Magnetic resonance imaging is often misleading when used as an adjunct to ultrasound in the management of placenta accreta spectrum disorders.

BACKGROUND: Magnetic resonance imaging is reported to have good sensitivity and specificity in the diagnosis of placenta accreta spectrum disorders, and is often used as an adjunct to ultrasound. But the additional utility of obtaining magnetic resonance imaging to assist in the clinical management of patients with placenta accreta spectrum disorders, above and beyond the information provided by ultrasound, is unknown.

OBJECTIVE: We aimed to determine whether magnetic resonance imaging provides data that may inform clinical management by changing the sonographic diagnosis of placenta accreta spectrum disorders.

STUDY DESIGN: In all, 78 patients with sonographic evidence or clinical suspicion of placenta accreta spectrum underwent magnetic resonance imaging of the abdomen and pelvis in orthogonal planes through the uterus utilizing T1- and T2-weighted imaging sequences at the University of Utah and the University of Colorado from 1997 through 2017. The magnetic resonance imaging was interpreted by radiologists with expertise in diagnosis of placenta accreta spectrum who had knowledge of the sonographic interpretation and clinical risk factors for placenta accreta spectrum disorders. The primary outcome was a change in diagnosis from sonographic interpretation that could alter clinical management, which was defined a priori. Diagnostic accuracy was verified by surgical and histopathologic diagnosis at the time of delivery.

RESULTS: A change in diagnosis that could potentially alter clinical management occurred in 28 (36%) cases. Magnetic resonance imaging correctly changed the diagnosis in 15 (19%), and correctly confirmed the diagnosis in 34 (44%), but resulted in an incorrect change in diagnosis in 13 (17%), and an incorrect confirmation of ultrasound diagnosis in 15 (21%). Magnetic resonance imaging was not more likely to change a diagnosis in the 24 cases of posterior and lateral placental location compared to anterior location (33% vs 37%, P = .84). Magnetic resonance imaging resulted in overdiagnosis in 23% and in underdiagnosis in 14% of all cases. When ultrasound suspected severe disease (percreta) in 14 cases, magnetic resonance imaging changed the diagnosis in only 2 cases. Lastly, the proportion of accurate diagnosis with magnetic resonance imaging did not improve over time (61-65%, P = .96 for trend) despite increasing volume and increasing numbers of changed diagnoses.

CONCLUSION: Magnetic resonance imaging resulted in a change in diagnosis that could alter clinical management of placenta accreta spectrum disorders in more than one third of cases, but when changed, the diagnosis was often incorrect. Given its high cost and limited clinical value, magnetic resonance imaging should not be used routinely as an adjunct to ultrasound in the diagnosis of placenta accreta spectrum until evidence for utility is clearly demonstrated by more definitive prospective studies.