Prenatal detection of placenta accreta spectrum using a sonographic checklist

Helena C. Bartels,Jennifer M. Walsh,Stephen Carroll,Paul Downey,Donal J. O'Brien,Fionnuala M. McAuliffe,Clare C'Connor,Claire Thompson,Jennifer Donnelly,Donal J. Brennan,Siobhan M. Corcoran
DOI: https://doi.org/10.1111/aogs.14943
2024-10-04
Acta Obstetricia Et Gynecologica Scandinavica
Abstract:In a multi‐site tertiary center of prospectively obtained cases, a standardized sonographic checklist had a high sensitivity and moderate specificity for the prenatal detection of PAS in a high risk population. Across two trimesters, we found progression of features from the second to third trimester and improved performance of the checklist when applied in the third trimester. Combining features as in the checklist is important as no individual feature had a good performance for prenatal detection of PAS. Features evolve and performance metrics are improved in the third trimester compared to the second, highlighting the need of repeat scanning as gestation advances so delivery and surgical plans can be optimized. Introduction The European Working Group for Abnormally Invasive Placenta proposed a checklist of ultrasound features for the antenatal detection of placenta accreta spectrum (PAS). This study aims to assess the performance of the checklist in identifying histopathologically confirmed PAS cases in a cohort with a high pre‐test probability of PAS, and identify if particular features are associated with PAS. Material and Methods This is a prospective multi‐site cohort study conducted between 2018 and 2023. Consecutive patients who underwent ultrasound assessment for suspicion of PAS were included, and the sonographic checklist was completed at the time of ultrasound. Cases were defined as PAS where they had intraoperative findings as described by the International Federation of Gynecology and Obstetrics (FIGO) grading, and histopathological findings for hysterectomy and myometrial resection cases. All non‐PAS cases in this study had placenta previa and at least one prior cesarean delivery. Results Seventy‐eight participants met inclusion criteria, of whom 63 (80.7%) were diagnosed with PAS. Cesarean hysterectomy was performed in 49 cases (62.8%). Overall, third‐trimester ultrasound performed at a median gestational age of 32 weeks (IQR 30–34 weeks) had a sensitivity of 0.84 (95% CI 0.73 to 0.92) and specificity of 0.73 (95% CI 0.45 to 0.92) for detecting PAS, with a positive and negative likelihood ratio of 3.15 (95% CI 1.35 to 7.35) and 0.22 (95% CI 0.11 to 0.41), respectively. Features most associated with PAS were abnormal placental lacunae (Odds Ratio [OR] 5.40 [95% CI 1.61 to 18.03] and myometrial thinning OR 6.87 [95% CI 1.93 to 24.4]). While many of the ultrasound features seen in PAS were also present in cases of placenta previa with prior Cesarean section, the median (IQR) number of features present in PAS cases was significantly higher than in the non‐PAS placenta previa group (six features [3–8] vs. two features [0–3] p = 0.001). No case of non‐PAS placenta previa had more than five features present. Conclusions The use of a standardized sonographic checklist had a high sensitivity and good specificity for the detection of PAS in this prospective cohort of well‐classified PAS cases.
obstetrics & gynecology
What problem does this paper attempt to address?