Ultrasound Imaging of Placenta AccretaWith MR Correlation
Winnie A. Mar, MD,* Senta Berggruen, MD,Þ Ukamaka Atueyi, MD,* Simran Sekhon, MD,þSteven A. Garzon, MD,§ Martha G. Knuttinen, MD, PhD,* and John P. McGahan, MDþ
Abstract: Placenta accreta is abnormal placental adherence or inva-sion of the myometrium or extrauterine structures. It is increasing inincidence because of increasing number of cesarean sections and isone of the main causes of excessive postpartum hemorrhage. Rec-ognition of this entity is crucial because improved outcomes havebeen shown when the antenatal diagnosis of placenta accreta is made.Ultrasound is the first-line tool; magnetic resonance imaging (MRI) iscomplementary. Ultrasound and MRI features and MRI protocols willbe reviewed.
Key Words: placenta accreta, ultrasound, magnetic resonanceimaging
(Ultrasound Quarterly 2015;31:23Y33)
P lacenta accreta is an important cause of excessive post-partum hemorrhage. It can result in significant maternal
morbidity and mortality. Because of the increased number ofcesarean sections, the incidence of placenta accreta is increasing.Improved outcomes have been shown when the antenatal diag-nosis has been made, a scheduled cesarean delivery can beperformed, and when patients are managed at a tertiary referralcenter with a multidisciplinary team.1
To facilitate early antenatal diagnosis, we shall discussthe magnetic resonance (MR) and ultrasound imaging find-ings, MR protocol, and pitfalls in imaging. A brief discussionon pathophysiology and treatment will be presented.
DEFINITIONYPATHOLOGYPlacenta accreta has been used as a general term, en-
compassing three different degrees of placental invasion: pla-centa accreta vera, placenta increta, and placenta percreta (Fig. 1).Placenta accreta vera is the mildest form in which chorionic villiare attached to, but do not invade, the myometrium. In placentaincreta, the chorionic villi partially extend through the myome-trium. In placenta percreta, placental tissue extends completelythrough the myometrium, to the uterine serosa, and may invadeadjacent structures, such as the bladder, parametrium, vessels,ureters, colon, or rectum.
The placenta is composed of maternal and fetal com-ponents. The decidua basalis is a maternal component arisingfrom the endometrium, whereas the chorionic villi are a fetalcomponent. With a normal placenta, the chorionic villi attachonto the decidua basalis. Placenta accreta occurs when a defectin the decidua basalis allows the chorionic villi to attach di-rectly to or invade into the myometrium.
CLINICAL IMPLICATIONSThere is a wide range in reported incidence of pla-
centa accreta between approximately 1 in 500 to 7 in 10,000births,2,3 which could be attributed to the lack of consensus inthe diagnostic criteria. Histologic diagnosis is the referencestandard; however, this likely underestimates the true inci-dence of placenta accreta because there are many clinical casesof adherent placenta that are not diagnosed as placenta accretaor its variants.
The two most important risk factors for placenta accretaare prior cesarean section and placenta previa. The incidenceof placenta accreta has increased dramatically during the past2 decades, which corresponds to the increasing rate of cesareandelivery. Wu et al2 reported a rise in the rate of cesarean de-livery from 12.5% in 1982 to 23.5% in 2002. According to theCenters for Disease Control and Prevention, the rate of cesareandelivery in 2011 is 33%. Prior cesarean delivery increases therisk of placenta accreta by almost 9-fold.2 The risk of placentaaccreta is increased when the placenta is implanted over thecesarean section scar.3
Placenta previa occurs when the placenta partially orcompletely covers the cervical os and by itself is also a leadingcause of antepartum hemorrhage. The risk of placenta accretawith placenta previa in combination with prior cesarean deliv-ery seems to be synergistic and is 11% for one prior cesareandelivery, increasing to 40% and 61% with three and fourcesarean deliveries, respectively.4,5
Other procedures such as myomectomy, dilation and cu-rettage, lysis of adhesions for Asherman syndrome, or uterusseptum removal also increase risk.6 Other risk factors includeadvanced maternal age, multiparity, submucosal uterine fibroids,smoking, and chronic hypertension.
IMAGINGUltrasound is the first-line tool for screening, which is
usually performed at 18 to 20 weeks’ gestation. Gray-scaleultrasound has been shown to have a sensitivity of 77% to87%, a specificity of 96% to 98%, a positive predictive valueof 65% to 93%, and a negative predictive value of 98% for
REVIEWARTICLE
Ultrasound Quarterly & Volume 31, Number 1, March 2015 www.ultrasound-quarterly.com 23
Received for publication July 25, 2014; accepted October 21, 2014.*Department of Radiology, University of Illinois at Chicago; and †Department
of Radiology, Northwestern Memorial Hospital, Chicago, IL; ‡Departmentof Radiology, University of California-Davis, Sacramento, CA; and §De-partment of Pathology, University of Illinois at Chicago, Chicago, IL.
The authors declare no conflict of interest.Reprints: Winnie A. Mar, MD, 1740 W Taylor St Room 2483 (MC 931),
Chicago, IL 60612 (e<mail: [email protected]).Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.
diagnosis of accreta.7,8 The presence of multiple findings in-creases the specificity of ultrasound.7
Currently, there is insufficient evidence for routine useof MRI on all suspected cases of placenta accreta. At someinstitutions, MRI is often performed when accreta is suspectedon ultrasound.9,10 Magnetic resonance imaging may betterdelineate the degree of placental invasion and may changesurgical management.11 A prospective review of 300 cases ofplacenta accreta showed that MRI modified detection of in-vasion levels in approximately 30% of cases and could accu-rately depict parametrial invasion11 when ultrasound could not.Another recent study with 13 patients showed MRI to have agreater sensitivity and specificity compared with ultrasound;however, transvaginal ultrasound was not used in this study.12
Earlier literature suggested that MRI was superior toultrasound only in evaluating posterior placentas.13,14 Otherstudies have shown that these modalities are comparable,8,15
and are complementary, because one modality may show thecorrect finding when the other modality is equivocal.15
NORMAL ULTRASOUND APPEARANCEOF THE PLACENTA
Prenatal ultrasound is generally performed transabdo-minally; however, if needed, transvaginal ultrasound is safeand can also be performed for thorough evaluation of the loweruterine segment.16
The placenta is uniformly hyperechoic with respect tothe hypoechoic myometrium. The retroplacental clear space(Fig. 2A) is a hypoechoic linear structure at the junction of theplacenta and the myometrium and is thought to correspond todilated vessels in the decidua basalis.17
In the third trimester, there is increased heterogeneitycaused by vascular lakes and calcification. Regular retropla-cental vascularity should be present, with occasional vesselsextending into the placenta (Fig. 2B) corresponding to thematernal spiral arteries entering between the cotyledons.
ULTRASOUND FINDINGSPlacental lacunae are multiple irregular vascular spaces in
the placenta. They are the most sensitive sign with the highestpositive predictive value for placenta accreta (Fig. 3).7,18 Theyhave been associated with placenta accreta in 87.5% of high-risk patients,19 and the likelihood increases with a greaternumber of lacunae.18 Lacunae should be differentiated fromplacental lakes, which are rounder, with no flow, and are seen innormal placentas. If numerous vascular lacunae are seen, theplacenta should be evaluated for other supporting findings.
The location of placental implantation is important toassess because accreta was found in 29% of cases with im-plantation over the cesarean section scar.3 Clinical historyis also important, and particular attention should be paid toknown areas of uterine surgery, such as myomectomy, shouldthe placenta be implanted there.6
Lobularity and irregularity of the placental myometrialjunction (Fig. 4A) has also been described with placentaaccreta9 and seems to be analogous to the irregularity and focalbulging seen on MRI.
FIGURE 2. Two different patients with a normal placenta. A, Transabdominal ultrasound image of the placenta shows a thinhypoechoic line between the placenta and myometrium corresponding to the retroplacental clear space (arrows). Nonvisualizationof this space can be seen in normal placentas (arrowhead) and is nonspecific in the absence of other findings. B, Transabdominaltransverse ultrasound with color Doppler showing a normal vascular pattern of the placenta (arrows).
FIGURE 1. Classification of placenta accreta. Placenta accretaveraY chorionic villi are attached to the myometrium but donot invade the myometrium. Placenta incretaY chorionic villiinvade the myometrium. Placenta percretaY chorionic villipenetrate through the entire myometrium or beyond.
Mar et al Ultrasound Quarterly & Volume 31, Number 1, March 2015
24 www.ultrasound-quarterly.com * 2015 Wolters Kluwer Health, Inc. All rights reserved.
In a small series, thinning of the myometrium to less than1 mm in the presence of large placental lacunae was reportedto have a sensitivity of nearly 100% and a specificity of 72%to 79%.20 Focal thinning and loss of the myometrium areuseful when seen in combination with other findings such aslacunae and irregularity of the placental myometrial junction.
As the retroplacental clear space is believed to representdecidua basalis, nonvisualization of this space is thought to beassociated with placenta accreta. However, loss of clear spacewhen observed in isolation is not a reliable diagnostic sign. Ahigh false-positive rate of up to 50% has been reported,7,21 asthe clear space is frequently not visible in anterior placentas,17
FIGURE 3. Lacunae and hypervascularity of the uterine-bladder interface in pathologically proven placenta increta in a 27-year-oldwoman at 31 weeks’ gestation with 1 prior cesarean section. A, Transabdominal transverse ultrasound with irregular vascularlacunae (arrows) within the placenta. B, Transvaginal sagittal ultrasound performed for better evaluation of the uterineYbladder wallinterface with color Doppler. Interface hypervascularity (white arrows) and turbulent flow in the lacunae (black arrow). Bladder (*).
FIGURE 4. Placental bulging and loss of the uterine bladder interface in a 33-year-old woman at 17 weeks’ gestation, with 3 priorcesarean sections, presenting with a gush of blood from her vagina. Surgery and pathology showed placenta percreta involving thelower uterine segment anteriorly and bladder invasion. A, Transabdominal sagittal ultrasound showing placenta previa and bulgingof the placenta into the posterior lower uterine segment and cervix (arrows). B, Transabdominal sagittal ultrasound showing bulgingof the placenta into the bladder with loss of the uterine bladder interface (arrow) and large lacunae (L). C, Transabdominal sagittalcolor Doppler ultrasound showing extension of placental vessels to the bladder (arrow) and flow within the lacunae (curved arrow).D, Transvaginal transverse ultrasound showing myometrial thinning (white arrow) to less than 1 mm, with large placental lacunae (L).E, Axial SSFSE T2 MRI showing interruption of the outer myometrial line representing uterine serosa (arrowhead). Hypointensebands along with marked myometrial thinning, without loss of the outer myometrial serosal line elsewhere (arrows). Laminaria wereplaced within the cervix for possible dilation and curettage and because of vaginal bleeding (curved arrow). F, Sagittal SSFSE T2MRI showing percreta at the anterior lower uterine segment cesarean section scar with bulging and interruption of the myometrium(thin arrow). Bulging of the placenta at the bladder with loss of overlying myometrium and bladder wall thickening (thick arrow).Myometrium (arrowheads).
Ultrasound Quarterly & Volume 31, Number 1, March 2015 Imaging of Placenta Accreta
* 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ultrasound-quarterly.com 25
and in the third trimester. This clear space is optimally iden-tified in posterior, rather than anterior, placentas because ofoptimization of the focal zone of the transducer. However, withcurrent technology, this clear space can be more easily iden-tified then noted in prior publications.
Disruption of the uterine serosaYbladder line (Fig. 4B)has a good specificity but poor sensitivity.7,18,22 Endovaginalscanning should also be performed with a partially full bladderto facilitate visualization of the uterine bladder junction.13
Compared with gray-scale ultrasound, power and colorDoppler have not been shown to improve the diagnostic sen-sitivity.23 However, several findings may be helpful. Placentaaccreta may be suggested by hypervascularity of the uterineserosaYbladder interface and markedly dilated vessels overthe peripheral subplacental zone.18,24 Transabdominal, or evenbetter endovaginal, ultrasound with a partially filled bladdermay be useful to demonstrate loss of the uterine serosal/bladder interface but also demonstrate vessels protruding intothe bladder in cases of placenta percreta (Fig. 4C). However,along the same lines, visualization of vascularity in the blad-der wall does not equate to placenta percreta. Bladder varices,which occur in the late second and early third trimester ofpregnancy, may be a potential pitfall (Fig. 5).23 Spectral Dopplercan also be used to assess for turbulent blood flow throughlacunae as a helpful sign. Focal discontinuity of Doppler flowmay be seen at the location of placental invasion.9 Extensionof Doppler flow from the placenta to extrauterine structuresmay also predict abnormal placentation.19
In the first trimester, findings that suggest placenta accretaare low implantation of the gestational sac, lacunae,25 and ir-regularity of the placental myometrial interface.26 It is importantto recognize first trimester findings because extensive hemor-
rhage during dilation and curettage and first trimester uterinerupture have been reported.27 There has also been an increase inthe frequency of cesarean section ectopic pregnancies (Fig. 6),which are thought to be more aggressive than placenta accretabecause there is early invasion into the myometrium.28
NORMAL MRI APPEARANCE OF THE PLACENTAThe placenta is normally homogeneous and moderately
T2 hyperintense.29 The myometrium can be intermediate orhigh signal intensity on T2-weighted images.9,30 Myometrial
FIGURE 5. Bladder varices leading to false-positive MRI in a 43-year-old woman at 35 weeks’ gestation with 3 prior cesarean sections,presenting with vaginal bleeding and complete previa. Sagittal transabdominal gray-scale (A) and color Doppler (B) image showsthick retroplacental myometrium (arrowheads in A) homogeneous placenta (* in A) and preserved retroplacental hypoechoic zone(arrow in B). C, Sagittal SSFSE T2MRI demonstrates apparent placental (P) invasion into the bladder (black arrows) with false impressionof percreta. Note faint hypointense flow voids (white arrows) of the bladder varices. D and E, Transabdominal sagittal ultrasoundwithout and with color Doppler shows varices (white arrow) in the bladder wall, leading to false-positive MRI. B indicates bladder;M, myometrium; *, placenta.
FIGURE 6. Cesarean section scar ectopic in a 41-year-oldwoman with 1 prior cesarean section presenting with a positivepregnancy test. Sagittal endovaginal ultrasound demonstratingearly pregnancywith a yolk sac (arrow). This pregnancy is locatedin the lower uterine segment at the site of prior cesarean section.
Mar et al Ultrasound Quarterly & Volume 31, Number 1, March 2015
26 www.ultrasound-quarterly.com * 2015 Wolters Kluwer Health, Inc. All rights reserved.
contractions consisting of transient, focal, ill-defined T2hypointense thickening can also be seen.31
The myometrium typically has a trilaminar appearanceon T2-weighted or bSSFP images (Fig. 7A).30 The thin innerhypointense layer corresponds to the decidual myometrial in-terface and represents the retroplacental clear space.32 A thickerhyperintense myometrial layer contains the arcuate vessels. Thethin outer hypointense layer corresponds to the serosa.30 Theplacental myometrial interface is best seen on single shot fastspin echo (SSFSE) techniques T2 images33 and is less distincton balanced steady state free precession (bSSFP) images.
Normal regular subplacental vessels are seen within themyometrium (Fig. 7B).9 Occasionally, a few small vessels alsocross the placenta, typically near the umbilical cord insertion.9
With increasing gestational age, the placenta becomesslightly more hypointense and slightly more lobular and het-erogeneous.29 In the third trimester, the cotyledons can be
visualized, with intervening fine septations that are regularlyspaced (Fig. 7C).9
MRI PROTOCOLTypically, patients are imaged supine; however, patients
may also be imaged in the left lateral decubitus position if theyprefer. The patient’s feet should be placed first into the scannerto minimize claustrophobia. Although not standard practice,oxygen can be given via nasal cannula, if needed, to decreasefetal motion.9 Breath holding should be performed whenpossible. The bladder should be partially full because it aids invisualization of the myometrial/bladder interface (Fig. 8).
T2-weighted images are the most important sequences inimaging placenta accreta and should be obtained in all 3 planes.Imaging should be monitored by a radiologist. Additional fo-cused oblique axial or coronal sequences perpendicular to the
FIGURE 7. Normal placenta. A, Sagittal SSFSE T2 MRI. The three layers of the normal myometrium are seen. The hypointense outermyometrial layer (arrows) and inner myometrial layer (arrowheads) surround the hyperintense inner layer. B, Sagittal FSE T2 MRIshowing normal subplacental vascularity (arrows). C, Coronal SSFSE T2 MRI shows a homogeneous placenta with normal thinseptations (arrows).
FIGURE 8. The importance of having a partially filled bladder in a 42-year-old woman at 30 weeks’ gestation with placenta percretaand 1 prior cesearean section. Magnetic resonance imaging performed because of suspicion of percreta on routine ultrasound. Atsurgery, bladder invasion and percreta of the anterior lower uterine segment was found. Sagittal SSFSE T2MRI with bladder partiallydecompressed (A) and full (B). Placental nodularity (thick arrow) indenting the bladder is more apparent with a full bladder (B).There is loss of the myometrium at the anterior lower uterine segment (thin arrow) and over the bladder (thick arrow) where there isalso bulging. Myometrium (arrowhead).
Ultrasound Quarterly & Volume 31, Number 1, March 2015 Imaging of Placenta Accreta
* 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ultrasound-quarterly.com 27
placental myometrial interface can be performed at questionableareas if needed. Fast sequences such as SSFSE (Single Shot FSEon General Electric scanners, HASTE on Siemens scanners,Single Shot TSE on Philips scanners)34 and bSSFP sequences(FIESTA on General Electric scanners, TrueFISP on Siemensscanners, and b-FFE on Philips scanners) have decreased fetaland maternal motion artifacts.30 Longer FSE T2 sequences havebetter resolution and signal-to-noise ratio, however, are moresusceptible to motion artifact because of their long acquisitiontimes.31 Parallel imaging techniques can also decrease motionartifact while retaining resolution.33
At least 1 fat-saturated T1 in phase gradient echo se-quence can be used to show hemorrhage as T1 hyperintense.33
T1-weighted images are not generally used for evaluating pla-cental invasion because the placenta and myometrium are similarin signal intensity.31,33
Slice thickness should range from4 to 6mm.35Y37 Thinnerslices can be used when focusing on an area of suspicion.Partial volume averaging is an issue that can also be avoided byverifying a finding on more than 1 plane of imaging.
Intravenous contrast has been used at some institutionsto better delineate the placental myometrial interface. Giventhat contrast is excreted into the amniotic fluid and swallowedby the fetus, its use is controversial because the long-term risksto the fetus are unknown.8 A recent study in mice revealedgadolinium-based contrast use showed only very small amountsof contrast in the fetus and amniotic fluid, and none was de-
tectable after 24 hours in the fetus and 48 hours in the amnioticfluid.38Y40 The ACR guidance document for safe MRI practicesrecommends that intravenous gadolinium should be avoidedduring pregnancy and should only be used if absolutely es-sential.40 At our institutions, contrast is not routinely used.Some authors report that they only use contrast in select cases,such as when delivery will occur a few days after contrast ad-ministration or if there will be pregnancy termination.9
Diffusion weighted imaging (DWI) is a relatively noveltechnique applied to evaluation of placenta accreta.41 As theplacenta and myometrium are often similar in T2 signal in-tensity, DWI can also be used to better delineate the placentalmyometrial border (Fig. 9). However, the spatial resolutionand signal-to-noise ratio of this sequence are generally low;B0/B1000 fusion images with color coding can be used tobetter show the placental/myometrial interface.41 We alsoperform B500 images and have found that these images haveimproved signal-to-noise ratio, although the contrast betweenthe myometrium and placenta is less. In addition, DWI is alsoless susceptible to motion artifact.42
MRI FINDINGS OF PLACENTA ACCRETAThe most useful MRI findings are dark intraplacental
bands, placental heterogeneity, and placental bulging.37,43 Grossextrauterine placental extension and tenting or nodularity of thebladder are also helpful when seen.9 Early MRI criteria of an
FIGURE 9. Utility of DWI to differentiate myometrium from placenta in a 29-year-old woman at 28 weeks’ gestation with 2 priorcesarean sections and placenta percreta confirmed at delivery and pathology. Magnetic resonance imaging performed because ofsuspicion of percreta on routine ultrasound. A, Sagittal FSE T2 MRI showing the placenta (arrow) and the myometrium (*), whichhave a similar signal intensity, making differentiation difficult. B, Sagittal DWI MRI at B500. The hyperintense placenta (arrow) iseasier to differentiate from the hypointense myometrium (*). No myometrium is seen anterior to the placenta, consistent withpercreta. C, Sagittal DWI MRI at B1000 shows slightly increased differentiation between the placenta (*) and myometrium (arrow),however, has more noise. Sagittal (D) and axial (E) SSFSE T2 MRI. Percreta is confirmed as a focal bulge (arrows) without overlyingmyometrium. A dark intraplacental band is also seen (curved arrow). Preserved myometrium (arrowheads).
Mar et al Ultrasound Quarterly & Volume 31, Number 1, March 2015
28 www.ultrasound-quarterly.com * 2015 Wolters Kluwer Health, Inc. All rights reserved.
indistinct myometrium and loss of the T2 dark uteroplacentalinterface have been found to be less helpful.35
PLACENTAL HETEROGENEITY/DARKINTRAPLACENTAL BANDS
Placental heterogeneity is usually secondary to dark bands.In 1 series, all patients with placental invasion had moderate ormarked placental heterogeneity.43
Dark bands may be related to increased vascularity orplacental infarcts/fibrin deposition based on the appearanceon images. These dark bands represent abnormal placentalvascularity if they correspond to hyperintense structures onbSSFP images (Fig. 10).35 If they remain hypointense onbSSFP images, they are thought to represent fibrin deposition inthe placenta and were seen to correspond histologically toplacental hemorrhage and infarction.12 However, normal het-erogeneity may be seen with increasing gestational age.29 Pla-cental heterogeneity also shows a high interobserver reliabilityin both experienced and inexperienced readers.44
As the placenta typically becomes more heterogeneousas pregnancy progresses, a mild to moderately heterogeneousplacenta in the late third trimester may be difficult to evaluatefor invasive placentation.26,43 Dark bands, in contrast to nor-mal heterogeneity in the late third trimester, typically extendfrom the placental myometrial interface (Fig. 11A), with arandom distribution and varying thickness.9
Alamo et al44 reported that dark intraplacental bandswere the single best MRI feature predicting placental invasion.In another study, the volume of dark intraplacental bands cor-related to increasing degrees of placental invasion.12 The sitesof the dark bands that represented hemorrhage and infarcts alsocorresponded to histologic areas of abnormal placentation.12 Inanother recent study, the greater experience of the reader cor-relates with improved specificity and sensitivity.44 However, thesensitivity for dark intraplacental bands was similar betweenjunior readers with less than 3 years of practical body MRIexperience and senior readers with greater than 5 years of bodyMRI experience.44
Some authors have stated that invasive placentation isunlikely when the placenta is homogeneous.9,31 However, there
have been instances of abnormal placentation in the absence oflacunae, which may correspond to dark bands.18,25 Certainly,invasive placentation is less likely in relatively homogeneousplacentas43; however, all imaging findings should be taken to-gether to come to a conclusion,31 as in Figure 12, where there isalso loss of the myometrium and lacunae seen on ultrasound.
INTERRUPTED MYOMETRIUMFocally interrupted myometrium has also been described
as a useful sign (Fig. 4E).31,33 A focally interrupted myo-metrial border was found to be the second most predictivefeature of invasive placentation in the study by Alamo et al44;however, accurate recognition of this finding showed the mostvariability between experienced and inexperienced readers.
PLACENTAL BULGINGLeyendecker et al31 suggest that placental bulging is the
most useful sign in isolation. They note that the contour of theplacenta should appear as a smooth arc; focal bulging that couldindicate abnormal placentation may disrupt this arc (Figs. 11, 12).Gentle lobulations may occur in normal placentas, but theplacental myometrial interface should not be disrupted.
Bulging of the lower uterine segment that was described byLax et al43 is also useful for diagnosis. Other authors have foundthis finding to be less helpful because it can be seen with normalpregnancies.44 There are two descriptions of bulging: (1) a diffusebulge in the uterine contour resulting in loss of the normal in-verted pear shape of the uterus (this can also be seen in normalpregnancies) or (2) a smaller focal bulging of the placenta into themyometrium.9 The latter has been reported to be more helpful.31
Focal nodular bulging into the bladder, analogous to thatseen on ultrasound, is also suspicious for placental invasion(Figs. 4F, 11B). Disruption of the normal zonal anatomy of thecervix also suggests invasion.31
EXTRAUTERINE PLACENTAL EXTENSIONInfiltration of the pelvic organs and tenting of the
bladder are findings of extrauterine invasion and are relativelyspecific when present.9
FIGURE 10. Differentiation between dilated intraplacental vessels and hemorrhage/infarct at 32 weeks’ gestation, with placentapercreta found in surgery and pathology, in a 37-year-old woman with 4 prior cesarean sections and suspicion of accreta on routineultrasound. A, Sagittal FSE T2 MRI shows a complete previa and multiple irregular dark bands (curved arrow and thin arrows).Nonvisualization of the myometrium anteriorly (thick arrows). Myometrium (arrowheads). B, Sagittal FIESTA MRI shows that one ofthe T2 dark bands is hyperintense (curved arrow), thus corresponding to a dilated vessel, whereas another dark band remains darkon FIESTA (thin arrows), corresponding to areas of fibrin deposition that may be related to hemorrhage or infarct. C, Transvaginalsagittal ultrasound shows multiple lacunae (arrows).
Ultrasound Quarterly & Volume 31, Number 1, March 2015 Imaging of Placenta Accreta
* 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ultrasound-quarterly.com 29
Magnetic resonance imaging has been reported to bebetter than ultrasound at depicting extrauterine involvementsuch as parametrial invasion.11
THINNING OF THEMYOMETRIUM/LOSS OF THEINNER MYOMETRIAL LAYER
Part or all of the normal trilaminar appearance of theplacenta may be lost in placenta accreta.30 However, loss ofthese layers may also be seen in normal placentation.31 Themyometrium becomes significantly thinned, often diffusely, inthe late third trimester.33 Thus, imaging in the second trimesteror early third trimester is preferred.44 There is also, typically,thinning of the myometrium over the spine.33
Thinning of the myometrium, when seen alone, maycontribute to false-positive examinations.9 Even in technicallyadequate examinations, the placental myometrial junction maybe lost, particularly in patients with prior cesarean delivery.33,43
When thinning of the myometrium is seen, the presence ofother imaging findings should be sought for diagnosis ofabnormal placentation.9
PITFALLSSuspicious findings should be confirmed in two planes
whenever possible.33 False-positives may occur when a find-ing is seen only on one plane because of the normal curvatureof the uterus.31
FIGURE 11. Dark bands and bulging in a 23-year-old woman at 25 weeks’ gestation with 3 prior cesarean sections andpathologically proven placenta increta. She presented with vaginal bleeding. A and B, Sagittal FSE T2MRI showing a heterogeneousplacenta with multiple irregular dark bands (thin arrows) many extending from the placental myometrial interface. There isbulging with loss of the anterior myometrium at the cesarean section scar (thick arrows). The placenta is nodular indenting theanterior-superior bladder in B (curved arrow). Myometrium (arrowheads).
FIGURE 12. Loss of the myometrium and necessity for ultrasound correlation in a 23-year-old woman at 27 weeks’ gestationwith 3 prior cesarean sections with pathologically confirmed placenta percreta anteriorly at the cesarean section scar. A, SagittalSSFSE T2 MRI showing that the placenta is only mildly heterogeneous, without dark bands. There is disruption of the normalsmooth arc of the placental myometrial contour, with a small bulge and focal interruption of the myometrium (arrow). Normalmyometrium (curved arrow) is hyperintense to the placenta (pl). B, Transabdominal ultrasound shows lacunae (arrow) that werenot well seen on MRI.
Mar et al Ultrasound Quarterly & Volume 31, Number 1, March 2015
30 www.ultrasound-quarterly.com * 2015 Wolters Kluwer Health, Inc. All rights reserved.
The differentiation between accreta and increta andbetween increta and percreta, without gross invasion intosurrounding structures, remains difficult.31,35,37,43 More ex-perienced readers had a sensitivity of 75% and a specificity of68% in differentiating the types of invasion in 1 study.44
These authors also showed that it was easier to differentiateplacenta percreta from placenta accreta and increta.44 Severalauthors have noted that the most difficult differentiation onimaging was between placenta accreta and increta.31,44
In addition, infarcts that develop in a normal matureplacenta can be misleading. In 1 study, dark intraplacentalbands seen before 30 weeks accurately corresponded to areasof abnormal placentation.12
MANAGEMENTOptimal management of placenta accreta involves a
multidisciplinary team approach with predelivery planning.Cesarean delivery is recommended before 36 weeks gesta-tional age to avoid complications. Cesarean hysterectomy,
without attempted removal of the placenta, has the lowestmorbidity and is the preferred treatment.45 Others havereported some success with conservative management withouthysterectomy; however, there was significant maternal mor-bidity.46,47 The American College of Obstetricians and Gy-necologists does not recommend this latter approach unlessthere is a strong desire for future pregnancy and the patient ishemodynamically stable and willing to accept the risks.16 In astudy that evaluated 167 women conservatively treated withretained placenta accreta, the majority of retained placentasspontaneously resorbed, whereas 10% needed hysterectomy.46
Ultrasound and MRI have also been used to monitorpatients after conservative treatment by leaving the abnormallyadherent placenta in the uterus.48 Ultrasound can be performedtransabdominally or with sonohysterography, showing theretained placenta attached to the myometrium and sometimeshaving Doppler vascularity.49Y51 Serial ultrasound used tomonitor conservatively treated patients shows a decreased sizeor the persistence of the retained placenta.47Magnetic resonance
FIGURE 13. Retained placenta percreta in a 26-year-old woman with a history of 3 prior cesarean sections and 2 prior dilationand curettage, most recent for a missed first trimester abortion presenting with menorrhagia and dysmenorrhea. A, Sagittaltransabdominal ultrasound shows a large heterogeneous mass in the lower uterine segment (arrows). B, Color and spectraltransverse Doppler shows large disorganized Doppler flow (arrows). C, Contrast-enhanced computed tomography shows the samemass causing a mass effect on the bladder, with large vessels extending to the superior bladder wall (arrows). D, MRI sagittalFSE T2 shows the retained percreta invading the thickened bladder wall (arrows).
Ultrasound Quarterly & Volume 31, Number 1, March 2015 Imaging of Placenta Accreta
* 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ultrasound-quarterly.com 31
imaging shows enhancement and T2 hyperintensity in the ma-jority of retained placenta (Fig. 13) and can help determinewhichare more likely to spontaneously pass by determining the surfacearea of attachment.52 A small study showed that an attachmentarea of less than a semicircle indicated that the retained placentawas more likely to pass spontaneously.52 Another study postu-lated that delayed enhancement between the retained placentaand myometrium was thought to correspond to decidua basalis,and thus, if this was absent, then the etiology of the retainedplacenta was accreta.53 A recent study showed that, in post-embolization patients, decreased placental enhancement in theearly postcontrast phase correlated with faster resorption.54
INTERVENTIONAL RADIOLOGYCurrent evidence is insufficient to firmly recommend
the endovascular techniques of intra-arterial balloon catheterplacement or embolization for reducing maternal morbidity.There are mixed results regarding preoperative endovascularplacement of internal iliac artery balloons.55 Some have re-ported success56 with decreased transfusion requirements anda shorter duration of surgery.57 However, others have reportedno difference in transfusion requirements58 or increased bloodloss after endovascular intervention.59 Although rare, com-plications do occur such as arterial aneurysm formation or legischemia. Collateral flow between the external iliac and fem-oral arteries with the internal iliac system could explain thelack of consistent success with internal iliac occlusion.60
Pelvic artery embolization has recently been reported to have aslightly higher success rate (90.7%) than balloon tamponade(84%), with only minor complications. However, this smallstudy encompassed only patients without antenatal diagnosisof invasive placentation.61
CONCLUSIONSPlacenta accreta is increasing in incidence likely be-
cause of the increased rates of cesarean delivery. The antenataldiagnosis of placenta accreta is important to decrease maternalmorbidity and mortality. Ultrasound remains the first-lineexamination. However, recent trends indicate a greater use ofplacental MRI as a complementary imaging modality. Inter-pretation of the MRI findings should be done in conjunctionwith review of the ultrasound findings. Multiple imagingfindings on both ultrasound and MRI should be then used tomake this diagnosis. The most reliable findings have beenshown to be placental lacunae on ultrasound and darkintraplacental bands and focal bulging on MRI.
ACKNOWLEDGMENTSThe authors thank Natalie Doolittle for the illustration
and David Botos for assistance with image editing.
REFERENCES1. Eller AG, Bennett MA, Sharshiner M, et al. Maternal morbidity in cases
of placenta accreta managed by a multidisciplinary care team comparedwith standard obstetric care. Obstet Gynecol. 2011;117(2 Pt 1):331Y337.
2. Wu S, Kocherginsky M, Hibbard JU. Abnormal placentation: twenty-yearanalysis. Am J Obstet Gynecol. 2005;192:1458Y1461.
3. Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for placentaprevia-placenta accreta. Am J Obstet Gynecol. 1997;177:210Y214.
4. Clark SL, Koonings PP, Phelan JPJ. Placenta previa/accreta and priorcesarean section. Obstet Gynecol. 1985;66:89Y92.
6. Al-Serehi A, Mhoyan A, Brown M, et al. Placenta accreta: an associationwith fibroids and Asherman syndrome. J Ultrasound Med.2008;27:1623Y1628.
7. Comstock CH, Love J, Joseph J, et al. Sonographic detection ofplacenta accreta in the second and third trimesters of pregnancy. Am JObstet Gynecol. 2004;190:1135Y1140.
8. Warshak C, Eskander R, Hull A, et al. Accuracy of ultrasonographyand magnetic resonance imaging in the diagnosis of placenta accreta.Obstet Gynecol. 2006;108:573Y581.
9. Baughman WC, Corteville JE, Shah RR. Placenta accreta: spectrum ofUS and MR imaging findings. Radiographics. 2008;28:1905Y1916.
10. Masselli G, Brunelli R, Casciani E, et al. Magnetic resonance imaging inthe evaluation of placental adhesive disorders: correlation with colorDoppler ultrasound. Eur Radiol. 2008;18:1292Y1299.
11. Jaraquemada JMP, Bruno CH. Magnetic resonance imaging in 300 casesof placenta accreta: surgical correlation of new findings. Acta ObstetGynecol Scand. 2005;84:716Y724.
12. Lim PS, Greenberg M, Edelson MI, et al. Utility of ultrasound andMRI in prenatal diagnosis of placenta accreta: a pilot study. AJR.2011;197:1506Y1513.
13. Levine D, Hulka CA, Ludmir J, et al. Placenta accreta: evaluation withcolor Doppler US, power Doppler US, and MR imaging. Radiology.1997;205:773Y776.
14. Maldjian C, Adam R, Pelosi M, et al. MRI appearance of placenta percretaand placenta accreta. Magn Reson Imaging. 1999;17:965Y971.
15. Dwyer BK, Belogolovkin V, Tran L, et al. Prenatal diagnosis of placentaaccreta sonography or magnetic resonance imaging? J Ultrasound Med.2008;27:1275Y1281.
17. McGahan JP, Phillips HE, Reid MH. The anechoic retroplacental area: apitfall in diagnosis of placentalYendometrial abnormalities duringpregnancy. Radiology. 1980;134:475Y478.
18. Finberg HJ, Williams JW. Placenta accreta: prospective sonographicdiagnosis in patients with placenta previa and prior cesarean section.J Ultrasound Med. 1992;11:333Y343.
19. Guy GP, Peisner DB, Timor-Tritsch IE. Ultrasonographic evaluation ofuteroplacental blood flow patterns of abnormally located and adherentplacentas. Am J Obstet Gynecol. 1990;163:723Y727.
20. Twickler DM, Lucas MJ, Balis AB, et al. Color flow mapping formyometrial invasion in women with a prior cesarean delivery. J MaternFetal Med. 2000;9:330Y335.
21. de Mendon0a LK. Sonographic diagnosis of placenta accreta. Presentationof six cases. J Ultrasound Med. 1988;7:211Y215.
22. Hoffman-Tretin JC, Koenigsberg MM, Rabin AA, et al. Placenta accreta.Additional sonographic observations. J Ultrasound Med. 1992;11:29Y34.
23. Chou MM, Ho ES, Lee YH. Prenatal diagnosis of placenta previa accretaby transabdominal color Doppler ultrasound. Ultrasound Obstet Gynecol.2000;15:28Y35.
24. Shih JC, Jaraquemada JMP, Su YN, et al. Role of three-dimensionalpower Doppler in the antenatal diagnosis of placenta accreta: comparisonwith gray-scale and color Doppler techniques. Ultrasound ObstetGynecol. 2009;33:193Y203.
25. Comstock CH. The antenatal diagnosis of placental attachment disorders.Curr Opin Obstet Gynecol. 2011;23:117Y122.
26. Ballas JJ, Pretorius D, Hull AA, et al. Identifying sonographic markersfor placenta accreta in the first trimester. J Ultrasound Med.2012;31:1835Y1841.
27. Dabulis SA, McGuirk TD. An unusual case of hemoperitoneum: uterinerupture at 9 weeks gestational age. J Emerg Med. 2007;33:285Y287.
28. Ash A, Smith A, Maxwell D. Caesarean scar pregnancy. BJOG : aninternational journal of obstetrics and gynaecology. 2007;114:253Y263.
29. Blaicher W, Brugger PC, Mittermayer C, et al. Magnetic resonanceimaging of the normal placenta. Eur J Radiol. 2006;57:256Y260.
30. Kim JA, Narra VR. Magnetic resonance imaging with true fast imagingwith steady-state precession and half-Fourier acquisition single-shot
Mar et al Ultrasound Quarterly & Volume 31, Number 1, March 2015
32 www.ultrasound-quarterly.com * 2015 Wolters Kluwer Health, Inc. All rights reserved.
turbo spin-echo sequences in cases of suspected placenta accreta. ActaRadiol. 2004;45:692Y698.
31. Leyendecker JR, Dubose M, Hosseinzadeh K, et al. MRI ofpregnancy-related issues: abnormal placentation. AJR Am J Roentgenol.2012;198:311Y320.
32. Elsayes KM, Trout AT, Friedkin AM, et al. Imaging of the placenta:a multimodality pictorial review. Radiographics. 2009;29:1371Y1391.
33. Masselli G, Gualdi G. MR imaging of the placenta: what a radiologistshould know. Abdom Imaging. 2013;38:573Y587.
34. Yamashita YY, Namimoto TT, Abe YY, et al. MR imaging of the fetus bya HASTE sequence. AJR Am J Roentgenol. 1997;168:513Y519.
35. Derman AY, Nikac V, Haberman S, et al. MRI of placenta accreta:a new imaging perspective. AJR Am J Roentgenol.2011;197:1514Y1521.
36. Allen B, Leyendecker J. Placental evaluation with magnetic resonance.Radiol Clin N Am. 2013;51:955Y966.
37. Teo TH, Law YM, Tay KH, et al. Use of magnetic resonance imaging inevaluation of placental invasion. Clin Radiol. 2009;64:511Y516.
38. Katzberg R, McGahan J. Will gadolinium-enhanced MR imaging beuseful in assessment of at-risk pregnancies? Radiology.2011;258:325Y326.
39. Muhler MR, Clement O, Salomon LJ, et al. Maternofetal pharmacokineticsof a gadolinium chelate contrast agent in mice. Radiology.2011;258:455Y460.
40. Kanal E, Barkovich AJ, Bell C, et al. ACR guidance document for safeMR practices: 2007. AJR Am J Roentgenol. 2007;188:1447Y1474.
41. Morita S, Ueno E, Fujimura M, et al. Feasibility of diffusion-weightedMRI for defining placental invasion. J Magn Reson Imaging.2009;30:666Y671.
42. Bonel HM, Stolz B, Diedrichsen L, et al. Diffusion-weighted MR imagingof the placenta in fetuses with placental insufficiency. Radiology.2010;257:810Y819.
43. Lax A, Prince MR, Mennitt KW, et al. The value of specific MRI featuresin the evaluation of suspected placental invasion. Magn Reson Imaging.2007;25:87Y93.
44. Alamo L, Anaye A, Rey J, et al. Detection of suspected placental invasionby MRI: do the results depend on observer’ experience? Eur J Radiol.2013;82:e51Ye57.
45. Eller AG, Porter TF, Soisson PP, et al. Optimal management strategies forplacenta accreta. BJOG. 2009;116:648Y654.
46. Sentilhes L, Ambroselli C, Kayem G, et al. Maternal outcome afterconservative treatment of placenta accreta. Obstet Gynecol.2010;115:526Y534.
47. Bretelle F, Courbiere B, Mazouni C, et al. Management of placentaaccreta: morbidity and outcome. Eur J Obstet Gynecol Reprod Biol.2007;133:34Y39.
48. Torrenga B, Huirne JA, Bolte AC, et al. Postpartummonitoring of retainedplacenta. Two cases of abnormally adherent placenta. Acta ObstetGynecol Scand. 2013;92:472Y475.
49. Zalel Y, Cohen SB, Oren M, et al. Sonohysterography for the diagnosisof residual trophoblastic tissue. J Ultrasound Med. 2001;20:877Y881.
50. Zalel Y, Gamzu R, Lidor A, et al. Color Doppler imaging in thesonohysterographic diagnosis of residual trophoblastic tissue. J ClinUltrasound. 2002;30:222Y225.
51. van den Bosch T, Daemen A, Van Schoubroeck D, et al. Occurrenceand outcome of residual trophoblastic tissue: a prospective study.J Ultrasound Med. 2008;27:357Y361.
52. Takahama JJ, Kitano SS, Marugami NN, et al. Retained placental tissue:role of MRI findings in diagnosis and clinical assessment. AbdomImaging. 2011;36:110Y114.
53. Tanaka YOY, Sohda SS, Shigemitsu SS, et al. High temporal resolutiondynamic contrast MRI in a high risk group for placenta accreta. MagnReson Imaging. 2001;19:635Y642.
54. Soyer P, Sirol M, Fargeaudou Y, et al. Placental vascularity and resorptiondelay after conservative management of invasive placenta: MR imagingevaluation. Eur Radiol. 2013;23:262Y271.
55. Knuttinen M, Jani A, Gaba R, et al. Balloon occlusion of the hypogastricarteries in the management of placenta accreta: a case report and reviewof the literature. Semin Intervent Radiol. 2012;29:161Y168.
56. Rao AP, Bojahr H, Beski S, et al. Role of interventional radiology in themanagement of morbidly adherent placenta. J Obstet Gynaecol.2010;30:687Y689.
57. Tan CH, Tay KH, Sheah K, et al. Perioperative endovascular internal iliacartery occlusion balloon placement in management of placenta accreta.AJR Am J Roentgenol. 2007;189:1158Y1163.
58. Levine AB, Kuhlman KK, Bonn JJ. Placenta accreta: comparison ofcases managed with and without pelvic artery balloon catheters. J MaternFetal Med. 1999;8:173Y176.
59. Bodner LJ, Nosher JL, Gribbin C, et al. Balloon-assisted occlusion ofthe internal iliac arteries in patients with placenta accreta/percreta.Cardiovasc Intervent Radiol. 2006;29:354Y361.
60. Palacios-Jaraquemada JM. Diagnosis and management of placentaaccreta. Best Pract Res Clin Obstet Gynaecol. 2008;22:1133Y1148.
61. Hwang SM, Jeon GS, Kim MD, et al. Transcatheter arterial embolisationfor the management of obstetric haemorrhage associated with placentalabnormality in 40 cases. Eur Radiol. 2013;23:766Y773.
Ultrasound Quarterly & Volume 31, Number 1, March 2015 Imaging of Placenta Accreta
* 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ultrasound-quarterly.com 33
