The value of a simple sonographic screening test for placenta accreta spectrum prediction: A case–control study

Placenta accreta spectrum (PAS) represents life‐threatening conditions; however, early diagnosis reduces complications and mortality rates.


INTRODUCTION
Placenta accreta spectrum (PAS) refers to a group of disorders characterised by abnormal adhesion or invasion of trophoblastic tissue, in which the villi adhere to the myometrium (accreta), penetrate the myometrium (increta), or invade into the serosa or surrounding structures (percreta). Accreta is the most common type (75%) of PAS, followed by increta (18%); the rarest and most severe type (7%) is percreta. [1][2][3][4] Increased rates of caesarean sections have led to increased incidences of invasive placenta, from 0.8 per 1000 births in the 1980s to three per 1000 births in more recent years. 5 Several risk factors for PAS have been identified, including advanced age, more births, uterine surgeries, fibroids and fetuses of the female gender. [6][7][8] The most significant risk factors are a previous caesarean section and placenta praevia. For women with placenta praevia, the risk of PAS increases with the number of previous caesarean sections: from 3% for one previous caesarean section to 67% for women with five or more. 9 PAS is a life-threatening condition associated with significant maternal morbidity, including postpartum haemorrhage, local organ damage and hysterectomy complications. 10,11 The most severe complications have mainly been described in women who were not suspected with PAS prenatally. 12 Diagnosis and identification of PAS before delivery reduce the complication and mortality rate by affording the opportunity to prepare, including surgical planning, through multispecialty consultations, such as with radiologists and experienced surgeons. 13 Prenatal ultrasound with grayscale and colour Doppler imaging is currently a promising diagnostic tool for PAS. 14,15 Welldocumented ultrasonographic signs of PAS in the second and third trimesters include loss of clear zone, myometrial thinning, placental lacunae, placental bed hypervascularity and bladder wall interruption. 16,17 However, diagnostic criteria of PAS have varied between studies. [16][17][18] The present single-institute study was carried out to review clinical and sonographic features, to analyse their value in diagnosing and predicting PAS, and to identify the most sensitive sonographic features, separately and in combination, for diagnosis. We developed a simple sonographic test that is accessible as screening for women with risk factors for PAS and that enables screening in routine pregnancy follow-up.

MATERIALSANDMETHODS
This study was endorsed by the Institutional Review Board Inclusion criteria included: women with a singleton pregnancy at 28-40 gestational weeks and risk factors for PAS, who had a sonography test to detect placenta accreta in the obstetrics and gynaecologic department of one regional medical centre, during 2010-2020. The included women also delivered at the same medical centre, and a clinical or pathological description of the placenta status was available.
Factors that were considered to increase the risk of PAS were: a history of one or more previous caesarean sections or of uterine curettage, myomectomy or scarred uteruses due to other reasons; or placenta praevia. Minor placenta praevia was classified when the lower edge of the placenta was inside the lower uterine segment or reached the internal os. Major placenta praevia was classified when the placenta partially or totally covered the cervix. 16 Gestational age was identified by calculating dates and by first-trimester ultrasound examinations.
All the sonographic tests were performed transabdominally, and also transvaginally when placenta praevia was detected. The ultrasound test included fetal viability, the well-being of the fetus and placenta location. Ultrasound and Doppler criteria suggestive of PAS included: lacunae (small/large), the location of the placenta (major or minor placenta praevia), and loss of a clear zone (or an irregular clear zone). In addition, women underwent Doppler assessment of hypervascularity of blood flow between the placenta and the bladder, which bridges vessels in the bladder interface, or retro-placental hypervascularity. 16,17 When PAS was suspected, during both caesarean and vaginal deliveries, obstetricians inspected the placenta for PAS according to clinical criteria of the FIGO classification. The latter included: no separation with synthetic oxytocin and cord traction; heavy bleeding from the placenta implantation site, requiring a mechanical or surgical procedure; an abnormal macroscopic finding over the placental bed; and significant amounts of hypervascularity in the uterine serosa. 19 Following delivery, the team suggested or excluded a PAS diagnosis, and this was confirmed with pathologic examination when available.
Data retrieved from the medical records of all the women included: age, gravidity, parity, previous caesarean section, uterine curettage, myomectomy or scarred uteruses due to other reasons, the ultrasound signs of PAS detailed above, the delivery mode and outcomes, placenta description, postpartum haemorrhage and other complications. For the entire cohort of women who met the inclusion criteria, we analysed demographic characteristics, ultrasonographic reports, delivery data and surgical/ pathological findings when present. We analysed and compared the data for women with confirmed PAS at delivery and women without confirmed PAS (control group).

Descriptiveanalysis
Categorical data were described using frequencies and percentages. Continuous variables with normal distributions were presented as means ± standard deviations. Median values and ranges were used to describe variables that did not distribute normally.

Inferentialanalysis
Categorical variables were compared between the groups with the χ 2 test or Fisher's exact tests (when expectancy was <5).
Continuous variables were compared between the groups using the independent t-test or the Mann-Whitney test (if normal distribution was found, the independent t-test was used). The distribution shape was determined mainly by a histogram. P < 0.05 was considered significant. Multivariable linear regression models were performed to examine correlations of PAS and sonographic signs; and sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and predictive probability were

Ethics approval
This study was endorsed by the Institutional Review Board

RESULTS
We included 481 women with risk factors for PAS who underwent ultrasound for PAS detection from January 2010 to December 2020. Of them, 422 (86.7%) had no accreta at delivery and 59 (12.3%) had a confirmed PAS.
Demographic characteristics are presented in Table 1.
Significant differences were not observed between the PAS and control groups regarding age (33 vs 34 years), the proportion of women with obesity (body mass index >30 kg/m 2 ), gravidity, parity, rates of caesarean section and uterine curettage, and a history of postpartum haemorrhage. The only significant difference was the higher rate of other uterine surgeries in the PAS than the control group: (four) 6.9% vs (three) 0.7%, P = 0.005.
Significant differences were not observed between the groups in pregnancy week at the sonographic test, while differences were observed in placenta location (Table 2). In the PAS compared to the control group, the rates were higher for major placenta praevia ((48) 81% vs (25) 5.9% P < 0.001) and minor placenta praevia ((five) 8.5% vs (ten) 2.4%, P = 0.001). Large and small vascular lacunae were more frequently observed in the PAS than the control group (55.6% vs 0.2%, P = 0.001 and 66.7% vs 5%, P = 0.001, respectively). A lack of a clear zone was more frequently observed in the PAS than the control group (67.3 vs 1.4%, P = 0.001).
The rate of caesarean section did not differ between the PAS

DISCUSSION
Delays in diagnosing PAS can significantly impact maternal morbidity and mortality. 13 Given the importance of early detection of placenta accreta, women with risk factors for PAS should be treated by a multidisciplinary medical team.
During the ten-year study period, PAS was confirmed, based on the FIGO classification, 19 in 59 of 481 (12.3%) women who underwent ultrasound for PAS detection. We evaluated the sonographic signs of PAS in such women to identify the most sensitive and specific sonographic signs and to develop, test and quantify   The area-under-the-curve was 0.87. These systems can be challenging to perform at a routine sonographic follow-up in the second and third trimesters.
Notably, we observed large and small placenta lacunae more frequently in women with PAS than in the control group (55.6% vs 0.2% and 66.7% vs 5%, respectively). This finding is compatible with a study that reported lacunae in 58.3% of women with PAS. 23 Moreover, the lack of a clear zone was more frequently seen in women with PAS than in women without PAS (67.3 vs 1.4%).
Similar data were reported in another study (60% vs 1.5%). 24 The main strength of our study is the specific evaluation of sonographic signs to identify and evaluate a simple screening test that includes a small number of clear sonographic signs with high predictive value for PAS detection. In addition, this is one of the largest cohorts of women with risk factors for PAS who had a specific sonographic test for PAS diagnosis. Limitations of the study include its retrospective design and the lack of complete documentation of the sonographic and postpartum findings. A prospective validation study should verify our results.
In conclusion, we developed a simple screening test for women with high risk of PAS. The test demonstrated high predictive value, sensitivity, specificity and NPV, and can be helpful for predicting PAS in routine sonographic follow-up at 28 weeks or later. A positive screening test is defined as a major or minor placenta praevia, together with one of two signs: large lacunae or the lack of a clear zone. Our model may assist clinicians in selecting women to perform a specific sonographic or magnetic resonance imaging test to diagnose antenatal PAS.