Computer-Generated 3D Models in Veterinary Student Education: Improving Radiographic Interpretation of Canine Congenital Heart Disease

Abstract Thoracic radiographic interpretation of congenital heart disease (CHD) is challenging for veterinary students, particularly when translating three-dimensional (3D) anatomic structures into two-dimensional (2D) images. This study evaluated the effectiveness of virtual 3D models in enhancing student interpretation of thoracic radiographs. We hypothesized that 3D models will improve students' spatial and anatomical understanding of healthy and CHD radiographs compared with traditional 2D methods. Three client-owned dogs diagnosed with patent ductus arteriosus, pulmonary valve stenosis, and a healthy control, were enrolled. Computed tomography angiography datasets were used to create 3D cardiac models using image segmentation software. Virtual models were overlaid onto each corresponding dog's thoracic radiographs. A workshop was conducted with third- and fourth-year veterinary students, divided into two interactive learning groups; 2D traditional methods versus virtual 3D models with radiographic overlays. Pre- and post-workshop multiple-choice tests and Likert scale surveys assessed the objective and subjective radiographic interpretation, and CHD understanding. Thirty students participated in the workshop. The mean overall pre-test score was 67.9, and the mean post-test score was 87.5 ( p < .001). The mean percentage increase in pre- to post-workshop test scores was 16.8 and 22.1 in the 2D and 3D groups, respectively; however, the difference in score improvement between the groups was not significant ( p = .42). Feedback was generally positive, with participants preferring 3D over 2D resources. Both traditional and 3D methods improved students' ability to interpret thoracic radiographs in dogs. Participants preferred 3D resources, highlighting their potential as a valuable adjunct tool in veterinary education.