Since its conception decades ago, pulse oximetry-the non-invasive measurement of
arterial blood oxygen saturation in real-time-has proven its worth by achieving and
maintaining its rank as a compulsory standard of patient monitoring. However, the
use of oversimplified models to describe and implement the technology has limited its
applicability and has had its evolution at a near standstill for the past decade.
Currently available technology relies on empirical calibrations that consist of the
correlation between simultaneous measurements from pulse oximeters and invasively
acquired arterial blood samples from test subjects, mainly because the mathematical
models underlying the technology are not sufficiently descriptive and accurate.
Advances in knowledge of human tissue optical properties, computing power and
sensing technology all contribute to a new realm of expansion for pulse oximetry
This research project aims to develop a methodology for improving optophysiological
models of pulse oximetry through the use of a validated Monte Carlo
simulation platform for optical propagation in arbitrary geometries. The platform
aims to arrive at a model that can predict the widest range of empirical outcomes
while maintaining the highest possible level of accuracy. To this end, an empirical
platform and a corresponding experimental protocol is developed towards an
increasingly repeatable standard, thus providing an empirical output for validation of
simulated data. Subsequently, the parameters and coefficients of the optophysiological
model at the core of the simulation platform are iterated until a high
level of correlation is achieved in their outputs. This gives way to a new approach to
the opto-physiological modelling of pulse oximetry.
A Doctoral Thesis. Submitted in partial fulfillment of the requirements for the award of Doctor of Philosophy of Loughborough University.