Accuracy of mobile devices for acute stroke head CT interpretation among Neurologists in training. (P1.083)

Abstract

Objective: To test accuracy of mobile devices (iPad and iPhone) in the interpretation of Head CT compared to standard Picture Archiving and Communication System (PACs) Radiology station among Neurologists in training. Background: Advances in technology has allowed physicians access to radiological images remotely facilitating their rapid interpretation. The diagnostic performance of this approach for acute stroke head CT scan interpretation among neurologists in training has not been evaluated. Design/ Methods: 9 readers (3 vascular neurology fellows, 3 PGY-4 and 3 PGY-3 neurology residents) independently interpreted 20 preselected acute stroke CTs. Images were viewed on iPad 2(1024x768 pixels) and iPhone 4 (960x640 pixels) using (ResMD® software (, Calgary Scientific, Calgary, CanadaA), and Radiology PACS station (gold standard). Readers recorded CT findings iIncluding Acute Ischemic Signs (AIS), Non-Acute Ischemic Signs (NAIS) and Hyperdense MCA (HMCA). A generalized, mixed linear model was constructed. Model-generated estimates of sensitivity & specificity (with 95[percnt] confidence intervals) were calculated. Results: For AIS, iPad had sensitivity of 83[percnt] (95[percnt]CI 69-92) and specificity of 60[percnt], (95[percnt]CI 41-76) compared to iPhone [sensitivity 79[percnt] (95[percnt]CI 60-90); specificity 58[percnt] (95[percnt]CI 36-77)]. For NAIS iPad sensitivity was 63[percnt] (95[percnt]CI 24-90) and specificity 77[percnt] (95[percnt]CI 37-95). iPhone sensitivity was 75[percnt] (95[percnt]CI 53-89) and specificity was 73[percnt] (95[percnt]CI 48-89). HMCA identification sensitivity on the iPad was 58[percnt] (95[percnt]CI 36-77) and specificity was 93[percnt] (95[percnt]CI 86-96) and iPhone sensitivity was 52[percnt] (95[percnt]CI 29-74) and specificity 90[percnt] (95[percnt]CI 81-95). Conclusions: Among neurologists in training (residents and stroke fellows) mobile devices had good to excellent sensitivity for identification of AIS. Mobile devices also has good to excellent specificity for the identification of the HMCA. Larger studies testing newer generation mobile devices and a greater breadth of findings are needed for comparison with traditional PACs interpretation prior to routine clinical practice.

Ähnliche Arbeiten