Placenta accreta spectrum disorders using standardised magnetic resonance imaging (MRI) descriptors; interobserver...

Accurate antenatal diagnosis of placenta accreta spectrum (PAS) disorders is essential for improving maternal outcomes. To address variability in magnetic resonance imaging (MRI) interpretation for suspected PAS, the International Society for Placenta Accreta Spectrum (IS-PAS) proposed nine standardised PAS-MRI descriptors. This study evaluates their interobserver agreement and diagnostic value.

A retrospective study was conducted using MRI scans of pregnancies suspected of PAS at Maastricht University Medical Center (2010-2024). Two radiologists independently assessed the MRI scans for IS-PAS descriptors, and interobserver agreement and diagnostic performance of each descriptor were evaluated. Ultrasound (US) reports leading to MRI referral were analysed for US PAS descriptors. Histopathological or intraoperative findings served as the reference standard.

Among 32 cases, 53% had no PAS, 37% had placenta accreta/increta, and 10% had placenta percreta. Interobserver agreement was highest for MRI descriptors 'bladder wall interruption' (κ = 0.52) and 'focal exophytic mass' (κ = 0.46), while 'heterogeneous placenta' (κ = 0.12) and 'dark intraplacental bands' (κ = -0.09) showed the lowest agreement. Sensitivity was highest for MRI descriptors 'loss of retroplacental dark zone' (100%) and 'myometrial thinning' (93%), while specificity was highest for 'bladder wall interruption' (94%) and 'focal exophytic mass' (100%). There were no significant differences in sensitivity and specificity between US and MRI for diagnosing PAS.

IS-PAS MRI descriptors show suboptimal interobserver agreement and limited diagnostic value. Prenatal diagnosis of placenta accreta/increta remains especially challenging. Future studies should focus on validating current MRI descriptors and investigate the use of newer imaging modalities.