Postpartum hemorrhage (PPH) is a leading cause of severe maternal morbidity¹ and accounts for approximately 25% of global maternal mortality².​​ Placenta accreta spectrum (PAS) disorders – encompassing placenta accreta, increta, and percreta - result from pathological trophoblastic invasion involving abnormal implantation with adhesion extending beyond the decidua basalis into the uterine myometrium³. PPH represents one a frequent and critical complications in PAS⁴, making prenatal risk stratification essential for optimizing obstetric management and maternal outcomes.

Ultrasound imaging is a well-established diagnostic tool for PAS⁵. Multiple studies have explored various ultrasound techniques to predict PPH risk in PAS patients^6–12. However, significant heterogeneity in placental invasion depth – even among cases with identical diagnoses¹³, limits the consistency of conventional ultrasound methods for PPH¹⁴.

Crystal Vue, a novel ultrasound rendering technology, enhances tissue contrast to improve differentiation of structures based on echogenicity¹⁵. While this modality has shown promise in PAS diagnosis by identifying characteristic sonographic signs^16–18, its utility for predicting PAS-associated remains largely unexplored. Therefore, this study aims to use Crystal Vue to quantitatively estimate the placental invasion area in PAS patients and access its association with PPH risk during delivery.

A total of 168 patients diagnosed with PAS disorders and meeting the study’s inclusion criteria were enrolled. These patients were stratified into the PPH group (n = 78) and the non-PPH group (n = 90). Baseline characteristics, stratified by study group, are presented in Table 2. Comparative analysis revealed significant differences between the two groups across multiple variables.​ Specifically, patients in the PPH group exhibited higher frequencies of gravidity (P = 0.01), parity (P < 0.001), and prior cesarean deliveries (P < 0.001) compared to those in the non-PPH group. Furthermore, the prevalence of smoking was significantly greater among patients in the PPH group (12.82% vs. 3.33%; P = 0.02). The PPH group also demonstrated a shorter gestational duration (242 days vs. 246 days; P < 0.001). Among patients in the PPH group, 61.64% were diagnosed with placenta previa. Regarding the specific types of PAS disorders, the PPH group showed a higher proportion of cases diagnosed as placenta increta (62/78, 79.49%) and placenta percreta (16/78, 20.51%). Notably, the estimated PAS area demonstrated a more than 2-fold elevation in the PPH group compared with the non-PPH cohort (17.28mm² vs. 7.36mm²; P < 0.001).

Based on collinearity analysis showing collinearity between parity and the number of previous cesarean sections, parity was excluded from the analysis, and the number of previous cesarean sections was retained, guided by a clinical perspective. Linear regression analysis, adjusted for all potential confounders, demonstrated that increased estimated PAS area was independently associated with a significant increase in EBL (40.13 (95% CI, 26.84–53.42) mL/mm²; P < 0.001). Likewise, multivariate logistic regression analysis revealed that each 1 mm² increment in estimated PAS area was associated with a 7.2% increase in the risk of PPH (OR 1.07; 95% CI 1.03–1.13; P = 0.003) and a 3% increase in the risk of severe PPH (OR 1.03, 95% CI 1.01–1.05; P = 0.01). No statistically significant associations were observed between estimated PAS area and maternal ICU admission or the need for transfusion (Table 3). ROC curve analysis, utilizing the Youden index, determined the optimal diagnostic threshold for estimated PAS area in predicting PPH (cut-off value, 10.12mm²; AUC (95% CI), 0.82 (0.76–0.88); P < 0.001) and severe PPH (cut-off value, 10.57mm²; AUC (95% CI), 0.83 (0.75–0.89); P < 0.001) (Fig. 3). Further evaluation of individual ultrasound parameters revealed that the estimated PAS area, when utilized with these cut-off values, demonstrated superior diagnostic performance for predicting PPH and severe PPH compared to classic ultrasound signs (Table 4, 5).

This study indicates that the estimated PAS area is an independent risk factor for both PPH and severe PPH. Furthermore, it demonstrates superior predictive efficacy compared to classic ultrasound signs in forecasting PPH and severe PPH occurrence in PAS patients.

Research on prenatal PAS has primarily focused on ultrasound’s diagnostic efficacy, with few studies exploring risk stratification for severe maternal mobility. Yang et al. developed a scoring system combining maternal history and classic ultrasound signs, showing reasonable predictive efficacy for EBL ≥ 1500 ml (AUC = 0.76), but it did not assess risk for EBL ≥ 1000 ml⁶. A Thai study reported a scoring system using 2D ultrasound grey scale and color Doppler imaging with good predictive performance for massive EBL (≥ 2500 ml) (AUC = 0.80); however, only 33.91% of enrolled participants had confirmed PAS⁷. Gali et al. established a PAS staging system based on prenatal ultrasound signs, demonstrating significant correlation with maternal outcomes (EBL and ICU admission rates)⁸. However, this system’s ability to predict life-threatening hemorrhage in PAS patients was not investigated.

Despite PAS being associated with hemorrhagic morbidities (massive blood loss, substantial transfusion) across severity levels^{9, 22}, current practice lacks effective methods to predict diverse adverse maternal outcomes. Blood exceeding 2500 mL is significantly associated transfusion need, ICU admission, and complications like acute renal failure or even maternal mortality^{23, 24}. Beyond immediate prognosis, PPH increases long-term risks like postpartum depression and anemia²⁵. As a critical cause of catastrophic PPH²⁶, accurate prenatal prediction of PPH risk in PAS patients could enable more comprehensive preoperative preparation, reducing maternal morbidity and mortality. While Crystal Vue is validated for prenatal PAS diagnosis¹⁵, limited evidence exists regarding its ability to assess PPH in PAS risk. In this study, we estimated the PAS area by measuring the length of tramline obliteration and its distance from the cervix, dividing the placental invasion topography into trapezoidal segments. Our findings indicate that each 1 mm² increase in the estimated PAS area is associated with an additional 40.13 mL blood loss during delivery, suggesting this novel method has clinical utility as a PPH risk indicator.

This study’s primary strength is the novel development of a Crystal Vue-based method to estimate PAS area. Strict enrollment criteria (prenatal and postnatal PAS diagnosis confirmation) minimized bias risk. Consistent prenatal management by a single multidisciplinary team following identical protocols, minimized bias related to operator experience and surgical approach. The remain limitation is the substantial topographical variability among PAS patients, which prevented establishing a standardized examination protocol (e.g., optimal timing, precise number of trapezoidal segments). Second, as a retrospective study, we did not assess whether clinical application of this method influence the prenatal management and PAS prognosis. Future research should explore integrating additional ultrasound signs to construct predictive models enhancing PPH risk prediction efficacy. Furthermore, investigating whether this method improves PAS prognosis by influencing prenatal management strategies is essential.

This study developed a novel Crystal Vue-based method to estimate PAS area and demonstrated its correlation with PPH risk in PAS patients. Further large prospective studies are needed to validate this method and to investigate whether its integration into clinical practice can aid in determining the optimal surgical approaches and improving outcomes for women with PAS disorders.