Diagnostic performance of android-based handheld device using endeavor mobile application in interpretation of a traumatic non- contrasted computed tomography (CT) brain

Title: Diagnostic Performance of Android-Based Handheld Device Using Endeavor Mobile Application in Interpretation of a Traumatic Non-Contrasted Computed Tomography (CT) Brain Purpose: To determine the image viewing quality of the handheld device of Android premium devices and investigate the usage of handheld device in interpretation of CT images of trauma cases. Materials and methods: Using Samsung Galaxy Note 10.1, we installed the AAPM TG-18 QC test pattern and recruited 30 candidates to do a subjective review of the QC test pattern on both handheld and workstation monitors. As for the investigation of using handheld device to interpret CT images of trauma cases, we recruited 2 observers, consisting of a radiologist and a final year resident to review 180 cases of CT brain.Their findings would then be cross-referenced to a result obtained by a consultant radiologist using workstation monitor. Kappa test were used to calculate the Interobserver agreement. Results: There was 100% reproducibility of the same level of luminance patches, grayscale continuation and spatial resolution in the handheld device when compared to the workstation monitor. Other components investigated produced similar results when the candidates were allowed to zoom in and change the window settings. The sensitivity of detecting lesions on the images using handheld devices ranging from 50.0% to 84.6%. The negative predictive values were generally high indicating that the handheld device has high accuracy to determine absence of certain lesion. There was substantial agreement regarding the findings of both observer compared to gold standard test on workstation monitor. Conclusions: Current generation of handheld devices has had viewing quality at par with workstation monitor and it is a reliable tool in reviewing CT for trauma cases