Teleradiology assessment of computerized tomographs online reliability study (TRACTORS) for acute stroke evaluation

Telemedicine has potential to increase the use of tissue plasminogen activator (t-PA) for ischemic stroke and is increasingly offered to provide stroke expertise to remote and underserved areas. The accuracy of stroke neurologists' use of telemedicine to read head computed tomography scans (CT) has not been demonstrated. We sought to determine the validity and reliability of neurologists' reading of acute stroke head CTs via teleradiology. The pilot study compared stroke neurologists' reading of CTs via teleradiology to gold standard readings of hard copies on a view box. Head CTs obtained during consecutive acute evaluations for thrombolytic therapy between February and July 2001 were used. For this study, each stroke neurologist was masked to all prior readings, clinical decisions regarding thrombolysis eligibility, and the reading of the other study neurologist. Each neurologist read one-half of the CT scans using teleradiology and the other one-half using a view box. We used the official reading by a neuroradiologist for comparison. The kappa statistic measured reliability. Validity was measured by sensitivity and specificity. Sixty scans were used for the analysis. The neurologists identified the same four hemorrhages using teleradiology that they and the neuroradiologist identified using the light board. Compared to the readings on a view box, the kappa statistic for eligibility for thrombolysis via teleradiology was 1.0. Sensitivity was 100%, 95% CI = (0.93, 1.0); specificity was 100%, 95% CI = (0.40, 0.98) in both comparisons. This pilot study provides encouraging preliminary evidence that neurologists with stroke expertise can determine radiological intravenous t-PA eligibility via teleradiology. Further analysis in a larger sample is necessary to validate these results.