This study aimed to investigate the diagnostic accuracy of the placenta accreta index (PAI) for predicting placenta accreta spectrum (PAS) in women with placenta previa. We analyzed 33 pregnancies with placenta previa at Keio University Hospital. The PAI was assessed in the early third trimester, and PAS was diagnosed histologically or clinically defined as retained placenta after manual removal attempts. The PAI and incidence of PAS were analyzed. Ten women (30%) were diagnosed with PAS and had higher volumes of perioperative bleeding (p = 0.016), higher rate of requiring uterine artery embolization (p = 0.005), and peripartum hysterectomy (p = 0.0002) than women without PAS. A PAI > 2 was the most useful cut-off point for predicting PAS and was more sensitive than prediction values using traditional evaluation (history of cesarean section and placental location). Post-hoc analysis revealed a higher rate of previous history of cesarean delivery (30% vs. 4.4%, p = 0.038), severe placental lacunae (≥grade2) (70% vs. 8.7%, p = 0.0003), thin myometrial thickness (90% vs. 22%, p = 0.0003), anterior placenta (100% vs. 30%, p = 0.0002), and presence of bridging vessels (30% vs. 0%, p = 0.0059) in PAS women. PAI could help predict the outcomes of women with placenta previa with and without a history of cesarean delivery to reduce PAS-induced perinatal complications.
Keywords: placenta accreta index; placenta accreta spectrum; placenta previa; ultrasonography.