نوع مقاله : مقاله پژوهشی
نویسندگان
1 گروه پژوهشی ارتوئدی-بیومکانیک، گروه مهندسی مکانیک ، دانشکده مهندسی، دانشگاه بیرجند، بیرجند،خراسان جنوبی،ایران
2 گروه پژوهشی ارتوپدی-بیومکانیک، گروه مهندسی مکانیک، دانشکده مهندسی، دانشگاه بیرجند، بیرجند، خراسان جنوبی، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Background and Aims: Finite element analysis has become one of the common approaches for predicting load distribution between the foot and different supporting structures, providing valuable insights into the internal stresses of bone structures. The construction of accurate models using this method relies on medical imaging data to reconstruct the geometry of bones and tissues. CT scanning is the most widely used imaging modality for bone visualization and serves as the main basis for biomechanical joint modeling. However, radiation-related concerns, particularly in repeated imaging of individuals, children, or sensitive populations, highlight the need for safer and radiation-free alternatives such as MRI. Despite the advantages of MRI, including superior soft tissue quality and radiation safety, direct extraction of bone geometry from MRI remains challenging due to the low contrast of bone boundaries, especially in complex joints such as the ankle. Recently, another imaging approach known as AI generated CT has emerged, in which CT-like images are generated from MRI data using artificial intelligence techniques. The aim of this study was to evaluate the accuracy and substitution capability of bone models obtained from a specific MRI sequence and AI generated CT, in comparison with reference CT scans for ankle joint stress analysis.
Materials and Methods: For this purpose, CT scans, MRI data, and AI generated CT images were acquired from three healthy ankle joint samples. A total of six three-dimensional models were constructed based on these imaging modalities. Bone segmentation was performed using Mimics software, and the resulting three-dimensional models were further processed in 3-matic for standardized positioning, cartilage generation, minor surface smoothing, and mesh generation. Finite element modeling and mechanical simulations were then conducted in the Abaqus software environment. Contact stress distribution and von Mises stresses in the talus bone and cartilage were analyzed and compared across the different imaging-based models.
Results: The findings demonstrated that AI generated CT models exhibited good agreement with reference CT scans in terms of maximum stress values and stress distribution patterns, with a maximum difference of 5% in contact stress. In contrast, bone MRI-based models showed noticeable deviations in stress distribution patterns, as well as differences of 12% in maximum contact stress and 45% in von Mises stress.
Conclusion: These results indicate that, under current conditions, direct modeling from bone MRI lacks sufficient accuracy for mechanical stress analysis. Nevertheless, generating AI generated CT images from bone MRI using artificial intelligence methods provides a promising pathway for replacing CT scans in future biomechanical modeling of the ankle joint.
کلیدواژهها [English]
)