kB] |
Page updated: April 28, 2005