Abstract
OBJECTIVE: To estimate the level of interobserver agreement in the calculation of placenta accreta index (PAI) as well as to evaluate the accuracy of PAI in prediction of morbidly adherent placenta. MATERIALS AND METHODS: This was a prospective study where 45 pregnant women (from 28 to 37 weeks of gestational age) with at least one previous Caesarean section and ultrasound-proven placenta previa were included. A known and previously published scoring system, the PAI, was evaluated independently by two radiologists and the cases were followed for the delivery and histopathology outcome. The accuracy of the PAI and the level of interrater agreement was analysed using cross-table analysis, intraclass correlation efficient and Cohen's kappa as statistical variables. RESULTS: Adherent placenta was found in 15 patients accounting for 33% of cases. The PAI showed nearly 90% sensitivity, specificity and the predictive values. Interrater agreement in calculation of PAI by the two radiologists was perfect with an intraclass correlation efficient of 0.959. An easy-to-use morbid adherent placenta score was also predicted to simplify the results of PAI, which showed moderate agreement (κ = 0.746). CONCLUSIONS: The PAI can be helpful in stratifying the individual risk of placental invasion above the baseline risk. The PAI-derived, simplified scoring system called morbid adherent placenta score can be used as a simple tool to interpret and convey the results of PAI.