Reproducibility of Detecting Silent Cerebral Infarcts in Pediatric Sickle Cell Anemia

Detecting silent cerebral infarcts on magnetic resonance images (MRIs) in children with sickle cell anemia is challenging, yet reproducibility of readings has not been examined in this population. We evaluated consensus rating, inter-, and intra-grader agreement associated with detecting silent cerebral infarct on screening MRI in the Silent Infarct Transfusion Trial. Three neuroradiologists provided consensus decisions for 1073 MRIs. A random sample of 53 scans was reanalyzed in blinded fashion. Agreement between first and second consensus ratings was substantial (κ = 0.70, P < .0001), as was overall intergrader agreement (κ = 0.76, P < .0001). In the test-retest sample, intragrader agreement ranged from κ of 0.57 to 0.76. Consensus decisions were more concordant when MRIs contained more than one larger lesions. Routine use of MRI to screen for silent cerebral infarcts in the research setting is reproducible in sickle cell anemia and agreement among neuroradiologists is sufficient.