IP029. Ultrasound Detection of Heterogeneous Accumulated Strain Within 3D Printed Patient Specific Abdominal Aortic Aneurysms

Abstract

In vivo material properties (ie, stiffness) of abdominal aortic aneurysms (AAA) determined by strain across the cardiac cycle typically requires custom imaging techniques not readily available in the clinical setting. Even fewer imaging techniques have been able to detect heterogeneous material properties within the circumferentially oriented aorta. We present a validation of our novel ultrasound (US)-based accumulated strain algorithm using three-dimensional (3D) printed idealized and patient specific AAA tissue mimicking phantoms with heterogeneous material properties. AAA phantoms with an axisymmetric geometry, an idealized AAA geometry, and a computed tomography-derived patient-specific geometry were manufactured using a 3D printer. Homogenous phantoms were created with a 10% and 25% polyvinyl alcohol (PVA) by mass hydrogel, polymerized using freeze/thaw (F/T) cycles, to simulate both healthy and aneurysmal tissue, respectively. Heterogeneous phantoms were produced by casting the anterior wall of the aorta with 25% PVA and the remainder of the vessel with 10% PVA. Lumbar vertebral bodies were 3D printed and embedded in a 7% PVA block and polymerized with two F/T cycles to simulate the abdominal cavity. All models were placed within the simulated abdominal cavity and tested under a physiologic pulse pressure of 60 mm Hg. Abdominal RF ultrasound images were obtained using an Ultrasonix Sonix Touch system at 65 FPS and synchronized with instantaneous pressure/flow measurements. Accumulated mean maximum principle strain (Ɛp) within the axisymmetric phantom was six-times greater (15.6% ± 7.9% vs 2.4% ± 1.0%) in the 10% vs 25% PVA. Total volume change, found via the instantaneous flow measurement, was in good agreement with our ultrasound US Ɛp measurements, with < 0.1% error. Ɛp was approximately two-times greater (10.6% ± 4.7% vs 5.7% ± 3.1%) in the idealized homogeneous 10% phantom compared to 25%. Ɛp measured in posterior wall of the heterogeneous idealized phantom was similar to the 10% homogenous phantom (9.1% ± 6.7% vs 10.6% ± 4.7%), but the Ɛp in the anterior wall was significantly decreased (3.4% ± 1.35% vs 10.3% ± 3.8%). Likewise, a marked difference in Ɛp (2.3% ± 1.4% vs 8.28% ± 5.0%) was detected within the anterior wall of the heterogeneous patient-specific vessel phantom as compared to the homogeneous (Fig). Our novel accumulated strain algorithm is able to detect material differences in the circumferential orientation within a heterogeneous patient specific vessel phantom under standard clinic imaging conditions. Future clinical applications of our accumulated US strain will help to elucidate how material properties of the aorta are modulated in vivo during aneurysmal formation and growth.