Potentialities and limitations of a database constructing three-dimensional virtual bone models.

ABSTRACT

BACKGROUND: 3D bone reconstructions performed during general clinical practice are of limited use for preclinical research, education, and training purposes. For this reason, we are constructing a database of human 3D virtual bone models compiled from computer tomography (CT) scans. MATERIALS AND METHODS: CT data sets were post-processed using Amira(®) 5.2 software. In each cut, bone structures were isolated using semiautomatic labeling program codes. The software then generated extremely precise 3D bone models in STL format (standard triangulated language). These bone models offer a sustainable source of information for morphologic studies and investigations of biomechanical bony characteristics in complex anatomic regions. Regarding educational value and student acceptance models were introduced during bedside teaching and evaluated by medical students. RESULTS: The current database is comprised of 131 pelvises and 120 femurs (ø 60 years, ø 172 cm, ø 76 kg), and is continuously growing. To date, 3D morphometric analyses of the posterior ring and the acetabulum have been successfully completed. Eighty students (96 %) evaluated instruction with virtual 3D bone models as "good" or "very good". The majority of students want to increase learning with virtual bone models covering various regions and diseases. CONCLUSION: With consistent and steadily increasing case numbers, the database offers a sustainable alternative to human cadaver work for practical investigations. In addition, it offers a platform for education and training.