Anatomical Model Generator Based on Published Clinical Data on Cardiovascular Anatomy

Anatomical variability in patient populations has presented a challenge to the verification and validation (V&V) process for medical devices. This lack of conformity is also present in the anatomical models used for device design and testing, leading to a limited ability for comparative analysis among study results. Other industries have adopted standardized models for benchmark analysis and the FDA has begun to seek conformity in the analytical tools used for flow analysis. The ability to generate anatomical models from a central source that prescribes taxonomy, much like a digital object identifier (DOI) number for a published document, would enable researchers in the medical device field to homogenize the model development process. It is possible to reconstruct these anatomies as parametric models driven by clinical measurement data. This solution presents a transformative approach to anatomical modeling by enabling large scale simulations, allowing comparative analysis, and lowering the cost barrier for smaller research groups to source these models. Investigatory models of the aorta, left ventricle and left atrium were constructed to design the appropriate parametric framework for reconstruction. The resulting models are presented as user selectable classes of patients, which are referencing lookup tables of the measurement data. Validation of this model generator method has shown acceptable results using volume subtraction analysis with models generated manually. This approach delivers a solution for anatomical model generation and traceability of test results derived from these geometries. It answers significant scientific and regulatory concerns regarding comparative analysis and creates a more egalitarian approach to developing anatomical models for medical device design.