Evaluation of the clinical efficacy of texture analysis in myocardial quantification

Traditional image analysis within clinical imaging has focussed largely on the visual inspection and analysis of images of the heart. However a wealth of information is present in each and every medical image that is simply beyond the reading of the human eye. One advance in medical image analysis is that of texture analysis which measures the non-uniformity of signal within body tissues. This has been most widely explored in the field of cancer imaging where tumour non-uniformity has been shown to provide diagnostic information on the malignancy of a tumour, its response to treatment and its long term prognosis. However texture analysis has been poorly analysed in the field of cardiac imaging. Three pre-existing well characterised research cohorts involving over 3,500 participants will be used to assess this new image analysis technique in a large group with minimal cost.
The goal of this research project is to evaluate this novel image analysis technique to determine how the heart muscle texture changes in health, in disease, and in response to drug therapy. It will also assess if this characterisation of the heart muscle improves on the information provided by current image analysis.
Potential benefits of this involve the development and validation of a novel image analysis technique that can be applied in routinely acquired clinical images thus expanding the use and diagnostic information that can be extracted from these without any extra burden to the patient and minimal extra cost to the health service.