Intraplacental Fetal Vessel Diameter May Help Predict for Placental Invasiveness in Pregnant Women at High Risk for Placenta Accreta Spectrum Disorders
Abstract
Background
Prenatal identification of placenta accreta spectrum (PAS) disorder is essential for treatment planning. More objective means for predicting PAS and clinical outcome may be provided by MRI descriptors.
Purpose
To investigate the association of intraplacental fetal vessel (IFV) diameter at MRI with PAS and peripartum complications.
Materials and Methods
Between March 2016 and October 2019, 160 gravid women suspected of having PAS underwent placental MRI as part of a prospective trial. Secondary analysis was performed by two experienced genitourinary radiologists for presence and maximum diameter of IFVs. Relative risk ratios were computed to test the association of IFVs with presence and depth of PAS invasiveness. Receiver operating characteristic analysis was used to evaluate the ability of IFV diameter to help predict PAS, placenta percreta, and peripartum complications and for comparison of the area under the curve (AUC) versus that from other combined MRI predictors of PAS (eg, myometrial thinning, intraplacental T2-hypointense bands, uterine bulge, serosal hypervascularity, and signs of extrauterine placental spread). Intraoperative and histopathologic findings were the reference standard.
Results
A total of 155 women were evaluated (mean age, 35 years ± 5 [standard deviation]; mean gestational age, 32 weeks ± 3). PAS was diagnosed in 126 of 155 women (81%) (placental percreta in 68 of 126 [54%]). At delivery, 30 of 126 women (24%) experienced massive blood loss (>2000 mL). IFVs were detected at MRI in 109 of 126 women with PAS (86%) and in 67 of 68 women with placental percreta (98%). The relative risk ratio was 2.4 (95% CI: 1.6, 3.4; P < .001) for PAS and 10 (95% CI: 1.5, 70.4; P < .001) for placental percreta when IFVs were visible. IFVs of 2 mm or greater were associated with PAS (AUC, 0.81; 95% CI: 0.67, 0.95; P = .04). IVFs of 3 mm or greater were associated with placenta percreta (AUC, 0.81; 95% CI: 0.73, 0.89; P < .001) and with peripartum complications, including massive bleeding (AUC, 0.80; 95% CI: 0.71, 0.89; P < .001). Combining assessment of IFVs with other MRI descriptors improved the ability of MRI to predict PAS (AUC, 0.94 vs 0.89; P = .009).
Conclusion
Assessment of intraplacental fetal vessels with other MRI descriptors improved the ability of MRI to help predict PAS. Vessel diameter of 3 mm or greater was predictive of placenta percreta and peripartum complications.
© RSNA, 2020
Online supplemental material is available for this article.
See also the editorial by Dighe in this issue.
Summary
The presence of one or more intraplacental fetal vessels with a diameter of 2 mm or greater at MRI was associated with placental invasiveness in pregnant women.
Key Results
■ Women with visible intraplacental fetal vessels at MRI had twice the risk for abnormal placentation compared with those without (P < .001); the risk for placental percreta was 10 times higher (P < .001).
■ The diameter of intraplacental fetal vessels was related to the extent of placental invasiveness; vessels 2 mm in diameter or greater were associated with placenta accreta spectrum disorders (area under the curve [AUC], 0.81; P = .04); vessels with a diameter of 3 mm or greater were associated with diagnosis of placental percreta (AUC, 0.81; P < .001).
■ Addition of intraplacental fetal vessels improved the diagnostic ability of placental MRI (AUC, 0.94 vs 0.89; P = .009).
Introduction
The incidence of placenta accreta spectrum (PAS) disorders has surged during the past decades, with a reported prevalence of one in 540 to one in 2500 deliveries in western countries because of the higher number of uterine interventions, in particular cesarean deliveries (1). Unexpected findings of abnormal placentation during delivery may be associated with uncontrollable obstetric hemorrhage, emergency hysterectomy, and, in aggressive forms, urinary tract injury, leading to higher morbidity and mortality rates for both mother and fetus.
Accurate prenatal diagnosis of this condition allows for appropriate preoperative consultation, planned preterm delivery, multidisciplinary care, or even conservative treatment options (such as a placenta-left-in-situ approach and uterine-sparing surgical techniques), which may all improve patient prognosis (2–4). Although color Doppler US remains the first-line imaging investigation (5), MRI has a complementary role in preoperative diagnosis of PAS and surgical planning because it offers a larger imaging field of view and higher reproducibility compared with US (6–10).
Several studies have suggested that the presence of abnormal intraplacental vasculature is an important MRI feature of PAS (11–13). However, the Society of Abdominal Radiology and European Society of Urogenital Radiology consensus statement (14) did not include abnormal intraplacental vasculature in the recommended signs for PAS because members disagreed about its definition. A recent study with MRI–pathologic correlation found an abnormal pathologic intraplacental vascular pattern in placenta percreta and demonstrated a fetal origin of these vessels (15). The authors claimed that these abnormal vessels were enlarged subchorionic and stem vascular trunks, originating from the umbilical cord and running deep into the placental parenchyma, often reaching its maternal surface. Normal placental fetal vessels become imperceptible soon after they enter the placenta, but in PAS, these vessels are more elongated and wider with deficient branching, surrounded by sparse chorionic tissue compared with the normal placenta (described pathologically as the stripped fetal vessel sign). Defective fetal vasculogenesis may account for this aberrant intraplacental vascular pattern in PAS (15).
The aim of our study was to investigate the association of MRI-depicted intraplacental fetal vessels (IFVs) with PAS and peripartum outcome.
Materials and Methods
The institutional review board approved this observational study (ethics registration number B-196/13.10.2016). All participants provided written informed consent; secondary analysis of prospectively collected data was performed in a random order. Subsets of the study sample were included in three previous reports (10,15,16) with distinct concept, design, and results. Konstantinidou et al (15) correlated intraplacental vessels detected at MRI in 11 patients with surgical pathologic findings of percreta and reported a fetal origin for these vessels. Two previous reports assessed the prognostic ability of several combined MRI signs for clinical outcome (n = 100) (16) and for placenta percreta (n = 49) in women suspected of having PAS (10); MRI signs included abnormal intraplacental vascularity (11), with a distinctly different definition from the IFVs described here.
Patient Selection
Between March 2016 and October 2019, 160 gravid women in the third trimester were referred for placental MRI from two obstetrical units dedicated to PAS treatment; all were at high risk for PAS because of placenta previa (n = 150) and/or suspicious findings at second-trimester US (n = 81). Five women did not proceed with MRI because of claustrophobia (n = 3) or obesity (n = 2). A flowchart of women who participated in the study is shown in Figure 1.
Figure 1: Flowchart of inclusion and exclusion criteria of the study sample. PAS = placenta accreta spectrum.
MRI Protocol
MRI examinations were conducted at 1.5 T (n = 100) or 3.0 T (n = 55) (MultiVaneXD; Philips Healthcare, Best, the Netherlands). The MRI protocol for 1.5-T imaging included the following sequences: T2-weighted single-shot turbo spin-echo sequence in the axial, sagittal, and coronal planes (repetition time msec/echo time msec, 510–568/80); T2-weighted turbo spin-echo sequence in all three planes (6300/90) and, when extrauterine spread was suspected, parallel and perpendicular to the cervical canal (2500/90); and axial T1-weighted turbo spin-echo fat-suppressed sequence (730/6.9). The MRI protocol for 3.0-T imaging included T2-weighted imaging in all three planes (4345–4666/100), T2-weighted turbo spin-echo sequence perpendicular and parallel to the cervix (4807/90), and axial T1-weighted turbo spin-echo fat-suppressed sequence (723/8.0). No intravenous paramagnetic contrast material was administered. Total imaging time was less than 35 minutes. Acquisition protocols are provided in Tables E1 and E2 (online).
MRI Analysis
All MRI readings were performed independently by two radiologists experienced in genitourinary MRI (L.A.M., with 22 years of experience, and C.B., with 11 years of experience) who were blinded to all information except patient and gestational age. The interval between readout sessions was 3 months. In cases of discordance, a consensus was reached by accepting the positive diagnosis of either reader.
Before the reading sessions, both readers agreed on the definitions of MRI terms (15). IFVs were diagnosed when one or more flow-void structures were observed within the placenta, originating from the umbilical cord or the chorionic and/or subchorionic fetal placental surface; the blanket term chorionic and/or subchorionic was used because discrimination between chorionic and subchorionic space at MRI is nearly impossible (Fig 2). To estimate extent, we recorded the course of vessels in the inner (close to the fetal surface, <50% of total placental thickness) and outer (close to the maternal surface, ≥50% of total placental thickness) placental halves (Fig 3) by carefully studying all planes. Maximal diameter was calculated at the thickest part of the vessel’s course (outer-outer wall) in any plane (Figs 3–5, Figures E1 and E2 [online]); we avoided measurements within 10 mm from the chorionic/subchorionic surface because vessels are normally larger close to their origin from the umbilical cord (Figure E3 [online]). IFVs were best shown on high-resolution T2-weighted images (Figure E4 [online]).
Figure 2a: Images in 39-year-old gravid woman with placenta percreta at gestational week 35. (a) Sagittal T2-weighted turbo spin-echo MRI scan (1.5 T) shows elongated flow-void structure (white arrow), originating from umbilical cord (black arrow) and extending deep into placental parenchyma. (b) Photograph of corresponding placental specimen confirms presence of long vascular trunk (white arrow) originating from umbilical cord (black arrow) with deficient side branching.
Figure 2b: Images in 39-year-old gravid woman with placenta percreta at gestational week 35. (a) Sagittal T2-weighted turbo spin-echo MRI scan (1.5 T) shows elongated flow-void structure (white arrow), originating from umbilical cord (black arrow) and extending deep into placental parenchyma. (b) Photograph of corresponding placental specimen confirms presence of long vascular trunk (white arrow) originating from umbilical cord (black arrow) with deficient side branching.
Figure 3a: Images in 39-year-old woman at gestational week 32 with dichorionic twin pregnancy diagnosed with placenta percreta. (a) Sagittal T2-weighted turbo spin-echo MRI scan (3.0 T) demonstrates large intraplacental fetal vessel (arrow) within nonprevia placenta percreta (P1). Note absence of intraplacental vascularity in normal second placenta (P2). UC = umbilical cord. (b) Photograph of corresponding formalin-fixed placental specimen from hysterectomy (P1) shows abnormal intraplacental fetal vessels (arrow) originating from UC.
Figure 3b: Images in 39-year-old woman at gestational week 32 with dichorionic twin pregnancy diagnosed with placenta percreta. (a) Sagittal T2-weighted turbo spin-echo MRI scan (3.0 T) demonstrates large intraplacental fetal vessel (arrow) within nonprevia placenta percreta (P1). Note absence of intraplacental vascularity in normal second placenta (P2). UC = umbilical cord. (b) Photograph of corresponding formalin-fixed placental specimen from hysterectomy (P1) shows abnormal intraplacental fetal vessels (arrow) originating from UC.
Figure 4a: Images in 34-year-old primiparous woman with placenta previa at gestational week 33. (a) Coronal oblique T2-weighted turbo spin-echo MRI scan (3.0 T) demonstrates few flow-void structures with maximal diameter of 2.2 mm, extending from chorionic and/or subchorionic placental surface deep into placenta (arrow). (b) Photograph of intraoperative findings confirms placenta increta (arrows) at right posterolateral uterine surface, close to internal cervical os. (c) Photograph of specimen obtained at surgical-pathologic examination shows aberrant intraplacental fetal vessels in placental specimen (arrows). Scale is in centimeters.
Figure 4b: Images in 34-year-old primiparous woman with placenta previa at gestational week 33. (a) Coronal oblique T2-weighted turbo spin-echo MRI scan (3.0 T) demonstrates few flow-void structures with maximal diameter of 2.2 mm, extending from chorionic and/or subchorionic placental surface deep into placenta (arrow). (b) Photograph of intraoperative findings confirms placenta increta (arrows) at right posterolateral uterine surface, close to internal cervical os. (c) Photograph of specimen obtained at surgical-pathologic examination shows aberrant intraplacental fetal vessels in placental specimen (arrows). Scale is in centimeters.
Figure 4c: Images in 34-year-old primiparous woman with placenta previa at gestational week 33. (a) Coronal oblique T2-weighted turbo spin-echo MRI scan (3.0 T) demonstrates few flow-void structures with maximal diameter of 2.2 mm, extending from chorionic and/or subchorionic placental surface deep into placenta (arrow). (b) Photograph of intraoperative findings confirms placenta increta (arrows) at right posterolateral uterine surface, close to internal cervical os. (c) Photograph of specimen obtained at surgical-pathologic examination shows aberrant intraplacental fetal vessels in placental specimen (arrows). Scale is in centimeters.
Figure 5: Images in 35-year-old pregnant woman with placenta percreta at gestational week 35. Coronal T2-weighted turbo spin-echo MRI scan (3.0 T) shows large intraplacental fetal vessel with 9-mm maximal diameter (arrow), which is also demonstrated on corresponding hysterectomy-placental specimen (arrow in inset photograph). Patient experienced major complications during surgery and required massive transfusion (16 units of red packed blood cells) and extensive bladder wall reconstitution. UC = umbilical cord.
Each reader also recorded the presence of the following MRI features of PAS (9–14,16–22): myometrial thinning, intraplacental T2-hypointense bands, uterine bulge, exophytic placental mass, serosal, bladder and parametrial vessel sign, and signs of bladder invasion. A detailed description of these MRI terms is provided in Table E3 (online).
Reference Standard
All cesarean deliveries were performed by two obstetricians dedicated to PAS treatment (S.F., with 25 years of experience, and G.D., with 25 years of experience). Surgical evidence was the standard of PAS invasiveness, according to the Fédération Internationale de Gynécologie et d’Obstétrique, or FIGO, classification (23). Histologic examination of the utero-placental specimen (following hysterectomy) or gross examination of the removed placenta (following conservative treatment) were performed by one perinatal pathologist (A.E.K., with 22 years of experience) to support surgical diagnosis (pathologic evidence). Both obstetricians agreed on the definitions of clinical terms related to the peripartum maternal course (16,24,25). Detailed descriptions of all surgical, clinical, and histologic terms are provided in Table E4 (online).
Statistical Analysis
Variables with approximately symmetric distributions were summarized as mean and standard deviation, variables with skewed distribution as median and interquartile range (IQR), and qualitative variables as absolute and relative frequencies. We compared κ coefficients as a measure of agreement testing the null hypothesis of no agreement (κ = 0) between intraoperative diagnosis and/or histologic examination and MRI results and between radiologists for the presence of IFVs (κ = 1, perfect; κ ≥0.75, excellent; κ = 0.40–0.75, fair to good, κ <0.40, poor agreement) (26). The interrater agreement for IFV diameter was evaluated with intraclass correlation coefficients (ICCs) testing the null hypothesis of no agreement (ρ = 0) (ICC ≤0.40, poor/fair; ICC of 0.41–0.60, moderate; ICC of 0.61–0.80, good; ICC ≥0.80, excellent agreement) (27). For comparisons of proportions, χ2 and Fisher exact tests were used. Student t tests were computed for comparison of mean values when the distribution was approximately symmetric, and the Mann-Whitney test was used when distribution was not approximately symmetric. The Spearman correlation coefficient explored the association of two continuous variables.
Relative risk ratios were computed to test the predictive ability of IFVs and other MRI signs for presence and depth of invasiveness of PAS disorders. Receiver operating characteristic analysis and area under the curve (AUC) were used to evaluate the predictive ability of IFV diameter for PAS, placenta percreta, and adverse peripartum outcomes (including massive intraoperative bleeding, hysterectomy, and bladder repair). Sensitivity and specificity were determined for optimal cut-offs and were evaluated using the Youden index (28).
To investigate whether there was evidence of a difference in the predictive ability of IFV compared with all other MRI signs combined, AUCs were compared using the Wald test of the null hypothesis that all AUC values are equal (29). A bootstrap resampling procedure (1000 samples) was used to cross-validate relative risk ratios for IFVs and AUCs calculated for the combination of MRI signs. All P values reported were two tailed. Statistical significance was set at P = .05, and analyses were conducted with Stata software (version 11.0; StataCorp, College Station, Tex).
Results
Patient Characteristics
One hundred fifty-five women (mean age, 35 years ± 5 [standard deviation]; range, 20–47 years); mean gestational age at MRI, 32 weeks ± 3; range, 24–39 weeks] completed the MRI examination. Demographic and clinical characteristics are described in Table 1. Sixty-four of 155 women (41%) had a history of cesarean delivery only; 37 of 155 (24%) had both a cesarean delivery and at least one other uterine intervention; 27 of 155 (17%) had no history of cesarean delivery but at least one other uterine intervention; and 27 of 155 (17%) had no previous uterine procedures. Placenta previa was diagnosed in 146 of 155 women (94%): complete in 96 of 146 (66%), partial in 23 of 146 (16%), marginal in 19 of 146 (13%), and low-lying in eight of 146 (5%). In nine of 155 women (6%), placenta location was normal. All women underwent cesarean delivery within a median interval of 2 weeks (IQR, 1–4 weeks) from MRI.
 |
PAS was diagnosed in 126 of 155 women (81%), placenta percreta in 68 of 126 women (54%), and placenta creta and/or increta in 58 of 126 women (46%). Bladder involvement was recorded in 54 of 155 women (35%) and parametrial invasion in 23 of 155 women (15%); 28 of 54 women (52%) required extensive bladder repair. Thirty of 155 women (19%) experienced massive hemorrhage (>2000 mL) intraoperatively. PAS was finally diagnosed in 20 of 27 women (74%) with previous noncesarean interventions.
Hysterectomy treatment was performed in 56 of 155 women (36%), and a uterine-sparing surgical approach with subsequent placental removal (modified triple-P procedure) was successfully performed in 70 of 155 women (45%), all of whom had surgical evidence of PAS. In 29 of 155 women (19%) who had a normal placenta during delivery, cesarean delivery and manual placental detachment were performed. Endovascular assisted hemostasis or a placenta-left-in-situ approach was not applied. No maternal deaths were reported; two neonatal deaths resulted from acute onset of delivery at gestational week 29 and 26, respectively.
In 56 women with placenta percreta and hysterectomy treatment, histologic findings were in accordance with surgical diagnosis. In 70 women with PAS (placenta creta in nine, placenta increta in 49, and placenta percreta in 12) and uterine preservation, pathologic classification of grade of invasiveness was not feasible. In 29 of 155 women (19%), the placenta was normal at surgery and histologic evaluation.
MRI Predictive Ability for PAS
Agreement between intraoperative and/or histologic findings and the combination of MRI signs, including IFVs, was excellent (κ = 0.81 for PAS detection, κ = 0.93 for bladder involvement, and κ = 0.90 for parametrial involvement) (Table E5 [online]).
Incidence of IFVs
IFVs were the most frequently detected MRI feature (113 of 155 women [73%]); in most women, IFVs extended deep into the placental parenchyma (107 of 113 women [95%]). Median IFV diameter was 2.9 mm (IQR, 2.0–4.8 mm). The incidence of all recorded MRI signs is provided in Table E6 (online).
Interrater Agreement for Presence, Extent, and Diameter of IFVs
Interrater agreement for presence and extent of IFVs was excellent (κ = 0.77 and 0.78, respectively). The ICC for IFV maximal diameter between the two radiologists was 0.81 (excellent).
Association of IFVs with Intraoperative and/or Histologic Results and Adverse Peripartum Outcome
Association of presence and extent of IFVs with intraoperative/histologic results and delivery outcome is described in Tables 2 and 3.
 |
 |
IFVs were associated with PAS (P < .001), depth of invasiveness (P < .001), bladder involvement (P < .001) or parametrial involvement (P = .001), bladder repair (P < .001), massive blood loss (>2000 mL) during surgery (P < .001), and hysterectomy (P < .001) (Table 2).
IFVs exhibited a relative risk ratio of 2.4 for detection of PAS (95% CI: 1.6, 3.4; P < .001) and of 10 for prediction of depth of invasiveness (95% CI: 1.5, 70.4; P < .001). After the bootstrap procedure, the relative risk ratio of IFVs was 2.4 (95% CI: 1.6, 3.6) for prediction of PAS and 10 (95% CI: 2.7, 14.4) for prediction of depth of invasiveness (Table 3).
The course of IFVs to both inner and outer placental half showed a relative risk ratio of 2 (95% CI: 0.9, 4.4; P < .001) for PAS and of 0.4 (95% CI: 0.3; 0.5, P = .027) for depth of invasiveness (Table 3) and was associated with bladder repair or hysterectomy (both P = .03) (Table 2).
Association of IFV Diameter with Intraoperative and/or Histologic Results and Adverse Peripartum Outcomes
IFV diameter was greater in women with placenta percreta (P < .001) than in those with placenta creta and/or increta and in women with poor clinical outcome, including all recorded adverse peripartum variables (P < .001). Association of IFV diameter with intraoperative and/or histologic results and peripartum complications is reported in Table E7 (online).
Table 4 presents the results from receiver operating characteristic analysis for prediction of PAS, placenta percreta, and adverse peripartum events, based on IFV diameter; the AUC was 0.81 (P = .04) for PAS, 0.81 (P < .001) for placenta percreta, and 0.80–0.84 (P < .001) for prediction of major peripartum complications. Optimal cut-off for IFV diameter for prediction of PAS was 2 mm, with a sensitivity of 72% (95% CI: 63, 81) and a specificity of 75% (95% CI: 19, 99). Optimal cut-off for prediction of percreta was 3 mm, with a sensitivity of 70% (95% CI: 58, 81) and a specificity of 81% (95% CI: 66, 91). Cut-offs for prediction of massive intraoperative blood loss, hysterectomy, and bladder repair were 3 mm, 2.9 mm, and 2.9 mm, respectively, with 76%–83% sensitivity and 64%–78% specificity ranges.
 |
Incremental Prognostic Value of IFVs for Predicting PAS
Table 3 shows the association of all recorded MRI signs with PAS and depth of placental invasiveness.
Receiver operating characteristic analysis revealed that the combination of all recorded MRI signs except IFVs was predictive for PAS, with an AUC of 0.89 (95% CI: 0.84, 0.93; P < .001). No evidence of difference existed (P = .53) in the predictive ability of IFVs alone (AUC, 0.86; 95% CI: 0.78, 0.94; P < .001) compared with all other MRI signs combined (Fig 6a). Addition of IFVs to the combined MRI signs increased (P = .009) the AUC to 0.94 (95% CI: 0.91, 0.98; P < .001) (Fig 6b). After the bootstrap procedure, the AUC was stable at 0.89 for the combined MRI signs except for IFVs and at 0.94 for all combined MRI signs, including IFVs.
Figure 6a: (a) Receiver operating characteristic curves show no evidence of difference (P = .53) in predictive ability of intraplacental fetal vessel (IFV) sign for placental invasiveness compared with that of combination of all other significant MRI signs of placenta accreta spectrum. (b) Receiver operating characteristic curves of model obtained from combination of all significant for placenta accreta spectrum MRI signs, including IFV sign, show increase in area under curve (P = .009), supporting incremental prognostic value of IFV sign for detecting placenta accreta spectrum.
Figure 6b: (a) Receiver operating characteristic curves show no evidence of difference (P = .53) in predictive ability of intraplacental fetal vessel (IFV) sign for placental invasiveness compared with that of combination of all other significant MRI signs of placenta accreta spectrum. (b) Receiver operating characteristic curves of model obtained from combination of all significant for placenta accreta spectrum MRI signs, including IFV sign, show increase in area under curve (P = .009), supporting incremental prognostic value of IFV sign for detecting placenta accreta spectrum.
IFV Predictive Ability for PAS in 1.5-T and 3.0-T Groups
IFVs were predictive of PAS in both 1.5-T (n = 100) and 3.0-T (n = 55) groups (P < .001). IFV diameter of 3 mm or greater had significant predictive ability for depth of invasiveness in both 1.5-T (P = .035) and 3.0-T (P = .001) groups (Table E8 [online]).
Discussion
We investigated the performance of intraplacental fetal vessels (IFVs) at MRI for predicting placental invasiveness and peripartum outcome in gravid women suspected of having placenta accreta spectrum (PAS). Our results suggest that the presence of IFVs is associated with a higher risk for PAS (relative risk ratio, 2.4; P < .001) and placenta percreta (relative risk ratio, 10; P < .001), and their diameter is proportionate to the extent of invasiveness and major peripartum complications; estimated optimal cut-offs were 2 mm (area under the curve [AUC], 0.81; P = .04) for prediction of PAS and 3 mm (AUC, 0.81; P < .001) for prediction of placenta percreta and poor outcome. IFVs added incremental value to the overall MRI diagnostic ability for PAS.
Abnormal intraplacental vasculature at MRI among patients with PAS was first reported by Derman et al (11) in three patients; they described many tortuous, dilated (≥6 mm) vessels with a chaotic distribution, without commenting on their origin. Ueno et al (12) assumed that the dilated vascular spaces seen at MRI of invasive placentas represent maternal vessels originating from the myometrium at the site of the invasion. We found that these abnormal vessels followed a characteristic course, from the cord or chorionic and/or subchorionic placental surface (fetal origin) deep into the placenta, and we confirmed this observation upon surgicopathologic examination. In our study, the presence of even one such vessel was diagnostic of PAS. The cut-off diameter of these aberrant vessels in our study (2 mm) was much lower than that reported by Derman et al (11); the larger number of patients in our study may have allowed a more confident statistical analysis.
Type (artery vs vein) of IFVs could not be determined at MRI. Microscopic examination of placental specimens demonstrated two vessels running in parallel within the large IFV trunks, apparently representing a pair of artery and vein; one vessel, probably the vein, was greatly dilated at the expense of the other; however, fetal arteries cannot be distinguished from veins when they enter the placenta because the walls of both contain a large amount of extracellular matrix and no elastic lamina between smooth muscle cells of the media, making their histologic identification at light microscopy nearly impossible.
Sensitivity of the IFV sign for PAS detection was much higher (72%) compared with previously reported MRI findings of abnormal intraplacental vascularity (46.0%–53.8%), without affecting specificity (18). The IFVs provide a more objective means for diagnosing invasiveness compared with other common MRI features of PAS, such as intraplacental T2-hypointense bands, explaining the high agreement rates for IFVs between readers in our study. IFVs were easily identified on both 1.5-T and 3.0-T images; inclusion of high-spatial-resolution images in our study protocol facilitated identification of IFVs.
IFVs were present in most women with PAS (109 of 126 [86%]) and in all but one with placenta percreta (67 of 68 [98%]), with a 98.8% positive predictive value. In all but one of 17 women with PAS and negative results for IFVs (16 of 17 [94%]), placenta creta and/or increta was diagnosed. The high number of false-negative results (17 of 126 [13%]) explains the low negative predictive value of IFVs for PAS detection in our study sample; in all 17 women, however, the peripartum course was uneventful.
IFVs were present in four of 29 women with normal placentas; in all four women, IFVs were subtle, with a diameter of 2 mm or less, and in three women were confined to the inner placental half. Gestational age of these women was 31, 32, 35, and 37 weeks, respectively. To our knowledge, there is no clear evidence that placental maturation considerably affects intraplacental vascularity, and extensive intraplacental vascularity is not a normal finding in any pregnancy trimester at either MRI or pathologic examination; the extent of this aberrant intraplacental vasculature seems to correlate with the degree of invasiveness (15).
We found that IFV diameter is proportionate to the depth of invasiveness, with a 3-mm cut-off value for the diagnosis of placenta percreta. IFV diameter was associated with important peripartum events, including intraoperative blood loss, hysterectomy, and bladder repair. IFV diameter cut-off values of 3 mm and 2.9 mm were predictive of intraoperative massive bleeding and hysterectomy with bladder repair, with 76%–83% sensitivity and 78%–91% negative predictive values. To our knowledge, this information is reported for the first time; IFV diameter at prenatal placental MRI seems to be a prognostic marker of poor clinical outcomes for patients with PAS and allows more careful planning of delivery.
In our study, addition of the IFV sign to all other recorded MRI features of PAS improved the overall MRI diagnostic performance for detection of invasive placenta, suggesting that its routine implementation may contribute to the care of women with PAS.
Our study had several limitations. First, MRI examinations were performed with two different units; increased signal-to-noise ratio of 3.0-T images may have influenced image interpretation. However, the predictive ability of IFV for detection and depth of PAS was significant for both groups. Second, our study sample was already at high risk for PAS; therefore, results may not be the same when applied to the general population. Future investigation with larger numbers of normal placentas will be required to confirm the rigor of the IFV sign. Third, experienced radiologists in gynecologic and obstetric imaging conducted this study; external validation of the performance of the IFV sign is necessary to confirm reproducibility of results. Finally, many women with PAS had their uterus preserved (n = 70), and we relied on intraoperative information for these women, subject to surgical expertise. However, it has recently been agreed that clinical diagnosis is acceptable as the reference standard for PAS diagnosis (3).
In conclusion, the presence of one or more intraplacental fetal vessels with a diameter of 2 mm or greater, deep within the placenta, is an accurate and independent MRI predictor of invasiveness in pregnant women suspected of having placenta accreta spectrum; intraplacental fetal vessels with a diameter of 3 mm or greater are highly suggestive of placenta percreta and a complicated peripartum course.
Acknowledgments
We thank Kyrillos Sarris, MD, and Elissaios Balis for the provided medical illustration.
Author Contributions
Author contributions: Guarantors of integrity of entire study, C.B., K.Z., M.T., G.D., A.L., L.A.M.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, C.B., A.E.K., K.Z., A.A., M.T., G.D., M.E.N., A.L., E.A.M.; clinical studies, C.B., K.Z., A.A., S.F., M.T., G.D., M.E.N., A.L., L.A.M.; experimental studies, M.T., A.L.; statistical analysis, C.T., A.L.; and manuscript editing, C.B., A.E.K., K.Z., A.A., S.F., M.T., A.L., E.A.M., L.A.M.
References
- 1. . Placenta accreta: pathogenesis of a 20th century iatrogenic uterine disease. Placenta 2012;33(4):244–251.
- 2. . Maternal morbidity in patients with morbidly adherent placenta treated with and without a standardized multidisciplinary approach. Am J Obstet Gynecol 2015;212(2):218.e1–218.e9.
- 3. . Evidence-based guidelines for the management of abnormally invasive placenta: recommendations from the International Society for Abnormally Invasive Placenta. Am J Obstet Gynecol 2019;220(6):511–526.
- 4. . Placenta accreta spectrum: a hysterectomy can be prevented in almost 80% of cases using a resective-reconstructive technique. J Matern Fetal Neonatal Med 2020. 10.1080/14767058.2020.1716715. Published online January 26, 2020.
- 5. . FIGO consensus guidelines on placenta accreta spectrum disorders: Prenatal diagnosis and screening. Int J Gynaecol Obstet 2018;140(3):274–280.
- 6. . Magnetic resonance imaging in 300 cases of placenta accreta: surgical correlation of new findings. Acta Obstet Gynecol Scand 2005;84(8):716–724.
- 7. . Accuracy of ultrasonography and magnetic resonance imaging in the diagnosis of placenta accreta. Obstet Gynecol 2007;109(1):203; author reply 203.
- 8. . MR imaging of the placenta: what a radiologist should know. Abdom Imaging 2013;38(3):573–587.
- 9. . Prenatal identification of invasive placentation using magnetic resonance imaging: systematic review and meta-analysis. Ultrasound Obstet Gynecol 2014;44(1):8–16.
- 10. . MRI features predictive of invasive placenta with extrauterine spread in high-risk gravid patients: a prospective evaluation. AJR Am J Roentgenol 2018;211(3):701–711.
- 11. . MRI of placenta accreta: a new imaging perspective. AJR Am J Roentgenol 2011;197(6):1514–1521.
- 12. . Novel MRI finding for diagnosis of invasive placenta praevia: evaluation of findings for 65 patients using clinical and histopathological correlations. Eur Radiol 2014;24(4):881–888.
- 13. . MRI of placenta accreta, placenta increta, and placenta percreta: pearls and pitfalls. AJR Am J Roentgenol 2017;208(1):214–221.
- 14. . Society of Abdominal Radiology (SAR) and European Society of Urogenital Radiology (ESUR) joint consensus statement for MR imaging of placenta accreta spectrum disorders. Eur Radiol 2020;30(5):2604–2615.
- 15. . Stripped fetal vessel sign: a novel pathological feature of abnormal fetal vasculature in placenta accreta spectrum disorders with MRI correlates. Placenta 2019;85:74–77.
- 16. . MRI prognosticators for adverse maternal and neonatal clinical outcome in patients at high risk for placenta accreta spectrum (PAS) disorders. J Magn Reson Imaging 2019;50(2):602–618.
- 17. . The value of specific MRI features in the evaluation of suspected placental invasion. Magn Reson Imaging 2007;25(1):87–93.
- 18. . Diagnostic accuracy of magnetic resonance imaging in detecting the severity of abnormal invasive placenta: a systematic review and meta-analysis. Acta Obstet Gynecol Scand 2018;97(5):507–520.
- 19. . Placenta accreta spectrum: value of placental bulge as a sign of myometrial invasion on MR imaging. Abdom Radiol (NY) 2019;44(7):2572–2581.
- 20. . Suspected invasive placenta: evaluation with magnetic resonance imaging. Eur Radiol 2014;24(12):3150–3160.
- 21. . Invasive placenta previa: placental bulge with distorted uterine outline and uterine serosal hypervascularity at 1.5T MRI - useful features for differentiating placenta percreta from placenta accreta. Eur Radiol 2018;28(2):708–717.
- 22. . A proposal for standardized magnetic resonance imaging (MRI) descriptors of abnormally invasive placenta (AIP) - From the International Society for AIP. Diagn Interv Imaging 2019;100(6):319–325.
- 23. ; FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel. FIGO classification for the clinical diagnosis of placenta accreta spectrum disorders. Int J Gynaecol Obstet 2019;146(1):20–24.
- 24. . Correlation of probability scores of placenta accreta on magnetic resonance imaging with hemorrhagic morbidity. J Clin Anesth 2016;34:261–269.
- 25. . Maternal and neonatal outcomes following abnormally invasive placenta: a population-based record linkage study. Acta Obstet Gynecol Scand 2017;96(11):1373–1381.
- 26. . Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: Wiley, 1981; 38–46.
- 27. . Practical Statistics for Medical Research. London, England: Chapman & Hall, 1991; 404.
- 28. . Index for rating diagnostic tests. Cancer 1950;3(1):32–35.
- 29. . Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44(3):837–845.
Article History
Received: Feb 2 2020Revision requested: Apr 8 2020
Revision received: Sept 14 2020
Accepted: Oct 15 2020
Published online: Nov 24 2020
Published in print: Feb 2021
