Clinically relevant misinterpretation of computed tomography (CT) images via telemedicine is rare in a system that uses a structured assessment of images, say German researchers.
Neurologists in the Stroke Eastern Saxony Network (SOS-NET) use the Alberta Stroke Program Early CT Score (ASPECTS) to assess ischemic changes in a consistent manner in images sent for teleconsultation.
In an editorial accompanying the study in Neurology, Peter Müller-Barna (Städtisches Klinikum München GmbH, Germany) and Heinrich Audebert (Charité-Universitätsmedizin Berlin, Germany) stress the importance of a structured imaging assessment in the event of suspected stroke. "Because neurologists tend to look for intracranial lesions where they suspect the cause of clinical deficits, they may miss pathologies remote from the suspected areas."
The interobserver agreement for ASPECTS was "substantial," with a weighted kappa statistic of 0.62 between the nine neurologists who read 536 sets of images at the time of their transmission and two neuroradiologists who reassessed the images for the current study.
All the neurologists involved in the SOS-NET received certified training in stroke imaging interpretation, delivered by a neuroradiologist.
The neuroradiologists disagreed with the neurologists' interpretations for 43 (8%) patients. In most cases (35 patients) this was a difference of more than 1 point in ASPECTS, report study author Volker Puetz (Dresden University Stroke Center) and co-workers.
There were also missed or falsely diagnosed early ischemic changes in vascular territories not covered by ASPECTS in six patients, a missed subdural hematoma in one patient, and a missed intracranial tumor in a patient in whom the neurologist had correctly identified subarachnoid hemorrhage.
Nine of these discrepancies were considered clinically relevant. The neurologists missed extensive early ischemic changes in the middle cerebral artery territory of eight patients and recommended thrombolysis for all eight, one of whom then had a symptomatic intracranial hemorrhage (ICH). The case of the patient with the missed subdural hematoma was also considered clinically relevant, although they did not undergo thrombolysis because of other contraindications.
In all, clinically relevant discrepancies affected just 1.7% of the cohort, and the neurologists detected brain tumors and primary or traumatic ICH with at least 95% accuracy.
"These findings, however, must not lead to the conclusion that neurologists or stroke physicians of other disciplines can provide this high level of expertise without training and structured assessment of imaging data," warn Müller-Barna and Audebert.
They highlight the need for experienced stroke experts who are trained to interpret neuroimaging, if telestroke services are to fulfill their potential.
Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.