Objective
To evaluate the diagnostic accuracy of ultrasound and in the diagnosis of Placenta accreta spectrum (PAS).
Data sources
Screening of MEDLINE, CENTRAL, other bases from inception to February 2022 using the keywords related to placenta accreta, increta, percreta, morbidly adherent placenta, and preoperative ultrasound diagnosis.
Study eligibility criteria
All available studies- whether were prospective or retrospective- including cohort, case control and cross sectional that involved prenatal diagnosis of PAS using 2D or 3D ultrasound with subsequent pathological confirmation postnatal were included. Fifty-four studies included 5307 women fulfilled the inclusion criteria, PAS was confirmed in 2025 of them.
Study appraisal and synthesis methods
Extracted data included settings of the study, study type, sample size, participants characteristics and their inclusion and exclusion criteria, Type and site of placenta previa, Type and timing of imaging technique (2D, and 3D), severity of PAS, sensitivity and specificity of individual ultrasound criteria and overall sensitivity and specificity.
Results
The overall sensitivity was 0.8703, specificity was 0.8634 with -0.2348 negative correlation between them. The estimate of Odd ratio, negative likelihood ratio and positive likelihood ratio were 34.225, 0.155 and 4.990 respectively.
The overall estimates of loss of retroplacental clear zone sensitivity and specificity were 0.820 and 0.898 respectively with 0.129 negative correlation. The overall estimates of myometrial thinning, loss of retroplacental clear zone, the presence of bridging vessels, placental lacunae, bladder wall interruption, exophytic mass, and uterovesical hypervascularity sensitivities were 0.763, 0.780, 0.659, 0.785, 0.455, 0.218 and 0.513 while specificities were 0.890, 0.884, 0.928, 0.809, 0.975, 0.865 and 0.994 respectively.
Conclusions
The accuracy of ultrasound in diagnosis of PAS among women with low lying or placenta previa with previous cesarean section scars is high and recommended in all suspected cases.
Trial registration
Number CRD42021267501.