201. Optimising approximate entropy for assessing cardiac dyssynchrony with radionuclide ventriculography
- Author
-
K. A. Jones, A. D. Small, C. A. Paterson, J. Robinson, W. Martin, N. E. R. Goodfield, and David Hamilton
- Subjects
Computer science ,business.industry ,0206 medical engineering ,Biomedical Engineering ,Environment controlled ,Health Informatics ,Radionuclide ventriculography ,Pattern recognition ,02 engineering and technology ,medicine.disease ,Appropriate use ,020601 biomedical engineering ,Approximate entropy ,Phase image ,03 medical and health sciences ,0302 clinical medicine ,Signal Processing ,medicine ,Artificial intelligence ,Ventricular dyssynchrony ,business ,030217 neurology & neurosurgery - Abstract
Left ventricular dyssynchrony can be assessed with phase parameters from radionuclide ventriculography (RNVG), including approximate entropy ( ApEn ). The input values used to calculate ApEn will affect the results significantly, so it is essential to optimise ApEn for the application. However to date, no optimisation for ApEn applied to images has been published. In this paper, generated data were used to simulate patient phase images, allowing the input parameters for ApEn to be tested and optimised in a controlled environment. Clinical images were then used to confirm that the selected parameters were appropriate. The results demonstrate the effect of input parameters for ApEn and the most appropriate use with RNVG phase images. This work demonstrates the importance of optimisation and standardisation when using ApEn as a measure of dyssynchrony.
- Published
- 2021