ABSTRACT
PURPOSE:
We aimed to verify whether combination of specific signs improves magnetic resonance imaging (MRI) accuracy in morbidly adherent placenta (MAP).
METHODS:
MRI findings for MAP were retrospectively evaluated in 27 women. Histopathology was the reference standard, showing MAP in eight of 27 cases. Specificity, sensitivity, positive predictive value, and negative predictive value were calculated for all MRI signs. Two skilled radiologists analyzed MRI findings, resolving discrepancies by consensus, using three alternative diagnostic criteria during three consecutive sections. First criterion: at least one of reported MRI signs indicates MAP and the absence of any sign is normal; second criterion: at least one statistically significant sign indicates MAP and no sign or nonsignificant sign is normal; third criterion: at least two statistically significant signs indicate MAP and no sign, nonsignificant sign, or only one significant sign is normal.
RESULTS:
Using the first criterion yielded an unacceptable rate of false positive results (78.9%). Using the second criterion there were less false positive results (31.5%), and diagnostic accuracy of the second criterion was significantly higher than the first; the third criterion correctly classified 100% of cases.
CONCLUSION:
Only specific MRI signs can correctly predict MAP at histopathology, particularly when multiple (at least two) specific signs are observed together.