Computer Modelling Informed Coronary Artery Bypass Grafting

Abstract

Coronary artery disease is the leading cause of mortality and morbidity worldwide. Coronary artery bypass grafting operations have been performed as a treatment for this disease for over 60 years. There has been recent interest in the exclusive use of arterial grafts as conduits for this operation, in particular bilateral internal mammary arteries, on account of their longevity. However, there still exists significant variability in grafting strategy amongst cardiac surgeons and there remains controversy as to the optimal grafting configuration for a given patient. This is due to the complexity of the biophysics of a particular individual’s coronary circulation and the resulting degree of uncertainty facing a surgeon. The overall aim of this thesis was to investigate whether predictive patient-specific computational modelling could aid cardiac surgeons in their selection of grafting configurations using bilateral internal mammary and radial arteries for patients with severe triple vessel disease. In particular, the purpose was to avoid the selection of unsatisfactory grafting configurations that would result in grafts with limited patency due to competitive flow and/or steal of flow. This study was conducted by recruiting 16 cardiac surgeons (8 experts and 8 novices). Surgeons selected their preferred grafting configuration for each of 5 patients, including an anaortic technique which could be performed without the need for cardiopulmonary bypass. Five patient-specific 1D-0D lumped parameter computational models were created to compare 12 total arterial grafting configurations used internationally. The hemodynamic predictions included mean graft flows, pulsatility index and degree of regional myocardial perfusion restored. Surgeons then selected their preferred grafting configurations after viewing the computational data. The use of computational models led to a significant decrease in the selection of unsatisfactory grafting configurations using anaortic techniques. It also resulted in more patient-specific tailoring of individual grafting configurations for these techniques. Computer modelling informed novices were more likely to choose optimal grafting configurations whereas experts were content with satisfactory grafting configurations. Novices were also more likely than experts to be influenced by the computer model in selecting grafting configurations that they would usually avoid the use of. The 1D-0D predictive patient-specific models demonstrated that the arterial grafting configurations using composite and sequential grafts are patient-specific. Based on these results it is concluded that grafting configurations should be adjusted specifically for each patient. The use of computer modelling can aid surgeons in making better decisions for patients particularly in the context of more technically challenging anaortic operations that use bilateral internal mammary and radial arteries in patients with triple vessel disease.