Retina and optic nerve diffusion restriction in acute central retinal artery occlusion
Daniel Gomez Ramos1, Ehab Harahsheh2, Nan Zhang3, Parth Parikh1, Omer Elshaigi1, Tanya Rath1, Joseph Hoxworth1, Oana Dumitrascu1
1Mayo Clinic, 2Mayo Clinic College of Medicine, 3Mayo Clinic Arizona
Objective:
To determine whether diffusion restriction (DR) of the retina and/or optic nerve (ON) in acute central retinal artery occlusion (CRAO) can be accurately identified on brain diffusion-weighted magnetic resonance imaging (DW-MRI) by junior versus senior neurology residents and vascular neurologists.
Background:
Retina and/or ON DR in CRAO was described with various accuracies and fair inter-rater reliability even between experienced neuroradiologists. The identification of these subtle DW-MRI findings by the forefront neurology residents or vascular neurologists was not previously investigated.
Design/Methods:
Retrospective cohort study of consecutive acute CRAO patients that had stroke-protocol DW-MRI at our tertiary academic center (2013-2020). Two neurology residents (junior and senior), vascular neurologist, and two neuroradiologists blinded to the CRAO site independently reviewed the DW-MRI for retina and/or ON DR. Their inter-rater reliability was calculated using kappa statistics. The neuroradiologists’ consensus agreement was the reference for the non-radiologists.
Results:
We included 128 patients (mean age 68 years; 55% right eye, median time from CRAO to DW-MRI 2 days (IQR 1-4). 28/128 (21.8%) had a discrepancy between the neuroradiologists. After neuroradiologist consensus, DR was correctly identified in the retina or ON in 51/128 (39.8%), retina alone 35/128 (27.3%), ON alone 31/128 (24.2%), both retina and ON 15/128 (11.7%). Upon comparing the neurology trainees and vascular neurologist, the positive agreement was low (22.2% to 43.9%) but the negative agreement was high (88.0% to 92.9%). The senior resident had the highest observer agreement. Overall substantial agreement (P-B adjusted Kappa > 0.6) between the readings, excepting for moderate (P-B adjusted kappa= 0.59) for ON DR for the junior resident.
Conclusions:
Incongruities in identifying retina and/or ON DR in CRAO are encountered by various specialists. When compared to experienced neuroradiologists, neurology trainees and attending have substantial agreement on the negative findings but low agreement for the positive findings.