Interobserver reliability and accuracy of cranial ultrasound scanning interpretation in premature infants.

Abstract

To assess interobserver reliability between 2 central readers of cranial ultrasound scanning (CUS) and accuracy of local, compared with central, interpretations. The study was a retrospective analysis of CUS data from the National Institute of Child Health and Human Development (NICHD) trial of inhaled nitric oxide for premature infants. Interobserver reliability of 2 central readers was assessed with kappa or weighted kappa. Accuracy of local, compared with central, interpretations was assessed by using sensitivity and specificity. CUS from 326 infants had both central reader and local interpretations. Central reader agreement for grade 3/4 intraventricular hemorrhage (IVH), grade 3/4 IVH or periventricular leukomalacia (PVL), grade of IVH, and degree of ventriculomegaly was very good (kappa = 0.84, 0.81, 0.79, and 0.75, respectively). Agreement was poor for lower grade IVH and for PVL alone. Local interpretations were highly accurate for grade 3/4 IVH or PVL (sensitivity, 87%-90%; specificity, 92%-93%), but sensitivity was poor-to-fair for grade 1/2 IVH (48%-68%) and PVL (20%-44%). Our findings demonstrate reliability and accuracy of highly unfavorable CUS findings, but suggest caution when interpreting mild to moderate IVH or white matter injury.