Introducing an efficient model for the prediction of placenta accreta spectrum using the MCP regression approach...

For the first time, we aimed to introduce a model for prediction of placenta accreta spectrum (PAS), using existing sonography indices.

Women with a history of Cesarean sections were included. Participants were categorized "high risk" for PAS if the placenta was previa or low-lying. Sonography indices including abnormal placental lacuna, loss of clear zone, bladder wall interruption, myometrial thinning, placental bulging, exophytic mass, utero-vesical hypervascularity, subplacental hypervascularity, existence of bridging vessels, and lacunar flow, were registered. To investigate simultaneous effects of 15 variables on PAS, Minimax Concave Penalty (MCP) was used.

Among 259 participants, 74 (28.5%) were high risk and 43 individuals had PASs. All sonography indices were higher among patient with PAS (p < 0.001) in the high risk group. Our model showed that utero-vesical hypervascularity, bladder interruption and new lacunae have significant contribution in PAS. Optimal cut off point was p = 0.51 in ROC analysis. Probability of PAS for women with lacunae was between 96 and 100% and probability of PAS for women without lacunae was between 0 to 7%, therefore accuracy of the proposed model was equal to 100%.

Using the introduced model based on three factors of abnormal lacuna structures (grades 2 and 3), bladder wall interruption and utero-vesical vascularity, 100% of all cases of PASs are diagnosable. If supported by future studies our model eliminates the need for other imaging assessments for diagnosis of invasive placentation among high risk women with previous history of Cesarean sections.