Development of a new generation of high-resolution anatomical models for medical device evaluation: the Virtual Population 3.0

The Virtual Family computational whole-body anatomical human models were originally developed for electromagnetic (EM) exposure evaluations, in particular to study how absorption of radiofrequency radiation from external sources depends on anatomy. However, the models immediately garnered much broader interest and are now applied by over 300 research groups, many from medical applications research fields. In a first step, the Virtual Family was expanded to the Virtual Population to provide considerably broader population coverage with the inclusion of models of both sexes ranging in age from 5 to 84 years old. Although these models have proven to be invaluable for EM dosimetry, it became evident that significantly enhanced models are needed for reliable effectiveness and safety evaluations of diagnostic and therapeutic applications, including medical implants safety. This paper describes the research and development performed to obtain anatomical models that meet the requirements necessary for medical implant safety assessment applications. These include implementation of quality control procedures, re-segmentation at higher resolution, more-consistent tissue assignments, enhanced surface processing and numerous anatomical refinements. Several tools were developed to enhance the functionality of the models, including discretization tools, posing tools to expand the posture space covered, and multiple morphing tools, e.g., to develop pathological models or variations of existing ones. A comprehensive tissue properties database was compiled to complement the library of models. The results are a set of anatomically independent, accurate, and detailed models with smooth, yet feature-rich and topologically conforming surfaces. The models are therefore suited for the creation of unstructured meshes, and the possible applications of the models are extended to a wider range of solvers and physics. The impact of these improvements is shown for the MRI exposure of an adult woman with an orthopedic spinal implant. Future developments include the functionalization of the models for specific physical and physiological modeling tasks.