The majority of day-to-day clinical decision making is guided by the laboratory measurement of a biomarker. Despite monumental advances in biomarker discovery, it is estimated that on average less than one novel biomarker is approved on a yearly basis. This thesis examines characteristics of current (acute) cardiac care biomarkers, and subsequently demonstrates various approaches to optimize current biomarkers rather than deploying ones. The first five chapters evaluate analytical and clinical characteristics of the cardiac biomarkers troponin (protein involved in substraction of the heart) and natriuretic peptides (hormones secreted by the heart). Chapter six attempts to improve the clinical specificity of cardiac troponins by enabling hypoxia-specific detection of troponin. Chapter seven to ten apply machine learning to combine multiple biomarker results into clinically relevant outputs. Future research should further examine the proposed approaches and study their beneficial value in day-to-day clinical decision making.
|Qualification||Doctor of Philosophy|
|Award date||20 Jan 2023|
|Place of Publication||Maastricht|
|Publication status||Published - 2023|
- cardial biomarkers
- machine learning
- clinical relevance