Intraoperative Transvaginal Ultrasonographic Evaluation for Placenta Accreta Spectrum in Placenta Previa: a retrospective observational study

Background To assess the utility of intraoperative transvaginal ultrasonography (TVUS) in evaluating the duration of blood flow in cervical varicosities as a real-time diagnostic marker for placenta accreta spectrum (PAS) in placenta praevia. Methods This single-centre historical cohort study included patients with placenta praevia who underwent caesarean section with intraoperative TVUS. The primary outcome was blood flow persistence time, defined as the interval from fetal delivery to disappearance of low-velocity (≤ 2.0 cm/s) colour Doppler signals in cervical varicosities—dilated venous structures in the cervical stroma commonly observed in abnormal placentation. TVUS was performed in the sagittal plane including the internal cervical os. Flow persistence time was retrospectively measured from surgical video recordings and was independently assessed by two obstetricians. Receiver operating characteristic curve analysis was performed to evaluate the diagnostic performance of flow persistence time for predicting PAS. The area under ROC curve (AUC) was calculated to assess discriminative ability. Optimal cutoff values were determined using sensitivity, specificity, and predictive values. Multivariate regression identified independent predictors of PAS, and model performance was evaluated using AUC. Results Among the 52 included cases, PAS was diagnosed in 20 (38.5%). Blood flow persistence time was longer in PAS cases (20.20 ± 14.31 min) than in non-PAS cases (5.69 ± 2.46 min). A cutoff of 7 min yielded a sensitivity of 1.00 and specificity of 0.72 for PAS diagnosis (AUC = 0.94), reliably ruling out PAS when flow ceased within this time. Conversely, a duration of ≥14 min had a positive predictive value of 1.00, strongly indicating the presence of PAS. Multivariable analysis confirmed that flow duration was an independent predictor of PAS (AUC = 0.95). For predicting hysterectomy, a threshold of ≥14 min achieved a specificity of 0.96 and a negative predictive value of 1.00. Conclusions Intraoperative measurement of cervical varicosity blood flow duration via TVUS offers a reproducible, real-time physiological marker for PAS. A flow duration under 7 min reliably excludes PAS, while durations exceeding 14 min strongly suggest its presence. Incorporating this marker into intraoperative decision-making may enhance preparedness and maternal safety during caesarean delivery for placenta praevia.