부산대학교 도서관

Despite the fact that cardiovascular disease (CVD) mortality rates have seen major declines during recent years compared to past decades, reflecting advances made in primary and secondary prevention, CVD remains a significant mortality contributor, and the so-called residual cardiovascular risk persists. For the purposes of my thesis, I used the Oxford Fat, Vessels and Fat (Ox-HVF) cohort, which constitutes a uniquely phenotyped cohort, including >1,000 patients undergoing cardiac surgery, with long-term follow-up, to perform observational association studies, functional ex vivo bioassays on human tissue, and discovery transcriptomics combined with radiomics. In the first part of my thesis, I focused on characterising the prognostic value of arterial and myocardial oxidative stress. I showed that increased production of free radical oxygen species in human arteries and myocardium is associated with a higher risk of future major adverse cardiac events (MACE). I further revealed S100A8 and S100A9 as the most dysregulated genes in the presence of oxidative stress in the vascular wall. Applying discovery network transcriptomics, I unveiled a transcriptomic signature that is redox related and predicts future risk of MACE. Pathways relevant to oxidative phosphorylation, glucose metabolism, and the mitochondrial respiratory chain assembly were enriched and therefore revealed as potential therapeutic targets. In the second part of my thesis, I developed a method that allows the integration of transcriptomic and radiomic data to produce biologically relevant imaging signatures from radiomic features calculated from segmented three dimensional volumes. I applied this concept to radiomic features of perivascular adipose tissue around the internal mammary artery and thoracic aorta imaged in computed tomography angiography (CTA) to generate an imaging signature that captures vascular inflammation defined by upregulation of key inflammatory pathways in arterial transcription. This signature was validated in the context of vascular inflammation induced by the SARS-CoV-2 virus, and was proven to predict in hospital death, admission to the intensive care unit, response to dexamethasone, and systemic thrombosis. Taken together, my findings propose oxidative stress, and specific redox-relevant genes and pathways as prognostic and therapeutic targets in CVD. Further, COVID-19 related vascular inflammation can be captured non-invasively by a novel radiotranscriptomic perivascular adipose tissue signature, which can maximise the diagnostic and prognostic yield of CTA.