Objective: To develop a software for virtual reality (VR) simulation of X-ray images based on perspective projections through a patient model derived from data from a CT examination and to evaluate the accuracy in the projection geometry obtained by the software.

Methods: A VR software was developed on a personal computer, with models of a patient, an X-ray machine and a detector. The model of the patient was derived from data from a CT examination of a dry skull. Simulated radiographic images of the patient model could be rendered as perspective projections based on the relative positions between the models. The projection geometry of the software was validated by developing an artificial CT data set containing high attenuation points as objects to be imaged. The accuracy in projection geometry was evaluated in a systematic way. The distances between two dots, representing the projected test points in the simulated radiographic images, were measured. They were compared with theoretical calculations of the corresponding distances using traditional mathematical tools.

Results: The difference between the simulated and calculated projected distances never exceeded 0.5 mm. The error in simulated projected distances was in most cases within 1%. No systematic errors were revealed.

Conclusion: The software, developed for personal computers, can produce simulated X-ray images with high geometric accuracy based on perspective projections through a CT data set. The software can be used for simulation of radiographic examinations.

Objectives: To investigate whether skill in the interpretation of three-dimensional (3D) information in radiographs utilizing the parallax phenomenon is associated with visual-spatial ability and whether development of this skill is related to visual-spatial ability.

Methods: Eighty-six individuals with a median age of 25 years participated in the study. It was organized into three parts: (1) assessment before training, (2) training in object depth localization utilizing parallax and (3) assessment after training. Before training, visual-spatial ability was assessed with a mental rotation test, MRT-A; skill in interpreting 3D information was assessed with two specifically designed proficiency tests: a radiography test, which assessed the ability to interpret 3D information in radiographs utilizing motion parallax and a principle test which assessed understanding of the principles of motion parallax. After training, skill in interpreting 3D information was reassessed. Improvement was defined as the difference between test scores after training and before training. Multiple linear regression was used to analyse the effect of student and training characteristics on proficiency test results and improvement.

Results: Radiography test results after training and improvement of radiography test results were significantly associated with MRT-A scores (P<0.001 and P=0.020, respectively). Principle test results were high before training and did not improve after training. The test results were associated with MRT-A both before (P=0.009) and after training (P=0.003).

Conclusions: Understanding of the parallax phenomenon is associated with visual-spatial ability. Development of the skill to interpret 3D information in radiographs utilizing parallax is facilitated for individuals with high visual-spatial ability.

Abstract Introduction: A radiology simulator has been developed. We tested the simulator with students in an oral radiology program for training interpretation of spatial relations in radiographs utilizing parallax. The aim of the study was to compare learning outcome regarding interpretative skill after training in the simulator vs. after conventional training.

Methods: Fifty-seven dental students voluntarily participated in a randomized experimental study. The participants' proficiency in interpretation of spatial information in radiographs and their visual-spatial ability was assessed. Proficiency was assessed by a test instrument designed by the authors and visual-spatial ability with the Mental Rotations Test, version A (MRT-A). Randomization to training group was based on pre-training proficiency test results. The experimental group trained in the simulator and the control group received conventional training. Training lasted for 90 minutes for both groups. Immediately after training a second proficiency test was performed.

Results: The proficiency test results were significantly higher after training for the experimental group (P <= 0.01), but not for the control group. Univariate variance analysis of difference in proficiency test score revealed a significant interaction effect (P = 0.03) between training group and MRT-A category; in the experimental group there was a stronger training effect among students with low level of MRT-A.

Conclusions: Training in the simulator improved skill in interpreting spatial information in radiographs when evaluated immediately after training. For individuals with low visual-spatial ability simulator based training seems to be more beneficial than conventional training.