A 38-year-old woman, who had undergone correction of congenital supravalvular aortic stenosis at 8 years of age, developed progressive exertional dyspnea and chest pain for more than a year prior to the current presentation. Electrocardiogram-gated contrast-enhanced computed tomography (CT) showed stenosis in the ascending aorta and highly dilated bilateral coronary arteries (Fig. 1). Echocardiography showed normal left ventricular (LV) wall motion with an ejection fraction of 59%, prominent LV hypertrophy, mild aortic regurgitation, and flow acceleration of 5.2 m/s in the ascending aorta. Thallium myocardial scintigraphy did not detect myocardial ischemia. Catheter examination demonstrated a pressure drop of 110 mmHg at the stenosis site of the ascending aorta with high pressure in the sinus of Valsalva. Three-dimensional time-resolved cine phase contrast magnetic resonance imaging (4D flow MRI) demonstrated preserved LV wall motion and cardiac output but highly accelerated flow (> 5.0 m/s) in the ascending aorta. Release of stenosis was necessary to resolve the patient’s complaint and avoid LV functional deterioration due to excessive afterload; however, when we had performed simply aortic root replacement with the in situ coronary revascularization during surgery, reduction of the blood pressure in the sinus of Valsalva may have led to demand–supply mismatch in coronary flow due to heavily dilated and tortuous coronary artery and the highly thickened LV wall muscle. Moreover, we also suspected to harvest dense adhesion around the sinus of Valsalva due to the old patch material in this reoperation. Therefore, we had to examine whether CABG could be a candidate or alternative option for coronary reconstruction. However, we could not determine which strategy of revascularization could supply exactly more sufficient coronary perfusion, CABG or in situ reconstruction with the Carrel patch technique. Therefore, we performed computerized virtual surgery based on CFD combined with 3D CG. The CFD method was based on our previous validation studies [4, 5]. We used the data obtained from thin-slice, early-phase, enhanced multidetector-row CT imaging. These imaging data were transformed to a 3D patient-specific geometry by Osirix (Osirix Foundation, Geneva, Switzerland) and 3D-Coat (PIGWAY, Kiev, Ukraine). Computational meshes were created with ANSYS-ICEM CFD 16.0 (ANSYS Japan, Tokyo, Japan). To simulate the blood flow, the inlet boundary conditions in the aortic root were set as the mass flow boundary condition with a pulsatile wave. Cardiac outputs were set at 5.0 l/min based on the cardiac catheterization data. The outlet boundary conditions were used as the pressure boundary conditions that represented the external forces outside of the analysis domain. The physiological external forces in the neck vessels and the descending aorta were composed of the reflection wave from the peripheral tissue, vascular inertance, and autonomic regulation. In coronary artery outlets, time-varying impedance that expressed the systolic ventricular muscle contraction and diastolic relaxation, can realize perfused muscle volume or peripheral vascular bed capacity, as reported previously [3, 6]. Finite volume method with backward Euler transient calculation was used to solve the Navier–Stokes equation with the convergence criteria of 1.0E-5 for all parameters. Calculated flow velocity distribution inside the aorta resulted in almost the same flow map as that detected in the 4D flow MRI. The calculated pressure drop in the ascending aorta was high and almost the same as that measured in the catheter examination. The computerized virtual postoperative models with 3D CG were of two types: (1) in situ coronary reconstruction with Carrel patch technique for root replacement; (2) CABG procedure using the bilateral internal thoracic artery (ITA) to the left anterior descending artery and circumflex branch and saphenous vein graft from the ascending aorta to the right coronary artery (RCA) in addition to root replacement. The predicted postoperative coronary flow supply is illustrated in Fig. 2. The shortage of blood supply to the left-side coronary arteries was detected in the CABG model despite using the bilateral ITA. Compared with the preoperative state, in situ coronary reconstruction with the Carrel patch technique decreased the coronary blood flow in the systole due to the reduction in the Valsalva pressure but obtained sufficiently increased coronary flow in the diastole because of the increased intensity of reflection wave from the peripheral vasculature (Fig. 3).
Based on these results, we performed aortic root replacement using a 24-mm Valsalva graft (Japan Lifeline Co. Ltd., Tokyo, Japan) implanted with a 21-mm bi-leaflet mechanical valve (St Jude Medical, St Paul, Minnesota, USA) with in situ coronary reconstruction using the Carrel patch technique. Before surgery, we planned to perform the valve-sparing aortic root replacement as possible. However, during surgery, we observed the severely degenerated and thickened aortic leaflet with the degeneration of the sinus of Valsalva. Thus, because these aortic leaflets could not be expected for long-term durability, we determined that the aortic valve should be replaced. The patient was discharged 15 days postoperatively without any complications in the coronary arterial system. A postoperative electrocardiogram revealed non-ST elevation. Moreover, echocardiography and contrast-enhanced CT showed preserved LV function and sufficiently widened ascending aorta without coronary arterial stenosis or thrombosis, respectively. Two years after the surgery, she was well with no angina symptoms.