Electrocardiographic techniques and methods in the detection of ischaemic heart disease

  • Michael Jennings

Student thesis: Doctoral Thesis

Abstract

Electrocardiographic analysis has the capability to diagnose and locate abnormalities relating to the heart. There are several lead systems, such as the 12-lead ECG, that are commonly used in clinical settings. One aspect of this thesis is to derive additional posterior and right-sided chest leads from the 12-lead ECG, and to evaluate the performance of the derived leads in the detection of ECG changes associated with myocardial ischaemia.

The 12-lead ECG is also not practicable for long term ambulatory monitoring, especially in the detection of paroxysmal cardiac abnormalities such as unstable angina. Therefore, the second study of this thesis introduces a novel patch-based short-spaced lead system sensitive to ST-segment changes associated with ischaemia.

With the increasing numbers of electronically-stored patient data, it is imperative that clinicians can develop their own algorithms. In the third study, a framework for biomedical algorithm development is introduced, with a focus on its use by non-coders to pass data through multi-lingual scripts.

Derived posterior (V7-V12) and right-sided chest leads (V3R-V6R) from the 12-lead ECG were closely correlated to those recorded. Myocardial infarction detection was improved as additional leads were added to the 12-lead ECG, however, this was not statistically significant.

A patch-based short spaced lead system that was sensitive to ST-segment changes associated with ischaemia was suggested. It consisted of two bipolar leads. Coefficients were generated to derive this short spaced lead system from the 12-lead ECG. ST-segment changes associated with ischaemia were detected with the highest F1 score (86.7%).

A web-based framework was introduced to reduce the barrier to entry in biomedical digital signal processing for non-coders. A Python framework was used with the MATLAB Engine to allow users to create algorithms consisting of multi-lingual scripts capable of processing patient data, without the need to write code. The framework was reproducable and scalable.
Date of AwardDec 2022
Original languageEnglish
SponsorsEastern Corridor Medical Engineering Centre
SupervisorJim McLaughlin (Supervisor), Dewar Finlay (Supervisor) & Colin Turner (Supervisor)

Keywords

  • Electrocardiogram
  • Myocardial infarction
  • Heart attack
  • DSP
  • ECG
  • Lead systems
  • Short-spaced lead

Cite this

'