Bad Blood: Combining Data Analytics and Chemical Kinetics to Study Human Blood Coagulation in Certain Diseases

Abstract

Complete description of blood coagulation pathways with respect to patientspecific characterization presents a major challenge. Characteristics of blood coagulation vary drastically between patients. It is essential to characterize abnormalities in blood coagulation to diagnose and treat cardiovascular diseases better.

Given the paucity of patient-specific data to characterize and model the system, there is a greater need to regularize patient-specific models and methods effectively. In this dissertation, we formulate actionable questions and describe our methodology and results.

First, we explore a practical application for using models to classify acute coronary syndrome and coronary artery disease. The classification models were built based on a chemical kinetics model reported in the literature. In a diagnostic setting, the classification models could be employed to screen thousands of patients with greater certainty every year.

Second, we propose a simplified model for a key part of the blood coagulation cascade that demonstrates robust predictive capabilities. The model predicts prolonged activity of thrombin, an important enzyme in the clotting process, in certain plasma factor compositions. The activity sustains beyond the time which is conventionally considered to be the end of clotting. This observation along with the simplified model is a necessary step towards effectively studying clotting in realistic geometries.