Abstract W P40: Assessing Criteria for Intra-Arterial Therapy: Inter-Rater Reliability of Clinical-Diffusion Mismatch Among Stroke Clinicians
Background: Optimal patient selection strategy for intra-arterial therapy (IAT) in acute ischemic stroke remains uncertain. Clinical-diffusion mismatch (CDM) and perfusion-diffusion mismatch (PDM) on MRI are two selection methods used in practice. PDM can be radiologically defined and its analysis automated but this requires additional time and contrast administration. CDM is less technically challenging, however, there is variation in its definition and its inter-rater reliability has not been reported.
Hypothesis: CDM has a higher inter-rater variability than PDM among vascular neurologists.
Methods: Fellowship-trained vascular neurologists from our high-volume stroke center were surveyed on ten cases. For each case, they were first shown diffusion-weighted and apparent diffusion coefficient imaging and NIHSS score components and were asked if they felt a CDM existed. Based on the CDM alone, they were next asked if they would pursue IAT in the presence of a persistent large vessel occlusion. A perfusion image was then shown and the participants were asked if they felt a PDM existed. A final slide was shown giving the age, co-morbidities and pre-morbid disability, and location of arterial occlusion, after which they were again asked if they would pursue IAT. Kappa statistics were used to assess agreement.
Results: Ten vascular neurologists participated in this study. The average inter-rater kappa for CDM was 0.35, 95% CI (0.28, 0.43) across the 10 cases. The average inter-rater kappa to take a patient for IAT based on CDM alone was 0.37, 95% CI (0.28, 0.46). The average inter-rater kappa for PDM was 0.61, 95% CI (0.53, 0.69). The average inter-rater kappa for final decision to take the patient to IAT was 0.30, 95% CI(0.22,0.37). The rate of IAT over the 10 cases varied from 10% to 70% among the vascular neurologists.
Conclusions: This study suggests that vascular neurologists have a marginal inter-rater reliability for determining CDM. Inter-rater reliability for PDM was good although this did not have an effect on the reliability of final decision to take a patient for IAT. Further multi-center studies are needed to assess the variability of decision making for IA therapy.
Author Disclosures: P. George: None. I. Katzan: None. J. Gebel: None. K. Uchino: None.
- © 2014 by American Heart Association, Inc.