Captur, G;
(2015)
The Fractal Complexity of Myocardial Trabeculae.
Doctoral thesis , UCL (University College London).
Abstract
Objectives: This thesis centres on the development of a technique for quantifying left ventricular (LV) trabeculation using fractal geometry. Fractals are natural phenomena or mathematical sets of infinitely complex patterns that are self-similar across different scales. In humans, my objectives were to use a fractal analysis to i) better diagnose LV noncompaction (LVNC); ii) describe population-based reference ranges; iii) appraise the role of trabeculae as a pre-phenotypic disease marker in hypertrophic cardiomyopathy (HCM). In the mouse, my objectives were to use a fractal analysis to i) track the transformation (“compaction”) in myocardial wall structure during embryogenesis; ii) describe the trabecular profile in LVNC models. Background: Quantification of trabeculation is important for the diagnosis of LVNC. Current criteria have limitations. The complexity of trabeculae led us to hypothesize that a fractal approach applied to imaging datasets of the heart (cardiovascular magnetic resonance [CMR] in humans; high-resolution episcopic microscopy [HREM] in mice) would be of value. Methods: We created a fractal tool for human application in MATLAB® and subsequently OsiriX, and then deployed it within the commercial software platform, cvi42 for distribution (see Chapter 6). Using the former I measured fractal dimensions (FD) in a number of key healthy and diseased cohorts including: 135 LVNC patients/controls, 176 HCM patients/matched controls (including subclinical HCM in a 13- centre study) and 2,547 participants from the Multi-Ethnic Study of Atherosclerosis (MESA). I adapted the methodology to analyse the trabecular development in 123 embryonic mouse hearts at stages, E14.5, E15.5, E16.5, E17.5 and E18.5. Results: Human CMR fractal analysis was reproducible at multicentre scale (Chapters 8 - 11). The FD varied in a characteristic pattern from LV base to apex. The maximal apical FD (FDMaxApical) was increased in LVNC (1.392 ± 0.010) and in overt HCM (1.370 ± 0.08), and in subclinical HCM it was predictive of sarcomere gene mutation carriage (β = 1.6, P < 0.001). In healthy hearts, it was higher in African-Americans, Hispanics, and hypertensives and lower in Chinese-Americans compared to Caucasians. In Chapter 12 I show how the mature embryo heart exhibits a base-to-apex trabecular pattern similar to that observed in humans, and how FD falls (especially basally) with development, but progressed abnormally in LVNC models. Conclusions: Fractal analysis represents a new method for objectively quantifying myocardial trabecular complexity. The tools I have created are simple to use and commercially/freeware available. They can demonstrate the difference between health and overt/preclinical disease, and between different ethnicities. New insights into cardiac development in humans and mouse, particularly in HCM are presented, as are insights into cardiac adaptation in health.
Archive Staff Only
 |
View Item |
