Comparison of the lower uterine segment in pregnant women with and without previous cesarean section in 3T MRI

BackgroundPrenatal risk stratification of women with previous cesarean section (CS) by ultrasound thickness measurement of the lower uterine segment (LUS) is challenging. There is a wide range of proposed cutoff values and a valuable algorithm for selection before birth is not available. Using 3T magnetic resonance imaging (MRI), we aimed to identify possible shortcomings of the current protocols used for birth selection after CS. Therefore, we evaluated anatomic and morphologic differences of the LUS and its thickness in patients with CS and those without. Possible impact factors on LUS thickness were studied.MethodsWe retrospectively analyzed 3T MRI scans of 164 pregnant women in their second or third trimester, with (patient group, n=60) and without previous CS (control group, n=104). Sagittal T2-weighted images were studied. Normal findings of the LUS in MRI, reliability of MRI measurements, as well as factors influencing LUS thickness were assessed. MRI findings were compared to intraoperative findings.ResultsMRI provided good intra- (ICC 0.872) and fair inter-rater reliability (ICC 0.643). The relationship of the LUS and the cesarean scar to the surrounding anatomical structures and also its morphology varied strongly in patients and controls. Scar identification was possible in only 9/60 (15.0%) patients. The LUS was thinner in patients (1.90.7mm) than in controls (2.7 +/- 1.3mm). An LUS thinning up to 1mm was observed in 23% of women without a previous CS and in 34% of women with normal intraoperative findings. Suspicion of a uterine dehiscence (LUS thickness<1mm) was only found in the patient group (5/59 (8.5%)) and was intraoperatively confirmed. In controls, LUS thickness was influenced by fetal weight, gestational age and amniotic fluid amounts.Conclusion Variability in anatomy, thickness and morphology seem to limit common prenatal LUS imaging diagnostics. Therefore, we consider that diagnostic protocols must be re-evaluated and imaging should be adjusted to the individual patient conditions. Due to its independency of ultrasound limitations, an additional MRI might be useful for altered anatomy and impaired ultrasound conditions. An LUS thinning up to 1mm might be a normal finding and should be further investigated as reference value.