The REliability of MObile TEchnologies for Acute Stroke Neuroimaging Data Interpretation: The REMOTE Study (S5.001)

Abstract

OBJECTIVE: To assess agreement in interpreting acute stroke head CT scan among readers from different specialties on mobile devices (iPad, iPhone), and on a standard radiology Picture Archiving and Communication System (PACs) workstation. BACKGROUND: Advances in technology allow physicians access to radiological images remotely, although the diagnostic performance of this approach for acute stroke head CT scan interpretation among different specialties is unknown. DESIGN/METHODS: 20 selected acute stroke head CTs were independently interpreted, in an intent-to-diagnose manner, by 16 readers from different specialties - neurologists, emergency medicine physicians and radiologists - on three modalities, iPad and iPhone [using ResolutionMD software (ResMD®, Calgary Scientific)] and PACs workstation. CT findings included: normal CT; any acute ischemic sign; any intracerebral hemorrhage (ICH); and any middle cerebral artery (MCA) hyperdense sign. Kappa (ĸ) statistics were used to assess the proportion of agreement beyond chance. RESULTS: Agreement for normal CT across all readers was best on the iPad (ĸ=0.29 vs. 0.19 for both iPhone and PACS); highest for neurologists (ĸ=0.42) and radiologists (ĸ=0.42) compared to emergency medicine physicians (ĸ=0.15). For any acute ischemic sign, the overall agreement was fair (PACs ĸ=0.38; iPhone ĸ=0.26; iPad ĸ=0.25). The overall agreement for presence or absence of ICH was fair (ĸ’s 0.34-0.40). Radiology reached excellent agreement in all three modalities (k’s 0.65-1.00) compared to neurologists (k’s 0.33-0.51). Overall agreement for the presence of hyperdense MCA sign was fair (k’s 0.33-.38), reaching moderate to good agreement for radiologists, which were consistently better than for neurologists (k’s 0.30-0.32). CONCLUSIONS: Among readers with different expertise, there is only fair agreement on identifying any head CT acute ischemic sign, ICH, and hyperdense MCA, independently from the modality used. However, among radiologists, good to excellent agreement in all modalities support that their remote head CT scan interpretation, using mobile devices, may have the potential to approach traditional PACs readings.

Ähnliche Arbeiten