Evaluation of the Use of a Tablet Computer with a High-Resolution Display for Interpreting Emergency CT Scans

Purpose: Evaluation of the potential usability of an iPad 3 with a high-resolution display in CT emergency diagnosis compared to a 3D PACS workstation. Materials and Methods: 3 readers used a 5-point Likert scale to evaluate 40 CCT scans and 40 CTPA scans to determine the detectability of early signs of infarction in CCT or segmental and subsegmental pulmonary embolisms in CT angiography of the pulmonary arteries (CTPA) on the iPad 3 (Apple Inc., USA) using an application for image viewing (Visage Ease, Visage Imaging GmbH, Berlin) and on a 3D PACS workstation (Visage 7.1, Visage Imaging, Berlin) using a certified monitor for image viewing. The results were compared using the Wilcoxon rank sum test, Spearman's correlation coefficient, and a kappa statistic. Results: There was no significant difference in the median evaluations for the readings of both the CCT scans and the CTPA scans on the iPad 3 and on the workstation (p > 0.05) for all three readers. The mean Spearman's correlation coefficient for CCT and CTPA was 0.46 (+/- 0.2) and 0.69 (+/- 0.16), respectively, for the comparison iPad/PACS, 0.41 (+/- 0.16) and 0.68 (+/- 0.06), respectively, for the interobserver agreement on the iPad, and 0.35 (+/- 0.05) and 0.68 (+/- 0.10), respectively, for the interobserver agreement on the PACS. Mean kappa values for CCT of 0.52 (+/- 0.17) for the comparison iPad/PACS and 0.33 (+/- 0.16) and 0.32 (+/- 0.16), respectively, for the interobserver agreement on the iPad and the PACS were achieved. For CTPA average kappa values of 0.67 (+/- 0.19) were calculated for the comparison iPad/ PACS and 0.69 (+/- 0.08) and 0.60 (+/- 0.14), respectively, for the interobserver concordance on the iPad 3 and the PACS. All differences were not statistically significant (p > 0.05). Conclusion: The variability of the interpretation of typical emergency scans on an iPad 3 with a high-resolution display and on a 3D PACS workstation does not differ from the interobserver variability.