Development of a training program for the ultrasound screening of placenta accreta spectrum disorders.

OBJECTIVE: To evaluate the impact of a training program using a systematic protocol on ultrasound signs of placenta accreta spectrum (PAS). METHODS: Intra- and inter-observer variability rates and sensitivity were tested, before and after additional training, by two research fellows with a prior basic training in obstetric ultrasound using digitally recorded second-trimester ultrasound images from cases of anterior placenta previa with and without PAS. RESULTS: Fifty-two cases of anterior placenta previa with PAS (n=26) and without PAS (n=26) were included in the study. The highest level of inter-observer agreement for ultrasound signs was found for the absence of placental bulge and/or focal exophytic mass on gray-scale imaging and the absence of subplacental hypervascularity, bridging vessels and lacunar feeder vessels on color Doppler imaging. The level of inter-observer agreement increased from 39% before training to 40% after training; the numbers agreed as PAS by both trainees increased from four to 20. No cases were classified as inconclusive after training. There was a significant (P<0.001) change in sensitivity for both trainees after training. CONCLUSION: Additional training in detecting the ultrasound signs associated with PAS using a standardized protocol improves the diagnostic accuracy of operators with only a basic obstetric ultrasound training.

Interobserver agreement on standardized ultrasound and histopathologic signs for the prenatal diagnosis of placenta accreta spectrum disorders.

OBJECTIVE: To evaluate interobserver agreement in assessment of ultrasound signs and histopathologic findings associated with placenta accreta spectrum (PAS) disorders. METHODS: A retrospective study was conducted using data for patients prenatally diagnosed with PAS disorders at a UK hospital between January 31, 2012, and March 30, 2017. Ultrasound images (including gray-scale and color Doppler imaging [CDI] parameters) and histopathologic slides were reviewed by two observers; the level of agreement was calculated. RESULTS: Among 25 patients, 11 had placenta creta, 10 had placenta increta, and four had placenta percreta. Interobserver agreement for ultrasound imaging in the second and third trimesters and histopathologic diagnosis of PAS was rated as good-to-excellent. The highest level of interobserver agreement for ultrasound signs was found for loss of clear zone (100%) and substantial myometrial thinning (96%-100%) on gray-scale imaging, the presence of lacunar feeder vessels (100%) on two-dimensional CDI, and crossing vessels and lacunae (92%-95%) on three-dimensional CDI. CONCLUSION: Standardized ultrasound signs might prove useful for prenatal screening of women at risk of PAS disorders and should enable remote evaluation of images when PAS is suspected.