Identification of in vivo constitutive parameters of thoracic aortic aneurysms based on the...

Accurate patient-specific aortic-wall constitutive parameters are essential for biomechanical risk assessment of thoracic aortic aneurysms (TAA), yet their non-invasive determination remains challenging. Here, we proposed an inverse finite element method (IFEM) based on our unified-fiber-distribution (UFD) model to identify in vivo constitutive parameters of ascending TAA from patient-specific CT images. The UFD model considers aortic tissue fibers as a unified distribution, which is more physically consistent with the real fiber distribution than existing models (e.g., the Gasser-Ogden-Holzapfel [GOH] model) that separate fibers into several fiber families. We performed ex vivo biaxial testing on specimens from 30 TAA patients and compared the UFD model with the GOH model and two reduced-order isotropic models. Among the four models, the UFD model achieved the best fitting performance, with the lowest mean squared error (9.5±1.1kPa²) and the highest coefficient of determination (R²=0.977±0.002), significantly outperforming all comparators (all p<0.01), with one fewer parameter than the GOH model. We then applied the UFD-based IFEM to estimate in vivo constitutive parameters from multi-phase CT data of the same 30 patients. The in vivo CT-predicted parameters demonstrated good agreement with ex vivo ground truth values, with a mean relative error of 8.5% ± 0.6% across all parameters and patients, and sub-millimeter geometric agreement (0.77 ± 0.03 mm) between predicted and CT-derived diastolic configurations. These results demonstrate that the UFD-based IFEM can accurately identify patient-specific constitutive parameters non-invasively, representing a significant step toward clinical translation of biomechanical risk assessment for TAA. STATEMENT OF SIGNIFICANCE: Accurate patient-specific constitutive parameters of the aortic wall are essential for biomechanical risk assessment of thoracic aortic aneurysms (TAA), yet their non-invasive determination remains a major challenge. This study presents an inverse finite element method based on the unified-fiber-distribution (UFD) constitutive model to non-invasively identify aortic wall mechanical parameters from routine CT scans. Validated against ex vivo tissue testing in 30 patients, our approach achieved a mean parameter error of only 8.5%-an order-of-magnitude improvement over prior methods. This framework could enable preoperative, patient-specific biomechanical risk assessment without tissue excision, advancing clinical translation of wall stress analysis for aneurysm management.