Utility of the angle between the cervical canal and the anatomical conjugate line for predicting pouch of Douglas obliteration in patients with posterior placenta previa.

Placenta previa is associated with significant intraoperative and postpartum bleeding [ 1 , 2 ]. According to previous studies [ 3 – 5 ], the risk of placenta previa is higher in women with endometriosis than in those without endometriosis (6.0% versus 1.0%). Moreover, endometriosis is widely known as a major cause of posterior extrauterine wall adhesions [ 5 ]. Reliable predictors of massive hemorrhage and appropriate treatment strategies are vital for the successful management of placenta previa. Previous studies have reported that placenta accreta spectrum (PAS), advanced maternal age, history of previous cesarean section, and pouch of Douglas obliteration can serve as predictive factors of massive hemorrhage in women with placenta previa [ 6 – 8 ]. Various surgical methods, such as intrauterine balloon tamponade, uterine compression suture, and hysterectomy, have been used to control bleeding in women with abnormal placental location or placental invasion anomalies [ 9 – 12 ]. These surgical procedures usually entail exteriorization of the uterus, thereby raising the risk of intraoperative complications such as intestinal injury in women with pouch of Douglas obliteration. Magnetic resonance imaging (MRI) is usually used to assess the risk of pouch of Douglas obliteration before gynecological surgery. Several MRI findings associated with posterior extrauterine wall adhesions have been reported to date, including uterine retroflexion, displacement of intraperitoneal fluid, adhesion of bowel loops, and others [ 13 ]. However, these MRI findings were obtained in non-pregnant women, rendering them unsuitable for the assessment of the enlarged uterus during pregnancy. Moreover, risk assessment of pouch of Douglas obliteration in pregnant women has received little attention. According to a previous study [ 5 ], posterior extrauterine adhesions are more likely to be present in cases of placenta previa where the cervical canal is visualized horizontally. MRI scans of all patients included in their study were composed of 1.5-T superconducting magnets manufactured by General Electric (GE Healthcare, Waukesha, WI, USA), Philips (Philips North America, Andover, MA, USA), or Siemens (Siemens Corporation, Washington, DC, USA) [ 5 ]. Moreover, that study [ 5 ] reported that the cervical canal angle, which is defined as the angle between the line perpendicular to the broad of the back and the line passing through the internal os to the external os, was useful for predicting the absence of pouch of Douglas obliteration in women with placenta previa. Although this avenue is promising, it is possible that this method of measurement of the cervical canal angle is affected by maternal posture and soft tissues such as subcutaneous fat; therefore, it might not have high reproducibility. The line perpendicular to the broad of the back is important in calculating the cervical canal angle; however, correctly drawing the line of the broad of the back is relatively difficult due to maternal posture, soft tissues, and the extent of the back depicted on MRI. In other words, the cervical canal angle is at risk of being easily underestimated or overestimated, depending on the extent of the back depicted on MRI and the mother’s physique.” To overcome this issue, we devised a new measurement method that focuses on the inclination of the cervical canal relative to the anatomical conjugate line based on pelvimetry to predict pouch of Douglas obliteration. Assessment of pouch of Douglas obliteration is necessary for all pregnant women who are scheduled for a cesarean section. It is especially reasonable in women with placenta previa as they are likely to require surgical procedures to control bleeding, which may entail exteriorization of the uterus [ 2 , 8 – 11 ]. Therefore, we conducted this retrospective study to evaluate the utility of the angle formed by the anatomical conjugate line (based on pelvimetry) and the cervical canal on MRI for predicting pouch of Douglas obliteration in patients with posterior placenta previa.

During the study period, 44 women with posterior placenta previa underwent MRI for the preoperative assessment of PAS. Thirty-nine women were included in this study, after excluding 5 women (twin pregnancy, n = 2; PROM, n = 1; strong cervical canal curvature, n = 1; gauze tamponade, n = 1). The median maternal age was 34 (range, 22–44) years; 15 (38.5%) women were nulliparous, and 26 (66.7%) had term deliveries. The median CIA was 98° (range, 71°–128°). The overall incidence pouch of Douglas obliteration was six (15.4%). In evaluating the orientation of the uterus, 32 of the 39 women with posterior placenta previa were included in this study, after excluding 7 women (data missing, n = 5; inappropriate for diagnosis, n = 2). Three (9.4%) of the 32 women had a retroflexed uterus. Although the sample size of the retroflexed group was too small for us to draw a conclusion, the frequency of pouch of Douglas obliteration (66.7% versus 6.9%, respectively; p = 0.04) differed significantly between the retroflexed ( n = 3) and non-retroflexed groups ( n = 29). ROC analysis ( Fig 2 ) revealed that the cut-off value of 102° for the CIA would allow for the maximum number of women to be correctly classified according to the extrauterine posterior adhesions. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the curve (AUC) for this cut-off value were 66.7%, 78.8%, 36.4%, 92.9%, and 0.83, respectively. The characteristics of CIA ≥102° and CIA <102° groups were similar, except that the frequency of pouch of Douglas obliteration was higher in the CIA ≥102° group and the cervical length was shorter in the CIA <102° group ( Table 1 ). The inter-rater reliability for the CIA was relatively high with an ICC (2,1) of 0.80 (95% CI: 0.57–0.91). The inter-rater reliability for the cervical canal angle showed an ICC (2,1) of 0.64 (95% CI: 039–0.79). Values are presented as median (range) or number (%). Abbreviations MRI, magnetic resonance imaging ART, assisted reproductive technology BMI, body mass index GDM, gestational diabetes mellitus HDP, hypertensive disorders of pregnancy. † The orientation of the uterus was examined in 32 patients. For more details, see the Results section.

CIA may have a high ability to rule out an obliteration of the pouch of Douglas. It may also be a useful indicator in the surgical management of women with posterior placenta previa. However, caution should be exercised when generalizing the results of this study because of the small sample size, which makes the results prone to bias.

This retrospective observational cohort study was conducted at Yamanashi Prefectural Central Hospital between April 2016 and December 2021. We enrolled women with posterior placenta previa, including those with low-lying placenta as defined by a previous study [ 5 ], who underwent MRI for the preoperative assessment of PAS, delivered via cesarean delivery, and whose surgical records were available. Women with multiple pregnancies and premature rupture of fetal membranes (PROM) before MRI were excluded. Moreover, we excluded cases where gauze tamponade was performed as a hemostatic procedure for massive genital bleeding during MRI. The study protocol was reviewed and approved by the Human Subjects Review Committee of Yamanashi Prefectural Central Hospital (approval number: rinshou 2021–21), which waived the requirement for informed consent because of the retrospective design of this study. Nevertheless, the participants were provided with the opportunity to opt out of providing their data on the hospital’s website. All procedures were performed in accordance with the principles of the Declaration of Helsinki of 1964 and its subsequent amendments. Obstetric data were collected from the medical and operative records. Gestational age was determined on the basis of the maternally reported last menstrual period and was confirmed by the crown-rump length measured on ultrasonography performed during the first-trimester. We recorded data on the mother’s age at delivery, orientation of the uterus in the early first trimester, gestational age at MRI examination, cervical length, history of previous cesarean section, use of in vitro fertilization, parity, amount of bleeding including amniotic fluid, gestational age at delivery, maternal stature, pre-pregnancy weight status, fetal sex, neonatal birth weight, presence of gestational diabetes mellitus (GDM), and hypertensive disorders of pregnancy (HDP). Placenta previa was defined as the presence of placental tissue extending over the internal cervical os [ 14 ]. Low-lying placenta was defined as the location of the inferior placental edge within 2 cm of the internal os but not overlying it [ 15 ]. HDP was defined as blood pressure ≥140/90 mmHg on at least two occasions during pregnancy [ 16 ]. The pre-pregnancy body mass index was calculated according to the World Health Organization standard [body weight (kg)/height (m) 2 ]. The presence of ≥1 abnormal plasma glucose value (≥92, 180, and 153 mg/dL for fasting, 1-h, and 2-h plasma glucose concentrations, respectively) after a 75-g oral glucose tolerance test was diagnosed as GDM [ 17 ]. Cervical length was measured using transvaginal ultrasonography within one week of MRI examination. The orientation of the uterus was classified as retroflexed ( S1 Fig ) or non-retroflexed (anteverted or retroverted) ( S2 and S3 Figs) based on transvaginal ultrasonography conducted during early pregnancy (from 5 to 6 weeks gestation), since a retroflexed uterus may be associated with posterior deep infiltrating endometriosis [ 18 , 19 ]. We defined the angle formed by the anatomical conjugate line (based on pelvimetry) and the cervical canal as the cervical inclination angle (CIA), which was measured using MRI. MRI was performed at approximately 34 weeks gestation, since the mean reported gestational age at delivery in patients with placenta previa is 34–35 weeks [ 17 ]. However, the timing of MRI was left to the judgment of the treating obstetrician, depending on the presence of warning bleeding and signs of impending preterm labor. We projected a straight line through the anatomical conjugate line on the sagittal T2-weighted MRI scan ( Fig 1A and 1B , Line A). Thereafter, we identified the line passing through the internal os to the external os ( Fig 1A and 1B , Line B). We used a protractor to measure the angle between Lines A and B ( Fig 1A and 1B ), which was designated as the CIA ( Fig 1A and 1B ). The cervical canal angle was also calculated from the MRI scan, which has been described in detail in a previous study [ 5 ]. In all cases, two obstetricians (SS and RS), unaware of the women’s clinical characteristics and surgical findings, independently measured the CIA and cervical canal angle. The mean values were used for analysis. (a) Typical MRI findings in a patient without pouch of Douglas obliteration. (b) Typical MRI findings in a patient with pouch of Douglas obliteration. Based on a previous study [ 5 ], we considered pouch of Douglas obliteration to be present if any of the following criteria were documented in the surgical records: (1) adhesions between the posterior extrauterine wall and the small bowel, colon, rectum, or pelvic wall; (2) exteriorization of the uterus was impossible because of posterior adhesions; and (3) adhesions were dissected for extracorporeal elevation. All patients included in our study underwent MRI at Yamanashi Prefectural Central Hospital using two 1.5-T MR scanners (Signa HDxt; GE Medical Systems, DC, USA and Siemens Magnetom Symphony; Siemens AG, Erlangen, Germany). First, the Mann-Whitney U test and χ 2 test were performed to compare the maternal and neonatal outcomes. Fisher’s exact test was used when the expected frequency was <5. As the histograms were not normally distributed for several continuous variables, we selected the Mann-Whitney U Test, which can be used to test whether there is a difference between two groups, and the data need not be normally distributed. Second, receiver operating characteristic (ROC) curve analysis was conducted to determine the best cut-off value for the CIA and cervical canal angle. We used the Youden index, which describes the maximum vertical distance between the ROC curve and the diagonal (random chance), to estimate the optimal cut-off values [ 20 ]. Finally, we calculated the intraclass correlation coefficient (ICC) to evaluate the inter-rater reliability of the CIA and cervical canal angle. All analyses were performed using Bell Curve for Excel (Social Survey Research Information Co., Ltd., Tokyo, Japan) and IBM SPSS Statistics version 25 (IBM Inc., Tokyo, Japan). p -values <0.05 were considered statistically significant.

(TIF) Click here for additional data file. (TIF) Click here for additional data file. (TIF) Click here for additional data file. (XLSX) Click here for additional data file.