Observer reliability in grading nephrocalcinosis on ultrasound examinations in children

Background. Nephrocalcinosis is often associated with a variety of hypercalcemic conditions. Diagnostic ultrasound is often used for assessing nephrocalcinosis in children, but its reliability has not been proven. Objective. To determine the reliability of expert interpretation of sonographic films with a grading scale of severity for nephrocalcinosis. Materials and methods. Fifty-eight ultrasonographic films of 30 children with Williams syndrome and other conditions know to be associated with nephrocalcinosis were assessed. We used a blinded randomized design to assess intra- and interobserver reliability. Results. Grades I, II, and III nephrocalcinosis were noted in 13%, 19%, and 27% of the examinations, respectively. The weighted kappa coefficient was 0.80 (standard error 0.12; 95% confidence interval 0.68-0.92) for intraobserver agreement and 0.76 (standard error 0.13; 95% confidence interval 0.63 to 0.89) for interobserver agreement. Reliability in assessing change from one examination to the next, with independently graded films, was fair with an unweighted kappa coefficient of 0.68 (95% confidence interval 0.38-0.96) and 0.51 (95% confidence interval 0.21-0.80) for intra- and interobserver reliability, respectively. Conclusion. The severity of nephrocalcinosis can be reliably interpreted with an ultrasonography grading scale.