The variability in the interpretation of prenatal diagnostic ultrasound

Objectives Although prenatal ultrasound is broadly used to detect abnormal fetuses, the variability in the interpretation of second-trimester prenatal ultrasound examinations is unknown. We sought to evaluate the consistency of the interpretation of prenatal ultrasound Examinations. Design Physicians who perform prenatal ultrasound and who participate in th e California Maternal Serum Expanded AFP program were asked to interpret se ries of rest ultrasound examinations. The series of cases Er,as selected to include a random sample of fetal structural abnormalities, ultrasound mark ers that have been associated with chromosomal abnormalities. and normal ca ses. Interobserver agreement was evaluated using a kappa statistic for each organ system. The sensitivity and false-positive rate were calculated for detecting specific anatomic abnormalities within each organ system. Results Of the 210 sonologists eligible for inclusion in the study, complet ed responses were received from 148 (70%). There was moderate to substantia l agreement between physicians in reporting the presence of fetal abnormali ties for all organ systems (kappa range 0.40-0.88, P < 0.001). The consiste ncy was highest for the central nervous system (CNS), neck. and face. Withi n each organ system, the consistency was similar for major structural abnor malities and ultrasound markers of chromosomal abnormalities. The sensitivi ty ranged from 62% (95% confidence interval (CI) 58-66%) for major renal ab normalities to 91% (95% CI 88-94%) for CNS abnormalities, with correspondin g false-positive rates of 7% (95% CI 6-9%) for renal abnormalities and 9% ( 95% CI 7-11%) for CNS abnormalities. For most organ systems, the sensitivit y for detecting ultrasound markers of chromosomal abnormalities was similar to the sensitivity for detecting structural abnormalities. Conclusion There is moderate to substantial agreement in the interpretation of second-trimester prenatal ultrasound examinations. Whether the identifi cation of specific ultrasound abnormalities and markers is overall benefici al to patients remains to be determined.