Enabling cardiovascular multimodal, high dimensional, integrative analytics

Roth Cardoso, Victor ORCID: 0000-0002-9588-6304 (2022). Enabling cardiovascular multimodal, high dimensional, integrative analytics. University of Birmingham. Ph.D.

Preview

RothCardoso2022PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (5MB) | Preview

Abstract

While traditionally the understanding of cardiovascular morbidity relied on the acquisition and interpretation of health data, the advances in health technologies has enabled us to collect far larger amount of health data. This thesis explores the application of advanced analytics that utilise powerful mechanisms for integrating health data across different modalities and dimensions into a single and holistic environment to better understand different diseases, with a focus on cardiovascular conditions. Different statistical methodologies are applied across a number of case studies supported by a novel methodology to integrate and simplify data collection. The work culminates in the different dataset modalities explaining different effects on morbidity: blood biomarkers, electrocardiogram recordings, RNA-Seq measurements, and different population effects piece together the understanding of a person morbidity. More specifically, explainable artificial intelligence methods were employed on structured datasets from patients with atrial fibrillation to improve the screening for the disease. Omics datasets, including RNA-sequencing and genotype datasets, were examined and new biomarkers were discovered allowing a better understanding of atrial fibrillation. Electrocardiogram signal data were used to assess the early risk prediction of heart failure, enabling clinicians to use this novel approach to estimate future incidences. Population-level data were applied to the identification of associations and temporal trajectory of diseases to better understand disease dependencies in different clinical cohorts.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):