Introduction
Cancer therapy related cardiotoxicity is an important factor when considering patient morbidity and mortality. It has long been recognised that irradiation of the heart can cause a spectrum of cardiomyopathies and certain chemotherapeutic agents are also known to be cardiotoxic. The current standard measure of cardiac function is Left Ventricular Ejection Fraction (LVEF), which measures the efficiency of cardiac output. A generally accepted healthy LVEF value of >= 55% is used in conjunction with other medical factors to evaluate cardiac health. However LVEF is a gross measure that is insensitive to regional variances. The heart is able to compensate for slight variances in stroke volume by enlarging, but large variances are irreversible, and once LVEF becomes reduced it can be difficult to return to a healthy value. If cancer therapy patients with early stage cardiomyopathies, before LVEF becomes reduced, can be identified then medical intervention can be used to reduce morbidity and prevent reduced LVEF.
Materials and Methods
Two quantities, synchrony and entropy, have been devised which may be sensitive enough to determine when such subtle changes in the heart occur. This study evaluates synchrony and entropy using retrospective data from cancer therapy patients, and compares it with LVEF and other established measures of cardiac synchrony. Approximate Entropy (ApEn) is a statistical measure of regularity and has been applied to the patient data to determine its usefulness in quantifying cardiac synchrony. ApEn was then localised to study spatial variations in the value.
Results
The synchrony and entropy results are inconclusive and warrant further investigation. Both parameters return non-significant results, but taken in the context of other results, these do not completely disprove their usefulness. ApEn was found to give a significant result which is in agreement with LVEF time-trend. The spatial variation of ApEn gives some interesting information which is difficult to interpret.
My dissertation can be found here as a pdf, or the LaTeX code is here. If you don't know what LaTeX is, look here.
IDL is a pretty easy language to learn. I was able to pick it up quickly and wrote my dissertation project using it. It is probably my favourite language because of it's ease of use plus the great usenet/google group that make it easy to ask questions and offer expert answer.
My dissertation project code can be found here, however without the actual data that it is used to analyse this may be a little dry.