You are on page 1of 190

xi

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) xi

Preface

Obstetric Ultrasound
Deborah Levine, MD Department of Radiology Beth Israel Deaconess Medical Center 330 Brookline Avenue Boston, MA 02215, USA E-mail address: dlevine@bidmc.harvard.edu
Deborah Levine, MD Guest Editor

It is with great pleasure that I write an introduction to the first issue on Obstetric Ultrasound for Ultrasound Clinics. We have gathered a group of articles that should be of interest to radiologists as well as to obstetricians and gynecologists. Included are two articles written regarding the first trimester. First trimester screening for aneuploidy is a growing field. The article by Dr. Nyberg and colleagues gives a detailed review of this important topic. We also have a review by Dr. Eyvazzedeh on pelvic pain in the first trimester that details the types of abnormalities that can present with pelvic pain and how to image these patients appropriately. For second and third trimester scanning, we have two articles

regarding the cardiovascular system. An article by Dr. Lee details imaging of the fetal heart. An article by Dr. Abuhammed gives an overview of use of Doppler imaging in obstetric ultrasound. Threedimensional ultrasound is another topic that is attracting growing interest in the ultrasound community. Drs. Pretorius and Tarsa give an overview on how to obtain volumes and how they can be useful in obstetric diagnoses. Finally, an article by Dr. Chie details imaging of the lower uterine segment in pregnancy. This article reviews important concepts of placenta accreta, cesarean scar pregnancies, and uterine dehiscence and rupture.

1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.cult.2006.03.001

ultrasound.theclinics.com

231

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 231255

First-Trimester Screening
David A. Nyberg, MD*, Jon Hyett, MD, Jo-Ann Johnson, MD, FRCSC, Vivienne Souter,
& &

MD

&

First-trimester aneuploidy screening Screening strategies First-trimester combined screen First-trimester combined screen plus other ultrasound markers First-trimester screening followed by second-trimester biochemistry First-trimester screening followed by second-trimester ultrasonography Other advantages of first-trimester screening

& &

Other chromosome abnormalities Birth defects in euploid fetuses who have increased nuchal translucency Twins and multiple gestations Structural defects detected during the first trimester Summary References

All patients have a 2% to 3% risk of birth defects, regardless of their prior history, family history, maternal age, or lifestyle [1]. Chromosome abnormalities account for approximately 10% of birth defects, but are important because of their high mortality and morbidity. Trisomy 21 (Down syndrome) is the most common serious chromosome abnormality at birth, occurring in approximately 1 of 500 pregnancies in the United States. The actual risk varies with maternal and gestational age and whether there is a history of previous pregnancies affected by chromosomal abnormality, although, as with other birth defects, all patients are at risk for fetal Down syndrome. A detailed fetal anatomic survey performed at 18 to 22 weeks remains the primary means for detecting the majority of serious structural birth defects; however, first-trimester screening at 11 to 14 weeks has developed into the initial screening test for many patients. A wealth of information can be obtained at this time, including detection of many structural defects, as well as screening for

fetal aneuploidy, including Down syndrome. The major advantage of first-trimester screening is the earlier gestational age of detection so that diagnostic testing (chorionic villous sampling [CVS] or genetic amniocentesis) can be made available for patients considered at highest risk for chromosome abnormalities. First-trimester screening can also help identify patients at increased risk for a variety of other abnormalities, including cardiac defects, that may be seen later. In this way, firsttrimester screening can help triage patients for subsequent testing. Older screening methods relied on clinical risk factors, particularly maternal age, to determine which patients might benefit from a diagnostic invasive test for fetal aneuploidy; however, maternal age alone is a poor screening method for determining who is at risk for chromosome abnormalities. First-trimester screening has proved to be very effective in screening for fetal aneuploidy. The accuracy of both first-trimester and second-trimester ultrasound can be improved by also considering various

University of Washington Medical Center, 1959 NE Pacific Street, Seattle, WA 98195, USA * Corresponding author. E-mail address: nyberg@u.washington.edu (D.A. Nyberg).
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.006

ultrasound.theclinics.com

232

Nyberg et al

biochemical markers. As a result, there are currently four main components to screening for fetal aneuploidy and other birth defects: (1) firsttrimester ultrasound, (2) first-trimester biochemistry, (3) second-trimester ultrasound, and (4) secondtrimester biochemistry. These four components of contemporary screening can be used in isolation or can be combined with one another for greater accuracy. This article focuses on first-trimester ultrasound screening, but also describes related screening protocols that can be used.

First-trimester aneuploidy screening


It is now well-known that increased fluid or thickening beneath the skin at the back of the neck is associated with a higher risk for fetal aneuploidy and other birth defects. This sonographic observation mirrors the clinical description of Down syndrome made more than 100 years ago by Dr. Langdon Down, who reported that the skin of affected individuals is too large for their bodies [2]. During the 1980s, many ultrasound studies described the typical appearance of cystic hygromas in the second trimester, and their association with aneuploidy, particularly Turners syndrome [212]. At the same time, it was observed that cystic hygromas seen during the first trimester may have different appearances (nonseptated), and different associations (trisomies) than those seen during the second trimester. It was also observed that cystic hygromas seen during the first trimester can resolve to nuchal thickening alone, or even normal nuchal thickness, and still be associated with aneuploidy [13,14]. In a related observation, Benacerraf and colleagues [15,16] noted that second-trimester nuchal thickening was associated with an increased risk of Down syndrome. In 1992, Nicolaides and colleagues [17] proposed the term nuchal translucency (NT) for the sonographic appearance of fluid under the skin at the back of the fetal neck observed in all fetuses during the first trimester [Fig. 1]. They further reported an association between the thickness of the translucency and the risk of fetal aneuploidy, especially trisomies. This concept of measuring NT in all fetuses formed the basis for first-trimester screening by ultrasound. By 1995, the first large study of NT was published [18]. Subsequent studies have confirmed that NT thickness can be reliably measured at 11 to 14 weeks gestation and, combined with maternal age, can produce an effective means of screening for trisomy 21 [19]. The mechanism for increased NT may vary with the underlying condition. The most likely causes
Fig. 1. Normal nuchal translucency measurement (arrows) at 12 weeks, 5 days.

include heart strain or failure [20,21] and abnormalities of lymphatic drainage [22]. Evidence for heart strain includes the finding of increased levels of atrial and brain natriuretic peptide mRNA in fetal hearts among trisomic fetuses [23]. Also, some Doppler ultrasound studies of the ductus venosus at 11 to 14 weeks in fetuses who have increased NT have reported absent or reversed flow during atrial contraction in the majority of chromosomally abnormal fetuses and in chromosomally normal fetuses who have cardiac defects [24,25]. Abnormal lymphatic drainage may occur because of developmental delay in the connection with the venous system, or a primary abnormal dilatation or proliferation of the lymphatic channels. Fetuses who have Turners syndrome are known to have hypoplasia of lymphatic vessels [26,27]. Lymphatic drainage could also be impaired by lack of fetal movements in various neuromuscular disorders, such as fetal akinesia deformation sequence [28]. An alternative explanation for increased NT is abnormal composition of the extracellular matrix. Many of the component proteins of the extracellular matrix are encoded on chromosomes 21, 18, or 13. Immunohistochemical studies of the skin of chromosomally abnormal fetuses have demonstrated specific alterations of the extracellular matrix that may be attributed to gene dosage effects [29,30]. Altered composition of the extracellular matrix may also be the underlying mechanism for increased fetal NT in certain genetic syndromes that are associated with alterations in collagen metabolism (such as achondrogenesis Type II), abnormalities of fibroblast growth factor receptors (such as achondroplasia and thanatophoric dysplasia), or disturbed metabolism of peroxisome biogenesis factor (such as Zellweger syndrome). All studies indicate that proper training is required to obtain reproducible, accurate data from

First-Trimester Screening

233

Box 1: Criteria for what constitutes an adequate NT measurement include 1. Crownrump length between 45 mm and 84 mm 2. Sagittal view that shows the nuchal measurement and face with the fetus in neutral position 3. Magnification so that only the upper two thirds of the fetus is included on the image 4. Distinguishing nuchal membrane from the amnion 5. Measuring maximal subcutaneous translucency overlying the neck 6. Identifying causes of falsely increased nuchal translucency measurements, including fetal extension, and nuchal cord

NT measurements [3133]. The Fetal Medicine Foundation (www.fetalmedicine.org) has outlined guidelines that have become the standard for measurement of NT throughout the world. These are listed in Box 1. They also offer a certificate of competency for those sonographers who successfully show they can adhere to them. Virtually identical guidelines have now been proposed by the Society for Maternal Fetal Medicine in the United States, and they also offer a certificate of competency. Use of the guidelines proposed by the Fetal Medicine Foundation have resulted in a high consistency in results [Table 1]. Monni and coworkers [34] reported that after modifying their technique of measuring NT, by following the guidelines established by The Fetal Medicine Foundation, their detection rate of trisomy 21 improved from 30% to 84%. The ability to measure NT and obtain reproducible results improves with training; good results are achieved after 80 and 100 scans for the transab-

dominal and the transvaginal routes, respectively [35]. The intraobserver and interobserver differences in measurements are less than 0.5 mm in 95% of cases [36]. NT is usually measured using a transabdominal approach; transvaginal scanning may be necessary in 5% to 10% of pregnancies when transabdominal scans are technically limited. The normal range for NT measurements is gestational age dependent. Pandya and colleagues [36] reported that the median NT increases from 1.3 mm at a crown-rump length (CRL) of 38 mm to 1.9 mm at a CRL of 84 mm. The 95th percentile increases from 2.2 mm at a crown rump length of 38 mm to 2.8 mm at a CRL of 84 mm. Sonographers should recognize that technical factors influence NT measurements. For example, extension of the neck increases NT thickness, whereas flexion reduces the measurement. The criteria for a positive NT scan have evolved since its first description. Initially a categorical cutoff measurement (usually 2.5 or 3 mm) was used by most centers; however, as noted above, NT increases with gestational age, and the degree of risk was found to vary with NT measurements. Therefore, it is more appropriate to express NT measurements relative to gestational age or CRL as a delta value or multiple of the median. Use of multiple of median data and derived likelihood ratios can then estimate the patient-specific risk. This also permits integration of risk based on NT with biochemical data to generate a combined risk. It should be noted that the median NT measurement for Down syndrome is about two multiples of the median. This is equivalent to about 2.5 mm at 12 weeks. The effectiveness of NT screening for detection of fetal Down syndrome has now been confirmed by a number of studies [see Table 1]. In the largest multicenter study published [19], 96,127 singleton pregnancies were examined, including 326 af-

Table 1: Studies examining the implementation of fetal nuchal translucency measurement at 1014 weeks of gestation in screening for trisomy 21 Author Pandya et al, 1995 [37] Szabo et al, 1995 [38] Taiplae et al, 1997 [39] Hafner et al, 1998 [40] Pajkrt et al, 1998 [41] Economides et al, 1998 [42] Zoppi et al, 2000 [43] Thilaganathan et al, 1999 [44] Schwarzler et al, 1999 [45] Theodoropoulos et al, 1998 [46] Total N 1763 3380 6939 4233 1473 2281 5210 11,398 4523 3550 44,750 Screening cutoff NT >2.5 mm NT >3.0 mm NT >3.0 mm NT >2.5 mm NT >3.0 mm NT >99th centile Risk >1 in 100 Risk >1 in 200 Risk >1 in 270 Risk >1 in 300 FPR 3.6% 1.6% 0.8% 1.7% 2.2% 0.4% 4.2% 4.7% 4.7% 4.9% 3.0% DR 3 of 4 28 of 31 4 of 6 3 of 7 6 of 9 6 of 8 33 of 47 16 of 21 10 of 12 10 of 11 (75%) (90%) (67%) (43%) (67%) (75%) (70%) (76%) (83%) (91%)

119 of 156 (76%)

Abbreviations: DR, detection rate; FPR, false-positive rate; N, number.

234

Nyberg et al

Fig. 2. Mildly increased nuchal translucency measurement associated with trisomy 21 (calibers). The nuchal measurement was 2.3 mm, which is about twice normal for gestational age. Biochemical values also indicated an increased risk for trisomy 21.

fected by trisomy 21 and 325 who had other chromosomal abnormalities. The median gestation at the time of screening was 12 weeks (range 1014 weeks), and the median maternal age was 31 years (range 1445 years). The fetal NT was above the 95th percentile for crown-rump length in 72% of the trisomy 21 pregnancies [Figs. 2, 3]. The estimated risk for trisomy 21 based on maternal age and fetal NT was above 1 in 300 in 8.3% of normal pregnancies and 82% of those affected by trisomy 21. For a screen positive rate of 5%, the sensitivity was 77% (95% CI: 72%82%). The cumulative data from a number of other studies demonstrate a sensitivity of 77% for a false positive rate of 3% [see Table 1] [3746]. The effectiveness of screening for fetal aneuploidy is further increased when nuchal trans-

lucency thickness is combined with biochemical markers [Table 2]. The two most effective maternal serum markers currently used in the first trimester are pregnancy-associated plasma protein A (PAPP-A) and free B-human chorionic gonadotrophin (B-hCG). Maternal serum free -human chorionic gonadotropin (-hCG) normally decreases with gestation after 10 weeks and maternal serum PAPP-A levels normally increase. Levels of these two proteins tend to be increased and decreased, respectively, in pregnancies affected by trisomy 21. There does not appear to any correlation between the rise in free -hCG and fall in PAPP-A seen in trisomy 21 pregnancies, so these markers may be combined for screening purposes [47]. Similarly, these biochemical markers are independent of fetal NT thickness, allowing combination of biochemical and ultrasound tests [48,49]. Some authorities believe it is important to distinguish cystic hygromas from increased NT [50], whereas others do not. Malone and coworkers [50] reported 132 cases of cystic hygroma with followup among 38,167 screened patients (1 in 289). Chromosomal abnormalities were diagnosed in 67 (51%), including 25 who had Down syndrome, 19 who had Turners syndrome, 13 who had trisomy 18, and 10 who had other types of chromosome abnormalities. Major structural fetal malformations, primarily cardiac and skeletal abnormalities, were diagnosed in 22 of the remaining 65 cases (34%). Of the remaining cases, 20 resulted in spontaneous fetal death (n = 5) or elective pregnancy termination (15). One of 23 normal surviTable 2: Studies examining the implementation of a combined first-trimester test using maternal age, fetal nuchal translucency thickness, free -hCG and PAPP-A to screen for trisomy 21 Author Orlandi et al, 1997 [53] Biagotti et al, 1998 [54] Benattar et al, 1999 [55] De Biasio et al, 1999 [56] De Graff et al, 1999 [57] Spencer et al, 1999 [47] Krantz et al, 2000 [58] N FPR DR

744 5.0% 6 of 7 (86%) 232 5.0% 24 of 32 (75%) 1656 5.0% 5 of 5 (100%) 1467 3.3% 11 of 13 (85%) 300 5.0% 31 of 37 (84%) 1156 5.0% 187 of 210 (89%) 5718 5.0% 30 of 33 (90%)

Fig. 3. Increased nuchal translucency and trisomy 21. The nuchal translucency measurement (NT ) exceeded 3 mm.

Total

11,273 4.8% 294 of 337 (87%)

First-Trimester Screening

235

vors (4%) was diagnosed with cerebral palsy and developmental delay at birth. Overall, survival with normal pediatric outcome was confirmed in 17% of cases (22 of 132). Compared with increased nuchal translucency (>3 mm), cystic hygromas carried a fivefold, 12 fold, and sixfold increased risk of aneuploidy, cardiac malformation, and perinatal death, respectively. On the other hand, cystic hygromas were associated with larger NT measurements than those that had increased NT but did not have cystic hygromas, so it remains uncertain whether cystic hygromas are an independent risk factor. Like patients who have increased NT, the vast majority of pregnancies that have normal evaluation at the completion of the second trimester resulted in a healthy infant and a normal pediatric outcome. Lateral neck cysts, also termed jugular lymphatic sacs, have been found by Bekker and coworkers

[51] to be associated with larger NT measurements and thus a higher risk for fetal aneuploidy [Fig. 4]. They found that among 26 fetuses with increased NT (>95th percentile), 22 had clearly visible jugular lymphatic sacs and 16 of 26 (62%) had aneuploidy. In comparison, two fetuses in the control group also showed jugular lymphatic sacs and their NT measurements were upper normal (2.8 mm and 2.9 mm). Although one might conclude that lateral neck cysts are associated with a high risk of fetal aneuploidy, Sharony and colleagues [52] found that the outcome of lateral neck cysts is associated with both the presence of other abnormalities and the NT measurement, but not with the presence of cysts themselves. On the other hand, these authors found a relatively high incidence of lateral neck cysts (2.4%) in the general population, suggesting that some of these cysts were very small and would have escaped general detection.

Fig. 4. Distended jugular lymphatic sacs. (A) Increased nuchal translucency measurement of 2.5 mm is noted. (B) Transvaginal scans show small bilateral fluid collections consistent with jugular lymphatic sacs. These are associated with increased nuchal translucency measurements.

236

Nyberg et al

Screening strategies First-trimester combined screen


The first-trimester combined screen uses maternal age, NT measurement, and biochemical markers (free -hCG and PAPP-A) to estimate the risk for fetal Down syndrome and trisomy 18. This is the most popular and effective screening strategy during the first trimester. A number of studies suggest a detection rate in the range of 85% to 90% for a screen positive rate of 5% [see Table 1] [5259]. Two large US studies have also been reported showing the effectiveness of first-trimester screening. The First-trimester Maternal Serum Biochemistry and Fetal Nuchal Translucency Screening (BUN) study found a 79% detection rate, for a 5% falsepositive rate [60]. The First- and Second-Trimester Evaluation of Risk (FASTER) Research Consortium trial [61] is the largest US-based study, and the only study that has compared first-trimester screening with second-trimester screening. The FASTER data [61] clearly confirm the pioneering work of Nicolaides and colleagues [17,19], with similar results. The overall detection rate was 85%, for a false-positive rate of 5%; however, the results clearly varied with gestational age, with detection rate of 87% at 11 weeks compared with 82% at 13 weeks.

poor during the first trimester [70,71]. This probably reflects the technical difficulty in obtaining accurate nasal bone measurements at this time. Imaging of the nasal bones requires a near-perfect midsagittal image and optimal angle of insonation with the fetal profile, whereas NT measurements can be obtained with minor variations off-center and differences in direction of imaging. Demonstrating the absence of a very small structure is even more difficult than detecting its presence, because it can be difficult to know for certain whether the nasal bones are absent or whether the images are simply suboptimal. Malone and coworkers [71] found that factors associated with an increased failure rate of nasal bone included early gestational age when the nasal bone is normally small, larger maternal body habitus, inadequate nuchal translucency sonography, and use of a transvaginal sonographic approach. Increased impedance of flow of the ductus venosus Abnormal Doppler flow patterns of the ductus venosus have been associated with an increased risk of fetal Down syndrome [Fig. 5] [24,25]. Matias and colleagues [24] performed ductus venosus Doppler measurements on 486 singleton fetuses, including 68 who had chromosomal abnormalities, at 10 to 14 weeks gestation. In 90.5% of the chromosomally abnormal fetuses there was reversed or absent flow during atrial contraction,

First-trimester combined screen plus other ultrasound markers


Although increased NT remains the primary ultrasound marker of fetal aneuploidy and other birth defects during the first trimester, several other ultrasound findings have been found to be helpful at this time. These include hypoplastic or absent nasal bone, and abnormal Doppler waveforms of the tricuspid valve and ductus venosus. Hypoplastic/absent nasal bone A small nasal bone was first noted to a common feature of patients who had trisomy 21 by Dr. Langdon Down [2]. Anthromorphic studies in patients who have trisomy 21 have shown a small nasal bone in approximately half of affected cases. A number of ultrasound studies have now also shown an association between sonographically absent nasal bone and trisomy 21 as well as other chromosome abnormalities [6269]. In the combined data of 15,822 fetuses, the fetal profile was successfully examined in 97.4%, and the nasal bone was absent in 1.4% of normal fetuses and in 69% of fetuses who had trisomy 21. A minority of studies have concluded that an absent nasal bone is not a useful feature to detect fetal Down syndrome, and that reproducibility is

Fig. 5. Abnormal ductus venosus Doppler and trisomy 21 (same fetus as Fig. 4). Duplex Doppler of the ductus venosus shows retrograde flow during atrial contractions.

First-Trimester Screening

237

whereas abnormal ductus flow was only present in 3.1% of the chromosomally normal fetuses. The height of the A-wave was found to be the only significant independent factor in multivariate regression analysis. Other researchers have also found that ductus venosus Doppler studies can substantially improve Down syndrome screening efficiency [72]. Tricuspid regurgitation has also been associated with an increased risk of fetal Down syndrome. In the largest study reported, Faiola and coworkers [73] reported that the tricuspid valve was successfully examined in 718 (96.8%) cases. Tricuspid regurgitation was found in 39 (8.5%) of the 458 chromosomally normal fetuses, in 82 (65.1%) of the 126 who had trisomy 21, in 44 (53%) of the 83 who had trisomy 18 or 13, and in 11 (21.6%) of the 51 who had other chromosomal defects. In chromosomally normal fetuses, tricuspid regurgitation was associated with increased NT measurements, suggesting that Doppler studies may be particularly useful in this group of patients. Fetuses who have abnormal flow patterns of the ductus venosus and tricuspid valve also appear to have a higher risk of cardiac defects. Among 142 chromosomally normal fetuses who had increased NT, 11 fetuses had reversed or absent flow on ductus venosus Doppler during atrial contraction, and 7 of these had major cardiac defects at subsequent echocardiography [25]. Similarly, Faiola and colleagues [73] found that in the chromosomally normal fetuses, tricuspid regurgitation was found in nearly half (46.9%) of fetuses who had cardiac defects and in 5.6% of those who did not have cardiac defects (likelihood ratio of 8.4). Nicolaides and coworkers [74] suggest that secondary findings of absent nasal bone or abnormal Doppler studies could be particularly useful in patients found to be in the intermediate risk group by the first-trimester screen. Using these secondary signs in patients with an intermediate risk group (risk of 1 in 100 to 1 in 1000) for fetal Down syndrome, the researchers reported detection rates of 92% for absent nasal bone, 94% for increased impedance of the ductus venosus, and 91.7% for tricuspid regurgitation, with each method showing an overall false-positive rate of less than 3% [74].

First-trimester screening followed by second-trimester biochemistry


Second-trimester biochemical screening can detect 70% to 80% of affected fetuses who had Down syndrome (at a false positive rate of 7%8%). The effectiveness appears to be clearly higher for the quad screen (HCG, alpha-fetoprotein, estriol, and inhibit-A), than the older triple screen that did not include inhibin-A [50]; however, the effective-

ness of second-trimester biochemical screening is more limited in a population that has already been screened, and in the authors experience, most patients who have undergone first-trimester screening will choose not to undergo second-trimester biochemical screening. For those patients who would like additional reassurance by way of a second-trimester biochemical screen, it should be done in a way that accounts for the first-trimester screening results rather than treating them as independent tests. One method is the so-called integrated screen, which combines the elements of the first-trimester combined screen with the elements of the second-trimester quad screen, providing a single, low false-positive result in the second trimester [75]. This is the most accurate screening method currently available, with detection rate of 92% in the FASTER study [76]; however, a major disadvantage of integrated screening is that patients do not receive results until after completion of the second-trimester biochemistry. Thus screen-positive women do not have the option of CVS for early definitive diagnosis [77]. In addition, it is considered unethical to suppress ultrasound information obtained in the first trimester. Stepwise sequential screening is an alternative approach that has been proposed; it interprets second-trimester results based on first-trimester risk assessment. A clear advantage of stepwise sequential screening is that it provides some women an earlier diagnosis while maintaining an extremely high detection rate. This method has gained rapid acceptance and it is expected to be widely adapted into clinical practice in the near future [78]. When patients in the FASTER trial underwent firsttrimester combined screening at 11 weeks and the false-positive rate of each component was set at 2.5%, stepwise sequential screening provided a 95% detection of Down syndrome, for a 4.9% false-positive rate. This compares to a 4.0% falsepositive rate for fully integrated screening. Incorporation of second-trimester biochemical as part of a stepwise sequential screen would be most effective for patients considered in an intermediate risk group (risk between 1 in 100 and 1 in 1000) [79]. The intermediate group includes 15% of affected fetuses who had Down syndrome and approximately 15% of normal fetuses. In comparison, high risk patients (risk >1 in 100) should probably consider diagnostic invasive testing without additional screening; this group includes 80% of affected fetuses who have Down syndrome but only 5% of normal fetuses. Also, low-risk patients (risk <1 in 1000) probably do not require additional screening in most cases; this group of patients includes less than 5% of affected fetuses

238

Nyberg et al

who have Down syndrome, but 80% of normal fetuses.

First-trimester screening followed by second-trimester ultrasonography


A second-trimester fetal survey remains the primary method of detecting the majority of birth defects that can be detected prenatally [80]. Because of the wide range of anomalies that can be detected at this time, this examination is unlikely to be replaced by any other screening test in the future. In addition to detection of structural defects, the presence or absence of various sonographic markers can further modify the risk for fetal aneuploidy, including Down syndrome. The estimated risk can be derived by multiplying the background risk (based on maternal age, gestational age, history of previously affected pregnancies, and, where appropriate, the results of previous screening by NT or biochemistry in the current pregnancy) by the likelihood ratio of the specific defect [81]. The most common second-trimester ultrasound markers that are systematically evaluated include nuchal thickening, echogenic intracardiac foci, absent or hypoplastic nasal bone, hyperechoic bowel, renal pyelectasis, and shortened femur and humerus lengths relative to the biparietal diameter. Nyberg and coworkers [82] and others have calculated likelihood ratios for many of these markers and have refined this for single markers [83]. In the vast majority of cases, second-trimester ultrasound markers such as echogenic intracardiac foci will be found in normal fetuses, especially when the marker is isolated. In this situation, a prior normal first-trimester screening result can be very reassuring. Because a normal first-trimester screening results permits significant reduction of

risk for fetal Down syndrome, and because isolated findings such as echogenic intracardiac foci only slightly increase the risk, most patients will remain at very low risk and do not require further testing. Ultrasound findings, however, can also improve the detection rate of fetuses who have Down syndrome in patients who have borderline normal results from first-trimester screening, or fetuses who show multiple markers or major defects. At the same time, a normal second-trimester ultrasound can reduce the risk of fetal Down syndrome approximately threefold, and this can normalize patients who have borderline positive results form first-trimester screening (risk 1 in 100 to 1 in 300). Results of the FASTER trial show that use of a second-trimester genetic sonogram can both improve the detection rate and lower the false positive rate in patients who have undergone first-trimester screening [84].

Other advantages of first-trimester screening Other chromosome abnormalities


Nuchal translucency is also increased with other chromosome abnormalities, including trisomies 13 and 18, Turners syndrome, triploidy, and unbalanced translocations [Fig. 6] [85]; however, firsttrimester biochemical markers may differ from those typically associated with trisomy 21. In trisomies 18 and 13, maternal serum free -hCG and PAPP-A are decreased [86,87]. In cases of sex chromosomal anomalies, maternal serum free -hCG is normal and PAPP-A is low [88]. Triploidy of paternal origin, which is associated with a partial molar placenta, has greatly increased levels of free -hCG, whereas PAPP-A is mildly decreased [89]. In con-

Fig. 6. Increased nuchal translucency and trisomy 18. Large nuchal translucency measurement was noted and cytogenetic testing revealed trisomy 18.

First-Trimester Screening

239

Table 3: Abnormalities and genetic syndromes reported in association with increased nuchal translucency and normal karyotype Central nervous sytem defect Anencephaly Craniosynostosis Dandy-Walker malformation Diastematomyelia Encephalocele Holoprosencephaly Hydrolethalus syndrome Joubert syndrome Microcephaly Macrocephaly Spina bifida Iniencephaly Trigoncephaly C Ventriculomegaly Agnathia/micrognathia Facial cleft Treacher-Collins syndrome Cystic hygroma Neck lipoma Di George syndrome Cystic adenomatoid malformation Diaphragmatic hernia Fryn syndrome Cloacal exstrophy Omphalocele Gastroschisis Crohns disease Duodenal atresia Esophageal atresia Small bowel obstruction Ambiguous genitalia Congenital nephrotic syndrome Hydronephrosis Hypospadius Infantile polycystic kidney disease Meckel-Gruber syndrome Megacystis Multicystic dysplastic kidney disease Renal agenesis Achondrogenesis Achondroplasia Asphyxiating thoracic dystrophy Blomstrand osteochondrodysplasia Campomelic dwarfism Jarcho-Levin syndrome Kyphoscoliosis Limb reduction defect Noonan-Sweeney syndrome Osteogenesis imperfecta Roberts syndrome

Table 3: (continued ) Central nervous sytem defect Anencephaly

Facial defect

Nuchal defect Cardiac defect Pulmonary defect

Abdominal wall defect Gastrointestinal defect

Genitourinary defect

Robinow syndrome Short rib polydactyly Sirenomelia Talipes equinovarus Split hand/foot malformation Thanatophoric dwarfism VACTER association Fetal anemia Blackfan-Diamond anemia Dyserthropoietic anemia Fanconi anemia Parovirus 19 infection Alpha thalassemia Neuromuscular Fetal akinesia defect deformation sequence Myotonic dystrophy Spina muscular atrophy Metabolic defect Beckwith-Wiedemann syndrome GM1 gangliosidosis Long-chain 3-hydroyacylcoenzyme A dehydrogenase deficiency Mucopolysaccharisosis Type VII Smith-Lemli-Opitz syndrome Vitamin D-resistant rickets Zellweger syndrome Other Body stalk anomaly (limb body wall complex) Brachmann-de Lange syndrome CHARGE association Deficiency of the immune system Congenital lymphedema EEC syndrome Neonatal myoclonic encephalopathy Noonan syndrome Perlman syndrome Stickler syndrome Unspecified syndrome Severe developmental delay
Abbreviation: EEC syndrome, ectrodactyly-ectodermal dysplasia-cleft palate syndrome. Adapted from Souka AP, Krampl E, Bakalis S, et al. Outcome of pregnancy in chromosomally normal fetuses with increased nuchal translucency in the first-trimester. Ultrasound Obstet Gynecol 2001;18(1):13, 14.

Skeletal defect

trast, digynic triploidy, characterized by severe asymmetrical fetal growth restriction, is associated with markedly decreased maternal serum free -hCG and PAPP-A. Screening by a combination of fetal NT, free -hCG, and PAPP-A can identify

240

Nyberg et al

Table 4: Nuchal translucency measurements and adverse outcomes Nuchal translucency measurement <95th percentile 95th99th 3.54.4 mm 4.55.4 mm 5.56.4% 6.5 mm
Data from Refs. [19,42,90,103].

Aneuploidy .2% 3.7% 21.1% 33.3% 50.5% 64.5%

Death 1.3% 1.3% 2.7% 3.4% 10.1% 19%

Major anomaly 1.6% 2.5% 10% 18.5% 24.2% 46.2%

Alive and well 97% 93% 70% 50% 30% 15%

about 90% of these anomalies for a screen positive rate of 1%.

Birth defects in euploid fetuses who have increased nuchal translucency


Extensive studies have now established that, in chromosomally normal fetuses, increased NT is associated with a wide range of fetal defects and genetic syndromes [Table 3]. The prevalence of birth defects and adverse outcome also increases with increasing NT measurements [Table 4]. Souka and colleagues [90] reported that the overall risk of adverse outcome, including miscarriage and intrauterine death, was 32% for those who had NT of 3.5 to 4.4 mm, 49% for NT of 4.5 to 5.4 mm, 67% for NT 5.5 to 6.4 mm, and 89% for those who had NT of 6.5 mm or more. Among 1080 surviving fetuses who had NT of 3.5 mm or more, 5.6% had abnormalities requiring medical or surgical treatment or leading to mental handicap. The chance of no defect among live births was 86% for those who had NT of 3.5 to

4.4 mm, 77% for those who had NT of 4.5 to 5.4 mm, 67% for those who had NT of 5.5 to 6.4, and 31% for those who had NT of 6.5 mm or more. An association between increased NT and cardiac defects was first noted by Hyett and coworkers [20] in both chromosomally abnormal and normal fetuses. This has subsequently been confirmed by a number of studies [91100]. A retrospective study of 29,154 chromosomally normal singleton pregnancies identified major defects of the heart and great arteries in 50 cases, and 56% of these had NT measurement translucency above the 95th percentile [101]. In chromosomally normal fetuses, the prevalence of major cardiac defects increases exponentially from 1.6 per 1000 for NT less than 95th percentile, 1% for NT between 2.5 and 34 mm, 3% for NT 3.5% to 4.4%, 7% for NT 4.5% to 5.4%, 20% for NT 5.5 to 6.4 mm, 30% for NT 6.5 mm or more. The clinical implication of these observations is that patients found to have increased NT should undergo formal fetal echocardiography. Certainly,

Fig. 7. Discrepant nuchal translucency measurements in monochorionic twins. (A) This fetus shows nuchal translucency measurement (NT ) of 2 mm at 12 weeks. The co-twin showed nuchal translucency measurement of 1.1 mm. (B) Velemenous cord insertion is also apparent. This monochorionic twin pregnancy showed signs of severe twin-twin transfusion syndrome by 18 weeks.

First-Trimester Screening

241

Fig. 8. Normal face at 13 weeks. (A) Sagittal view shows normal facial profile including nasal bone. (B) 3D multiplanar ultrasound with surface rendering shows normal facial features.

the overall prevalence of major cardiac defects in such a group of fetuses (about 2%) is similar to that found in pregnancies affected by maternal diabetes mellitus or who have a history of a previously affected offspring, which are well-accepted indications for fetal echocardiography. Improvements in the resolution of ultrasound machines have now made it possible to undertake detailed cardiac scanning as early as 14 weeks [87,102].

It should be emphasized to the parents that increased NT per se does not constitute a fetal abnormality, and that, once chromosomal defects have been excluded, nearly 90% of liveborns who have fetal translucency below 4.5 mm have healthy live births. If the fetus survives until midgestation, and if a targeted ultrasound at 20 to 22 weeks fails to reveal any abnormality, the risk of adverse outcome is not statistically increased [103]. The rate of

Fig. 9. Normal brain at 12 weeks. (A) Transabdominal scans show that the normal choroid plexus dominates the cerebral hemispheres. (B) Transvaginal scan on the same patient better shows normal anatomy.

242

Nyberg et al

development delay is also not statistically increased among fetuses who have increased NT [104].

Twins and multiple gestations


First-trimester screening can be effectively used for twin pregnancies [105]. Detection rates for Down syndrome are in the range of 75% to 85%, with a 5% false-positive rate [106]. Therefore, effective screening and diagnosis of major chromosomal abnormalities can be achieved in the first-trimester, allowing the possibility of earlier and therefore safer selective feticide for those parents that choose this option. Discrepant NT measurements also appear to be a nonspecific early marker of twin-twin transfusion syndrome among monochorionc twins [Fig. 7]. In a study of 132 monochorionic twin pregnancies, including 16 that developed severe twin-to-twin transfusion syndrome at 1522 weeks of gesta-

tion, increased NT (above the 95th percentile of the normal range) at the 11 to 14 week scan was associated with a fourfold increase in risk for the subsequent development of severe twin-to-twin transfusion syndrome [107]. It is possible that increased NT thickness in the recipient fetus may be a manifestation of heart failure caused by hypervolemic congestion. With advancing gestation and the development of diuresis that would tend to correct the hypervolemia and reduce heart strain, both the congestive heart failure and NT resolve. Severe complications unique to monochorionic pregnancies, such as reversed arterial perfusion syndrome or acardiac twin, and conjoined twins, can be diagnosed during the first trimester. Twin reversed arterial perfusion (TRAP) has been reported at 10 to 12 weeks using both TVS and color Doppler [108,109]. Conjoined twins have also been frequently diagnosed during the first trimester,

Fig. 10. Normal anatomy. (A) Transvaginal scan at 13 weeks shows normal four-chamber view of the heart. (B) Transabdominal scan at 13 weeks shows normal fluid-filled stomach. (C ) Transabdominal scan of the pelvis at 12 weeks shows a normal urinary bladder between the two umbilical arteries, seen with color flow Doppler. A normal urinary bladder is less frequently seen than the stomach.

First-Trimester Screening

243

Fig. 11. Normal genitalia. (A) Male genitalia at 13 weeks. (B) Female genitalia at 12 weeks.

and have been detected as early as 8 to 9 weeks [110118].

Structural defects detected during the first trimester


Use of a systematic survey can demonstrate normal anatomic development in the first trimester, similar to the fetal survey performed during the second trimester. Normal structures that can be visualized include the brain, choroid plexi, posterior fossa, face, heart, thorax, abdomen, stomach, urinary bladder, and all four extremities, including both feet and hands [Figs. 810]. In addition, the individual digits of each hand can usually be counted by 12 weeks. Fetal gender can be reliably determined by 13 weeks, and by 12 weeks in most cases [Fig. 11] [119]. When deviation from normal anatomy is recognized, a number of birth defects can be detected during the first trimester. Detection varies significantly between centers, with increasing detection by a thorough systematic survey and greater use of transvaginal ultrasound and three-dimensional (3D) multiplanar ultrasound.

Fig. 12. Anencephaly/acrania at 12 weeks. The normal calvarium is not visualized and the shape of the brain is slightly abnormal. Anencephaly/acrania can be easily missed at this gestational age.

244

Nyberg et al

First-Trimester Screening

245

Fig. 14. (A,B) Omphalocele associated with trisomy 18 at 10.5 weeks. Chorionic villus sampling showed trisomy 18.

Ossification of the fetal cranium begins and accelerates after 9 weeks [120,121],so that anencephaly can be diagnosed as early as 9 to 10 weeks [122]. Ancephaly can also be easily overlooked during the first trimester, however, because it initially is seen as acrania with absent calvarium but relatively normal amount of brain. Careful scrutiny will show an abnormal shape and appearance of the brain caused by the lack of the supporting calvarium [Fig. 12]. The sagging appearance of the brain may show Mickey Mouse ears. Posterior cephaloceles have been diagnosed as early as 12 weeks [123], and alobar holoprosencephaly has been diagnosed as early as 10 weeks [124,125], but other brain abnormalities cannot reliably detected until later. Spina bifida can occasionally be detected before the 12th postmenstrual week by noting irregularities of the bony spine or a bulging within the posterior contour of the fetal back [126]. There are also well-established additional sonographic findings that can enhance the detection of spina bifida, namely the lemon sign or the banana sign [127,128], and these may be evident as early as 12 weeks, although they can be initially subtle [129131]. With high quality imaging, which may include tansvaginal scans, a normal posterior cere-

bellum and cisterna magna should be apparent, and this finding excludes all but the mildest forms of spina bifida. Cleft lip and palate have been diagnosed in utero as early as the 13 to 14 weeks [132]. Bilateral cleft lip and palate may appear initially only as a an echogenic median mass, which actually is the premaxillary protrusion, made up of soft tissue, and at times of osseous and dental structures [Fig. 13] [133]. Because bilateral cleft lip and palate is associated with a high rate of aneuploidy and other birth defects, close follow-up, genetic counseling, and amniocentesis should be offered. Ocular abnormalities such as hyper- and hypotelorism, anophthalmia and microphthalmia, have been diagnosed from 12 to 16 weeks [134136]. Congenital cataracts has been diagnosed as early as 12 to 14 weeks [137,138]. By 12 to 14 weeks, a four-chamber view of the heart can be consistently imaged [139142]. The great arteries can also be imaged by 11 to 12 weeks in many cases. As with normal anatomy later in the second trimester, the right and left ventricles should be of approximately the same size, the heart should not occupy more than one third of the thoracic cavity, and the heart apex should be oriented obliquely to the left anterior thorax.

Fig. 13. Bilateral cleft lip associated with trisomy 13 at 13 weeks. (A) Sagittal view shows abnormal soft tissue protruding just below the nose (arrow). (B) Transverse view confirms this finding. Bilateral cleft lip and palate was diagnosed (arrow). (C ) Umbilical cord cyst (arrow, C ) was also noted. (D) Follow-up 3D rendered image at 17 weeks confirms bilateral cleft lip and palate with premaxially protrusion. Other findings identified on the follow-up ultrasound, but not seen on the first-trimester scan, included echogenic intracardiac focus in the left ventricle, mildly hypoplastic left ventricle and atrium, micro-opthalmia, echogenic kidneys, and polydactyly.

246

Nyberg et al

Fig. 15. Normal hands. (A) 2D ultrasound at 12 weeks, 3 days shows normal hand with four fingers and one thumb. (B) Another fetus at 13 weeks shows normal hand and extremities with 3D surface rendering.

Achiron and colleagues [143] reported eight cases of heart defects among approximately 1000 fetuses scanned by transvaginal ultrasound between 10 and 12 weeks. Only one fetus had an abnormal karyotype (45XO), but all fetuses showed other anomalies. Based on this experience, detection of isolated heart abnormalities is likely to remain difficult before 14 weeks. Abdominal and truncal defects may be diagnosed during the first trimester, and these include omphalocele, gastroschisis [144,145], ectopia cordis

[146,147], and body-stalk anomaly [148,149]. Omphaloceles may be categorized as those containing both bowel and liver (extracorporeal liver) and those containing only bowel (intracorporeal liver). Intracorporeal omphalocele can only be reliably diagnosed after 12 postmenstrual weeks, because of the difficulty in distinguishing it from physiologic midgut herniation [150,151]. Such omphaloceles have a high rate of fetal aneuploidy [152,153]. Extracorporeal omphalocele can be diagnosed as early as 9 to 10 weeks [154156], and these may

Fig. 16. Normal limbs. (A) 3D surface rendering image shows poor visualization of extremities. (B) Transvaginal scans better shows normal extremities.

First-Trimester Screening

247

also be associated with fetal aneuploidy and other birth defects, including cardiac defects [Fig. 14]. The kidneys assume their final position within the renal fossa by 11 weeks [157]. Using transvaginal ultrasound, the kidneys can be consistently imaged by 12 to 13 weeks [158160]. Cystic kidneys can sometimes be diagnosed during the first trimester. Multicystic dysplastic kidney disease has been diagnosed as early as 12 to 15 weeks [160]. Infantile polycystic kidney disease has also been diagnosed by 13 to 16 weeks by demonstration of enlarged, echogenic kidney. [161,162], although oligohydramnios may not develop until after 16 weeks.

The urinary bladder becomes apparent at 10 to 12 weeks, but like the kidney, it does not become consistently imaged until the 13th week [162], at which time cyclical filling and emptying of the fetal bladder should be apparent. Obstructive uropathy at the level of the urethra results in an enlarged urinary bladder (megacystis), which has been diagnosed as early as the 11th week [163,164]. It has been suggested that the diagnosis of megacystis can be reliably diagnosed when the urinary bladder measures more than 15 mm during the first trimester [165]. Affected fetuses seen during the first trimester have a high rate of associated anomalies and aneuploidy [166].

Fig. 17. Normal extremity movements at 13 weeks. Three images obtained within a few seconds of one another (A,B,C ) show normal extremities with active normal movement.

248

Nyberg et al

The limbs begin to develop toward the end of the sixth week with development of the upper limbs before the lower limbs [167], and they can be imaged by the eighth week [168]. By 12 weeks the hands, fingers, feet, and toes can be consistently imaged [Fig. 15]. Use of transvaginal sonography and 3D ultrasound with surface rending can aid in visualization of the extremities [Fig. 16]. By the 12th week, the long bones, phalanges, ilium, and scapula begin to ossify; the metacarpals and metatarsals ossify by 12 to 16 weeks [169]. Active fetal movements can be observed after 10 weeks [170]. Normal fetal activity is particularly apparent using

real-time 3D (4D) ultrasound with surface rendering [Fig. 17]. A variety of skeletal abnormalities can be detected during the first trimester, including amputation defects and certain lethal skeletal dysplasias [Fig. 18]; however, their detection clearly varies with gestational age. In one of the largest reported series of prenatally diagnosed skeletal abnormalities in the first and early second trimesters, Bronshtein and coworkers [171] were able to detect 96% of the anomalies between 14 to 16 weeks, 3% between 12 to 14 weeks, and 1% at 10 to 12 weeks. Osteogenesis imperfecta (OI) is one of the lethal skele-

Fig. 18. Clubfeet at 12 weeks. (A) Transabdominal scan at 12 weeks, 4 days shows clubbed foot (arrow, F ). (B) 3D surface rendered image confirms severe bilateral clubfeet (arrow). This was also confirmed on follow-up scans at 18 weeks.

First-Trimester Screening

249

tal dysplasias that has been diagnosed as early as 13 to 15 weeks [172175]. Sirenomelia has been diagnosed as early as 11 to 14 weeks using transvaginal ultrasound [176179]. It is expected that akensia can be detected during the first trimester. Polydactyly can also be detected during the first trimester, and this can be aided by use of 3D multiplanar ultrasound.

Summary
Screening for fetal chromosome abnormalities, particularly for trisomy 21, has made dramatic advances in the last 15 years. These advances have both complicated screening and provided couples with more effective screening options. More effective screening has demonstrated that patients who traditionally were considered high riskparticularly patients aged 35 or oldercan be at lower risk for aneuploidy and other birth defects than a 20-year-old woman who does not undergo screening. This has resulted in a clear trend in the reduction of amniocentesis for these patients, and at the same time has made screening available for younger patients who share the 2% to 3% risk of birth defects that all pregnancies carry. More effective screening translates into lower proceduralrelated losses of normal fetuses, and better use of resources. The trend toward earlier detection of structural defects during the first trimester will undoubtedly continue as ultrasound resolution and 3D multiplanar ultrasound continue to improve. Conversely, a normal systematic survey at this time can be reassuring and can help to exclude a variety of major defects. Based on the presence or absence of findings, patients can then be triaged into early follow-up and possible amniocentesis at 14 to 16 weeks, or a later detailed anatomic survey at 18 to 20 weeks.

References
[1] Harper PS. Practical genetic counselling. 5th edition. Boston: Butterworth Heinemann; 1998. p. 5670. [2] Langdon Down J. Observations on an ethnic classification of idiots. Clin Lectures and Reports, London Hospital 1866;3:25962. [3] Phillips HE, McGahan JP. Intrauterine fetal cystic hygromas: sonographic detection. AJR Am J Roentgenol 1981;136(4):799802. [4] Toftager-Larsen K, Benzie RJ, Doran TA, et al. Alpha-fetoprotein and ultrasound scanning in the prenatal diagnosis of Turners syndrome. Prenat Diagn 1983;3(1):3540.

[5] Chervenak FA, Isaacson G, Blakemore KJ, et al. Fetal cystic hygroma. Cause and natural history. N Engl J Med 1983;309(14):8225. [6] Newman DE, Cooperberg PL. Genetics of sonographically detected intrauterine fetal cystic hygromas. J Can Assoc Radiol 1984;35(1):779. [7] Brown BS, Thompson DL. Ultrasonographic features of the fetal Turner syndrome. J Can Assoc Radiol 1984;35(1):406. [8] Redford DH, McNay MB, Ferguson-Smith ME, et al. Aneuploidy and cystic hygroma detectable by ultrasound. Prenat Diagn 1984;4(5): 37782. [9] Marchese C, Savin E, Dragone E, et al. Cystic hygroma: prenatal diagnosis and genetic counselling. Prenat Diagn 1985;5(3):2217. [10] Garden AS, Benzie RJ, Miskin M, et al. Fetal cystic hygroma colli: antenatal diagnosis, significance, and management. Am J Obstet Gynecol 1986;154(2):2215. [11] Brambati B, Simoni G. Diagnosis of fetal trisomy 21 in first-trimester. Lancet 1983; 1(8324):586. [12] Gustavii B, Edvall H. First-trimester diagnosis of cystic nuchal hygroma. Acta Obstet Gynecol Scand 1984;4:3836. [13] Reuss A, Pijpers L, Schampers PTFN, et al. The importance of chorionic villus sampling after first-trimester diagnosis of cystic hygroma. Prenat Diagn 1987;7:299301. [14] Bronshtein M, Rottem S, Yoffe N, et al. Firsttrimester and early second-trimester diagnosis of nuchal cystic hygroma by transvaginal sonography: diverse prognosis of the septated from the nonseptated lesion. Am J Obstet Gynecol 1988;161(1):7882. [15] Benacerraf BR, Barss V, Laboda LA. A sonographic sign for detection in the secondtrimester of the fetus with Downs syndrome. Am J Obstet Gynecol 1985;151:10789. [16] Benacerraf BR, Gelman R, Frigoletto FD. Sonographic identification of second-trimester fetuses with Downs syndrome. N Engl J Med 1987;317:13716. [17] Nicolaides KH, Azar G, Byrne D, et al. Fetal nuchal translucency: ultrasound screening for chromosomal defects in first-trimester of pregnancy. BMJ 1992;304(6831):8679. [18] Pandya PP, Kondylios A, Hilbert L, et al. Chromosomal defects and outcome in 1015 fetuses with increased nuchal translucency. Ultrasound Obstet Gynecol 1995;5:159. [19] Snijders RJM, Noble P, Sebire N, et al. UK multicentre project on assessment of risk of trisomy 21 by maternal age and fetal nuchaltranslucency thickness at 1014 weeks of gestation. Lancet 1998;352:3436. [20] Hyett JA, Moscoso G, Nicolaides KH. Abnormalites of the heart and great arteries in firsttrimester chromosomally abnormal fetuses. Am J Med Genet 1997;69:20716. [21] Hyett JA, Perdu M, Sharland GK, et al. Increased nuchal translucency at 1014 weeks of gestation

250

Nyberg et al

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

as a marker for major cardiac defects. Ultrasound Obstet Gynecol 1997;10:2426. Carvalho JS. The fetal heart or the lymphatic system or ? The quest for the etiology of increased nuchal translucency. Ultrasound Obstet Gynecol 2005;25(3):21520. Hyett JA, Brizot ML, von Kaisenberg CS, et al. Cardiac gene expression of atrial natriuretic peptide and brain natriuretic peptide in trisomic fetuses. Obstet Gynecol 1996;87:50610. Matias A, Gomes C, Flack N, et al. Screening for chromosomal abnormalities at 1114 weeks: the role of ductus venosus blood flow. Ultrasound Obstet Gynecol 1998;12:3804. Matias A, Huggon I, Areias JC, et al. Cardiac defects in chromosomally normal fetuses with abnormal ductus venosus blood flow at 10 14 weeks. Ultrasound Obstet Gynecol 1999;14: 30710. Chitayat D, Kalousek DK, Bamforth JS. Lymphatic abnormalities in fetuses with posterior cervical cystic hygroma. Am J Med Genet 1989; 33:3526. von Kaisenberg CS, Nicolaides KH, BrandSaberi B. Lymphatic vessel hypoplasia in fetuses with Turner syndrome. Hum Reprod 1999;14: 8236. Hyett J, Noble P, Sebire NJ, et al. Lethal congenital arthrogryposis presents with increased nuchal translucency at 1014 weeks of gestation. Ultrasound Obstet Gynecol 1997;9:3103. von Kaisenberg CS, Brand-Saberi B, Christ B, et al. Collagen Type VI gene expression in the skin of trisomy 21 fetuses. Obstet Gynecol 1998; 91:31923. von Kaisenberg CS, Krenn V, Ludwig M, et al. Morphological classification of nuchal skin in fetuses with trisomy 21, 18 and 13 at 12 18 weeks and in a trisomy 16 mouse. Anat Embryol (Berl) 1998;197:10524. Kornman LH, Morssink LP, Beekhuis JR, et al. Nuchal translucency cannot be used as a screening test for chromosomal abnormalities in the first-trimester of pregnancy in a routine ultrasound practice. Prenat Diagn 1996;16:797805. Bower S, Chitty L, Bewley S, et al. First-trimester nuchal translucency screening of the general population: data from three centres [abstract]. Presented at the 27th British Congress of Obstetrics and Gynaecology. Dublin, Ireland, Royal College of Obstetrics and Gynaecology; 1995. Roberts LJ, Bewley S, Mackinson AM, et al. Firsttrimester fetal nuchal translucency: problems with screening the general population 1. Br J Obstet Gynaecol 1995;102:3815. Monni G, Ibba RM, Zoppi MA. Antenatal screening for Downs syndrome. Lancet 1998; 352:16312. Braithwaite JM, Kadir RA, Pepera TA, et al. Nuchal translucency measurements: training of potential examiners. Ultrasound Obstet Gynecol 1996;8:1925.

[36] Pandya PP, Altman D, Brizot ML, et al. Repeatability of measurement of fetal nuchal translucency thickness. Ultrasound Obstet Gynecol 1995;5:3347. [37] Pandya PP, Goldberg H, Walton B, et al. The implementation of first-trimester scanning at 1013 weeks gestation and the measurement of fetal nuchal translucency thickness in two maternity units. Ultrasound Obstet Gynecol 1995; 5:205. [38] Szabo J, Gellen J, Szemere G. First-trimester ultrasound screening for fetal aneuploidies in women over 35 and under 35 years of age. Ultrasound Obstet Gynecol 1995;5:1613. [39] Taipale P, Hiilesmaa V, Salonen R, et al. Increased nuchal translucency as a marker for fetal chromosomal defects. N Engl J Med 1997;337: 16548. [40] Hafner E, Schuchter K, Liebhart E, et al. Results of routine fetal nuchal translucency measurement at 1013 weeks in 4233 unselected pregnant women. Prenat Diagn 1998;18:2934. [41] Pajkrt E, van Lith JMM, Mol BWJ, et al. Screening for Downs syndrome by fetal nuchal translucency measurement in a general obstetric population. Ultrasound Obstet Gynecol 1998; 12:1639. [42] Economides DL, Whitlow BJ, Kadir R, et al. First-trimester sonographic detection of chromosomal abnormalities in an unselected population. Br J Obstet Gynaecol 1998;105:5862. [43] Zoppi MA, Ibba RM, Putzolu M, et al. Assessment of risk for chromosomal abnormalities at 1014 weeks of gestation by nuchal translucency and maternal age in 5210 fetuses at a single centre. Fetal Diagn Ther 2000;15:1703. [44] Thilaganathan B, Sairam S, Michailidis G, et al. First-trimester nuchal translucency: effective routine screening for Downs syndrome. Br J Radiol 1999;72:9468. [45] Schwarzler P, Carvalho JS, Senat MV, et al. Screening for fetal aneuploidies and fetal cardiac abnormalities by nuchal translucency thickness measurement at 1014 weeks of gestation as part of routine antenatal care in an unselected population. Br J Obstet Gynaecol 1999;106:102934. [46] Theodoropoulos P, Lolis D, Papageorgiou C, et al. Evaluation of first-trimester screening by fetal nuchal translucency and maternal age. Prenat Diagn 1988;18:1337. [47] Spencer K, Souter V, Tul N, et al. A screening program for trisomy 21 at 1014 weeks using fetal nuchal translucency, maternal serum free beta-human chorionic gonadotropin and pregnancy-associated plasma protein-A. Ultrasound Obstet Gynecol 1999;13:2317. [48] Brizot ML, Snijders RJM, Bersinger NA, et al. Maternal serum pregnancy associated placental protein A and fetal nuchal translucency thickness for the prediction of fetal trisomies in early pregnancy. Obstet Gynecol 1994;84:91822.

First-Trimester Screening

251

[49] Brizot ML, Snijders RJM, Butler J, et al. Maternal serum hCG and fetal nuchal translucency thickness for the prediction of fetal trisomies in the first-trimester of pregnancy. Br J Obstet Gynaecol 1995;102:122732. [50] Malone FD, Ball RH, Nyberg DA, et al. FASTER Trial Research Consortium. First-trimester septated cystic hygroma: prevalence, natural history, and pediatric outcome. Obstet Gynecol 2005;106(2):28894. [51] Bekker MN, Haak MC, Rekoert-Hollander M, et al. Increased nuchal translucency and distended jugular lymphatic sacs on first-trimester ultrasound. Ultrasound Obstet Gynecol 2005; 25(3):23945. [52] Sharony R, Tepper R, Fejgin M. Fetal lateral neck cysts: the significance of associated findings. Prenat Diagn 2005;25(6):50710. [53] Orlandi F, Damiani G, Hallahan TW, et al. Firsttrimester screening for fetal aneuploidy: biochemistry and nuchal translucency. Ultrasound Obstet Gynecol 1997;10:3816. [54] Biagiotti R, Brizzi L, Periti E, et al. First-trimester screening for Downs syndrome using maternal serum PAPP-A and free beta-hCG in combination with fetal nuchal translucency thickness. Br J Obstet Gynaecol 1998;105: 91720. [55] Benattar C, Audibert F, Taieb J, et al. Efficiency of ultrasound and biochemical markers for Downs syndrome risk screening. A prospective study. Fetal Diagn Ther 1999;14:1127. [56] De Biasio P, Siccardi M, Volpe G, et al. Firsttrimester screening for Down syndrome using nuchal translucency measurement with betahCG and PAPP-A between 10 and 13 weeks of pregnancythe combination test. Prenat Diagn 1999;19:3603. [57] De Graaf IM, Parkrt E, Bilardo CM, et al. Early pregnancy screening for fetal aneuploidy with serum markers and nuchal translucency. Prenat Diagn 1999;19:45862. [58] Krantz DA, Hallahan TW, Orlandi F, et al. Firsttrimester Down syndrome screening using dried blood biochemistry and nuchal translucency. Obstet Gynecol 2000;96:20713. [59] Nicolaides KH. First-trimester screening for chromosomal abnormalities. Semin Perinatol 2005;29(4):1904. [60] Wapner R, Thom E, Simpson JL, et al. Firsttrimester Maternal Serum Biochemistry and Fetal Nuchal Translucency Screening (BUN) Study Group. First-trimester screening for trisomies 21 and 18. N Engl J Med 2003;349(15): 140513. [61] Malone FD, Canick JA, Ball RH, et al, First- and Second-Trimester Evaluation of Risk (FASTER) Research Consortium. First-trimester or secondtrimester screening, or both, for Downs syndrome. N Engl J Med 2005;353(19):200111. [62] Cicero S, Curcio P, Papageorghiou A, et al. Absence of nasal bone in fetuses with trisomy

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

[76]

21 at 1114 weeks of gestation: an observational study. Lancet 2001;358:16657. Cicero S, Rembouskos G, Vandecruys H, et al. Likelihood ratio for trisomy 21 in fetuses with absent nasal bone at the 1114 week scan. Ultrasound Obstet Gynecol 2004;23:21823. Prefumo F, Sethna F, Sairam S, et al. Firsttrimester ductus venosus, nasal bones, and Down syndrome in a high-risk population. Obstet Gynecol 2005;105(6):134854. Otano L, Aiello H, Igarzabal L, et al. Association between first-trimester absence of fetal nasal bone on ultrasound and Down syndrome. Prenat Diagn 2002;22:9302. Orlandi F, Bilardo CM, Campogrande M, et al. Measurement of nasal bone length at 11 14 weeks of pregnancy and its potential role in Down syndrome risk assessment. Ultrasound Obstet Gynecol 2003;22:369. Viora E, Masturzo B, Errange G, et al. Ultrasound evaluation of fetal nasal bone at 11 to 14 weeks in a consecutive series of 1906 fetuses. Prenat Diagn 2003;23:7847. Zoppi MA, Ibba RM, Axiana C, et al. Absence of fetal nasal bone and aneuploidies at firsttrimester nuchal translucency screening in unselected pregnancies. Prenat Diagn 2003;23: 496500. Nicolaides KH. Nuchal translucency and other first-trimester sonographic markers of chromosomal abnormalities. Am J Obstet Gynecol 2004;191:4567. Bekker MN, Twisk JW, van Vugt JM. Reproducibility of the fetal nasal bone length measurement. J Ultrasound Med 2004;23(12):16138. Malone FD, Ball RH, Nyberg DA, et al, FASTER Research Consortium. First-trimester nasal bone evaluation for aneuploidy in the general population. Obstet Gynecol 2005;105(4):901. Borrell A, Gonce A, Martinez JM, et al. Firsttrimester screening for Down syndrome with ductus venosus Doppler studies in addition to nuchal translucency and serum markers. Prenat Diagn 2005;25(10):9015. Faiola S, Tsoi E, Huggon IC, et al. Likelihood ratio for trisomy 21 in fetuses with tricuspid regurgitation at the 11 to 13 + 6-week scan. Ultrasound Obstet Gynecol 2005;26(1):227. Nicolaides KH, Spencer K, Avgidou K, et al. Multicenter study of first-trimester screening for trisomy 21 in 75 821 pregnancies: results and estimation of the potential impact of individual risk-orientated two-stage first-trimester screening. Ultrasound Obstet Gynecol 2005; 25(3):2216. Wald NJ, Watt HC, Hackshaw AK. Integrated screening for Downs syndrome based on tests performed during the first and secondtrimesters. N Engl J Med 1999;341:4617. Wald NJ, Rodeck C, Hackshaw AK, et al. SURUSS in perspective. Semin Perinatol 2005; 29(4):22535.

252

Nyberg et al

[77] Copel J, Bahado-Singh RO. Prenatal screening for Downs syndromea search for the familys values. N Engl J Med 1999;341:5212. [78] Benn P, Wright D, Cuckle H. Practical strategies in contingent sequential screening for Down syndrome. Prenat Diagn 2005;25(8):64552. [79] Wapner RJ. First-trimester screening: the BUN Study. Semin Perinatol 2005;29(4):2369. [80] Taipale P, Ammala M, Salonen R, et al. Twostage ultrasonography in screening for fetal anomalies at 1314 and 1822 weeks of gestation. Acta Obstet Gynecol Scand 2004;83(12): 11416. [81] Snijders RJM, Nicolaides KH. Assessment of risks. In: Snijders RJM, Nicolaides KH, editors. Ultrasound markers for fetal chromosomal defects. Carnforth (UK): Parthenon Publishing; 1996. p. 63120. [82] Nyberg DA, Luthy DA, Resta RG, et al. Ageadjusted ultrasound risk assessment for fetal Downs syndrome during the second-trimester: description of the method and analysis of 142 cases. Ultrasound Obstet Gynaecol 1998; 12:814. [83] Nyberg DA, Souter VL, El-Bastawissi A, et al. Isolated sonographic markers for detection of fetal Down syndrome in the second-trimester of pregnancy. J Ultrasound Med 2001;20(10): 105363. [84] Malone F, Nyberg DA, Vidaver J, et al. First and second trimester evaluation of risk: The role of second trimester genetic sonography. Am J Obstet Gynecol 2004;191(6):S3. [85] Cheng PJ, Chang SD, Shaw SW, et al. Nuchal translucency thickness in fetuses with chromosomal translocation at 1112 weeks of gestation. Obstet Gynecol 2005;105:105862. [86] Tul N, Spencer K, Noble P, et al. Screening for trisomy 18 by fetal nuchal translucency and maternal serum free beta hCG and PAPP-A at 1014 weeks of gestation. Prenat Diagn 1999; 19:103542. [87] Spencer K, Ong C, Skentou H, et al. Screening for trisomy 13 by fetal nuchal translucency and maternal serum free beta hCG and PAPP-A at 1014 weeks of gestation. Prenat Diagn 2000; 20:4116. [88] Spencer K, Tul N, Nicolaides KH. Maternal serum free beta hCG and PAPP-A in fetal sex chromsome defects in the first-trimester. Prenat Diagn 2000;20:3904. [89] Spencer K, Liao A, Skentou H, et al. Screening for triploidy by fetal nuchal translucency and maternal serum free -hCG and PAPP-A at 10 14 weeks of gestation. Prenat Diagn 2000;20: 4959. [90] Souka AP, Krampl E, Bakalis S, et al. Outcome of pregnancy in chromosomally normal fetuses with increased nuchal translucency in the firsttrimester. Ultrasound Obstet Gynecol 2001; 18(1):917. [91] Hyett J. Does nuchal translucency have a role

[92]

[93]

[94]

[95]

[96]

[97]

[98]

[99]

[100]

[101]

[102]

[103]

[104]

[105]

in fetal cardiac screening? Prenat Diagn 2004; 24(13):11305. McAuliffe FM, Hornberger LK, Winsor S, et al. Fetal cardiac defects and increased nuchal translucency thickness: a prospective study. Am J Obstet Gynecol 2004;191(4):148690. Bahado-Singh RO, Wapner R, Thom E, et al, First-trimester Maternal Serum Biochemistry and Fetal Nuchal Translucency Screening Study Group. Elevated first-trimester nuchal translucency increases the risk of congenital heart defects. Am J Obstet Gynecol 2005;192(5): 135761. Makrydimas G, Sotiriadis A, Huggon IC, et al. Nuchal translucency and fetal cardiac defects: a pooled analysis of major fetal echocardiography centers. Am J Obstet Gynecol 2005;192:8995. Gembruch U, Knopfle G, Chatterjee M, et al. First-trimester diagnosis of fetal congenital heart disease by transvaginal two-dimensional and Doppler echocardiography. Obstet Gynecol 1990;75:4968. Bronshtein M, Siegler E, Yoffe N, et al. Prenatal diagnosis of ventricular septal defect and overriding aorta at 14 weeks gestation, using transvaginal sonography. Prenat Diagn 1990;10: 697702. Gembruch U, Knopfle G, Bald R, et al. Early diagnosis of fetal congenital heart disease by transvaginal echocardiography. Ultrasound Obstet Gynecol 1993;3:3107. Achiron R, Rotstein Z, Lipitz S, et al. Firsttrimester diagnosis of fetal congenital heart disease by transvaginal ultrasonography. Obstet Gynecol 1994;84:6972. Zosmer N, Souter VL, Chan CSY, et al. Early diagnosis of major cardiac defects in chromosomally normal fetuses with increased nuchal translucency. Br J Obstet Gynaecol 1999;106: 82933. Atzei A, Gajewska K, Huggon IC, et al. Relationship between nuchal translucency thickness and prevalence of major cardiac defects in fetuses with normal karyotype. Ultrasound Obstet Gynecol 2005;26(2):1547. Hyett JA, Perdu M, Sharland GK, et al. Using fetal nuchal translucency to screen for major congenital cardiac defects at 1014 weeks of gestation: population based cohort study. BMJ 1999;318:815. Carvahlo JS, Moscoso G, Ville Y. First-trimester transabdominal fetal echocardiography. Lancet 1998;351:10237. Souka AP, Von Kaisenberg CS, Hyett JA, et al. Increased nuchal translucency with normal karyotype. Am J Obstet Gynecol 2005;192:100521. Hiippala A, Eronen M, Taipale P, et al. Fetal nuchal translucency and normal chromosomes: a long-term follow-up study. Ultrasound Obstet Gynecol 2001;18:1822. Sebire NJ, Snijders RJM, Hughes K, et al. Screening for trisomy 21 in twin pregnancies

First-Trimester Screening

253

[106]

[107]

[108]

[109]

[110]

[111]

[112]

[113]

[114]

[115]

[116]

[117]

[118]

[119]

[120]

[121]

by maternal age and fetal nuchal translucency thickness at 1014 weeks of gestation. Br J Obstet Gynaecol 1996;103:9991003. Bush MC, Malone FD. Down syndrome screening in twins. Clin Perinatol 2005;32(2): 37386. Sebire NJ, Hughes K, DErcole C, et al. Increased fetal nuchal translucency at 1014 weeks as a predictor of severe twin-to-twin transfusion syndrome. Ultrasound Obstet Gynecol 1997; 10:869. Shalev E, Zalele Y, Ben Ami M, et al. Firsttrimester ultrasonic diagnosis of twin reversed arterial perfusion sequence. Prenat Diagn 1992; 2:212. Langlotz H, Sauerbrei E, Murray S. Transvaginal Doppler sonographic diagnosis of an acardiac twin at 12 week gestation. J Ultrasound Med 1991;10:1759. van Eyndhoven HW, ter Brugge H. The firsttrimester ultrasonographic diagnosis of dicephalus conjoined twins. Acta Obstet Gynecol Scand 1998;77:4646. Tongsong T, Chanprapaph P, Pongsatha S. First-trimester diagnosis of conjoined twins: a report of three cases. Ultrasound Obstet Gynecol 1999;14:4347. Bonilla-Musoles F, Raga F, Bonilla Jr F, et al. Early diagnosis of conjoined twins using two-dimensional color Doppler and threedimensional ultrasound. J Natl Med Assoc 1998;90: 5526. Goldberg Y, Ben-Shlomo I, Weiner E, et al. Firsttrimester diagnosis of conjoined twins in a triplet pregnancy after IVF and ICSI: case report. Hum Reprod 2000;15:14135. Hill LM. The sonographic detection of early first-trimester conjoined twins. Prenat Diagn 1997;17:9613. Lam YH, Sin SY, Lam C, et al. Prenatal sonographic diagnosis of conjoined twins in the first-trimester: two case reports. Ultrasound Obstet Gynecol 1998;11:28991. Maymon R, Halperin R, Weinraub Z, et al. Three-dimensional transvaginal sonography of conjoined twins at 10 weeks: a case report. Ultrasound Obstet Gynecol 1998;11:2924. Meizner I, Levy A, Katz M, et al. Early ultrasonic diagnosis of conjoined twins. Harefuah 1993; 124:7414. Schmidt W, Heberling D, Kubli F. Antepartum ultrasonographic diagnosis of conjoined twins in early pregnancy. Am J Obstet Gynecol 1981; 139:9613. Efrat Z, Akinfenwa OO, Nicolaides KH. Firsttrimester determination of fetal gender by ultrasound. Ultrasound Obstet Gynecol 1999;13: 3057. Inman V, Saunders JB de CM. The ossification of the human frontal bone. J Anat 1937;71: 38394. Kennedy KA, Flick KJ, Thurmond AS. First-

[122]

[123]

[124]

[125]

[126]

[127]

[128]

[129]

[130]

[131]

[132]

[133]

[134]

[135]

[136]

[137]

trimester diagnosis of exencephaly. Am J Obstet Gynecol 1990;162:4613. Johnson SP, Sebire NJ, Snijders RJ, et al. Ultrasound screening for anencephaly at 10 14 postmenstrual weeks. Ultrasound Obstet Gynecol 1997;9:146. Bronshtein M, Timor-Tritsch I, Rottem S. Early detection of fetal anomalies. In: Timor-Tritsch I, Rottem S, editors. Transvaginal sonography. New York: Chapman & Hall; 1991. p. 32771. Gonzalez-Gomez F, Salamanca A, Padilla MC, et al. Alobar holoprosencephalic embryo detected via transvaginal sonography. Eur J Obstet Gynecol Reprod Biol 1992;47:26670. Turner CD, Silva S, Jeanty P. Prenatal diagnosis of alobar holoprosencephaly at 10 postmenstrual weeks. Ultrasound Obstet Gynecol 1999; 13:3602. Baxi L, Warren W, Collins MH, et al. Early detection of caudal regression syndrome with transvaginal scanning. Obstet Gynecol 1990; 75:4869. Nicolaides KH, Campbell S, Gabbe SG, et al. Ultrasound screening for spina bifida: cranial and cerebellar signs. Lancet 1986;2:724. Campbell J, Gilbert WM, Nicolaides KH, et al. Ultrasound screening for spina bifida: cranial and cerebellar signs in a high-risk population. Obstet Gynecol 1987;70:24750. Sebire NJ, Noble PL, Thorpe-Beeston JG, et al. Presence of the lemon sign in fetuses with spina bifida at the 1014-week scan. Ultrasound Obstet Gynecol 1997;10:4035. Bernard JP, Suarez B, Rambaud C, et al. Prenatal diagnosis of neural tube defect before 12 weeks gestation: direct and indirect ultrasonographic semeiology. Ultrasound Obstet Gynecol 1997; 10:4069. Blumenfeld Z, Siegler E, Bronshtein M. The early diagnosis of neural tube defects. Prenat Diagn 1993;13:86371. Merz E, Weber G, Bahlmann F, et al. Application of transvaginal and abdominal threedimensional ultrasound for the detection or exclusion of malformations of the fetal face. Ultrasound Obstet Gynecol 1997;9:23743. Nyberg D, Mahoney B, Kramer D. Paranasal echogenic mass: sonographic sign of bilateral complete cleft lip and palate before 20 menstrual weeks. Radiology 1992;184(3):7579. Bronshtein M, Zimmer E, Gershoni-Baruch R, et al. First- and second -trimester diagnosis of fetal ocular defects and associated anomalies: report of eight cases. Obstet Gynecol 1991; 77:4439. Feldman E, Shalev E, Weiner E, et al. Microphthalmia prenatal ultrasonic diagnosis: a case report. Prenat Diagn 1985;5:2057. Jeanty P, Dramaix-Wilmet M, Van Gansbeke D, et al. Fetal ocular biometry by ultrasound. Radiology 1982;143:5136. Monteagudo A, Timor-Tritsch IE, Friedman AH,

254

Nyberg et al

[138]

[139]

[140]

[141]

[142]

[143]

[144]

[145]

[146]

[147]

[148]

[149]

[150]

[151]

[152]

[153]

et al. Autosomal dominant cataracts of the fetus: early detection by transvaginal ultrasound. Ultrasound Obstet Gynecol 1996;8: 1048. Zimmer E, Bronshtein M, Ophir E, et al. Sonographic diagnosis of fetal congenital cataracts. Prenat Diagn 1993;13:50311. Gembruch U, Knopfle G, Bald R, et al. Early diagnosis of fetal congenital heart disease by transvaginal echocardiography. Ultrasound Obstet Gynecol 1993;3:3107. Dolkart LA, Reimers FT. Transvaginal fetal echocardiography in early pregnancy: normative data. Am J Obstet Gynecol 1991;165:68891. Gembruch U, Knopfle G, Chatterjee M, et al. First-trimester diagnosis of fetal congenital heart disease by transvaginal two-dimensional and Doppler echocardiography. Obstet Gynecol 1990;75:4968. Bronshtein M, Siegler E, Eshcoli Z, et al. Transvaginal ultrasound measurements of the fetal heart at 11 to 17 weeks. Am J Perinatol 1992;9: 3842. Achiron R, Rotstein Z, Lipitz S, et al. Firsttrimester diagnosis of fetal congenital heart disease by transvaginal ultrasonography. Obstet Gynecol 1994;84:6972. Guzman ER. Early prenatal diagnosis of gastroschisis with transvaginal ultrasonography. Am J Obstet Gynecol 1990;162:12534. Kushnir O, Izquierdo L, Vigil D, et al. Early transvaginal sonographic diagnosis of gastroschisis. J Clin Ultrasound 1990;18:1947. Fleming AD, Vintzileos AM, Rodis JF, et al. Diagnosis of fetal ectopia cordis by transvaginal ultrasound. J Ultrasound Med 1991; 10:4135. Bennett TL, Burlbaw J, Drake CK, et al. Diagnosis of ectopia cordis at 12 postmenstrual weeks using transabdominal ultrasonography with color flow Doppler. J Ultrasound Med 1991; 10:6956. Forrester MB, Merz RD. Epidemiology of abdominal wall defects, Hawaii, 19861997. Teratology 1999;60:11723. Becker R, Runkel S, Entezami M. Prenatal diagnosis of body stalk anomaly at 9 weeks. case report. Fetal Diagn Ther 2000;15:3013. Cyr DR, Mack LA, Schoenecker SA, et al. Bowel migration in the normal fetus: US detection. Radiology 1986;161:11921. Curtis JA, Watson L. Sonographic diagnosis of omphalocele in the first-trimester of fetal gestation. J Ultrasound Med 1988;7:97100. Nyberg DA, Fitzsimmons J, Mack LA, et al. Chromosomal abnormalities in fetuses with omphalocele. Significance of omphalocele contents. J Ultrasound Med 1989;8:299308. Benacerraf BR, Saltzman DH, Estroff JA, et al. Abnormal karyotype of fetuses with omphalocele: prediction based on omphalocele contents. Obstet Gynecol 1990;75:3179.

[154] Pagliano M, Mossetti M, Ragno P. Echographic diagnosis of omphalocele in the first-trimester of pregnancy. J Clin Ultrasound 1990;18:65860. [155] Brown DL, Emerson DS, Shulman LP, et al. Sonographic diagnosis of omphalocele during 10th week of gestation. AJR Am J Roentgenol 1989;153:8256. [156] Gray DL, Martin CM, Crane JP. Differential diagnosis of first-trimester ventral wall defect. J Ultrasound Med 1989;8:2558. [157] Moore K. The urogenital system. In: Moore K, editor. The developing human. Clinically oriented embryology. Philadelphia: WB Saunders Company; 1988. p. 24685. [158] Rosati P, Guariglia L. Transvaginal sonographic assessment of the fetal urinary tract in early pregnancy. Ultrasound Obstet Gynecol 1996;7: 95100. [159] Green JJ, Hobbins JC. Abdominal ultrasound examination of the first-trimester fetus. Am J Obstet Gynecol 1988;159:16575. [160] Bronshtein M, Kushnir O, Ben-Rafael Z, et al. Transvaginal sonographic measurement of fetal kidneys in the first-trimester of pregnancy. J Clin Ultrasound 1990;18:299301. [161] Bronshtein M, Bar-Hava I, Blumenfeld Z. Clues and pitfalls in the early prenatal diagnosis of late onset infantile polycystic kidney. Prenat Diagn 1992;12:2938. [162] Bronshtein M, Bar-Hava I, Blumenfeld Z. Differential diagnosis of the nonvisualized fetal urinary bladder by transvaginal sonography in the early second-trimester. Obstet Gynecol 1993;82:4903. [163] Zimmer EZ, Bronshtein M. Fetal intraabdominal cysts detected in the first and early second-trimester by transvaginal sonography. J Clin Ultrasound 1991;19:5647. [164] Stiller RJ. Early ultrasonic appearance of fetal bladder outlet obstruction. Am J Obstet Gynecol 1989;160:5845. [165] Liao AW, Sebire NJ, Geerts L, et al. Megacystis at 1014 weeks of gestation: chromosomal defects and outcome according to bladder length. Ultrasound Obstet Gynecol 2003;21(4):33841. [166] Favre R, Kohler M, Gasser B, et al. Early fetal megacystis between 11 and 15 weeks. Ultrasound Obstet Gynecol 1999;14:4026. [167] Moore K. The limbs. In: Moore K, editor. The developing human. Clinically oriented embryology. Philadelphia: WB Saunders Company; 1988. p. 35563. [168] Timor Tritsch IE, Farine D, Rosen MG. A close look at early embryonic development with the high-frequency transvaginal transducer. Am J Obstet Gynecol 1988;159:67681. [169] Mahoney B. Ultrasound evaluation of the fetal musculoskeletal system. In: Callen P, editor. Ultrasonography in obstetrics and gynecology. Philadelphia: WB Saunders Company; 1994. p. 25490. [170] Timor-Tritsch IE, Monteagudo A, Peisner DB.

First-Trimester Screening

255

[171]

[172]

[173]

[174]

High-frequency transvaginal sonographic examination for the potential malformation assessment of the 9-week to 14-week fetus. J Clin Ultrasound 1992;20:2318. Bronshtein M, Keret D, Deutsch M, et al. Transvaginal sonographic detection of skeletal anomalies in the first and early second-trimesters. Prenat Diagn 1993;13:597601. DOttavio G, Tamaro LF, Mandruzzato G. Early prenatal ultrasonographic diagnosis of osteogenesis imperfecta: a case report. Am J Obstet Gynecol 1993;169:3845. DiMaio MS, Barth R, Koprivnikar KE, et al. Firsttrimester prenatal diagnosis of osteogenesis imperfecta Type II by DNA analysis and sonography. Prenat Diagn 1993;13:58996. Brons JT, van der Harten HJ, Wladimiroff JW, et al. Prenatal ultrasonographic diagnosis of osteogenesis imperfecta. Am J Obstet Gynecol 1988;159:17681.

[175] Stephens JD, Filly RA, Callen PW, et al. Prenatal diagnosis of osteogenesis imperfecta Type II by real-time ultrasound. Hum Genet 1983;64:1913. [176] Sepulveda W, Romero R, Pryde PG, et al. Prenatal diagnosis of sirenomelus with color Doppler ultrasonography. Am J Obstet Gynecol 1994; 170:13779. [177] Sepulveda W, Corral E, Sanchez J, et al. Sirenomelia sequence versus renal agenesis: prenatal differentiation with power Doppler ultrasound. Ultrasound Obstet Gynecol 1998; 11:4459. [178] van Zalen-Sprock MM, van Vugt JM, van der Harten JJ, et al. Early second-trimester diagnosis of sirenomelia. Prenat Diagn 1995;15:1717. [179] Valenzano M, Paoletti R, Rossi A, et al. Sirenomelia. Pathological features, antenatal ultrasonographic clues, and a review of current embryogenic theories. Hum Reprod Update 1999; 5:826.

257

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 257271

Imaging of Pelvic Pain in the First Trimester of Pregnancy


Aimee D. Eyvazzadeh,
& & & & & & & &

MD

, Deborah Levine,

MD

b,*

Normal pregnancy Subchorionic hemorrhage Spontaneous abortion Molar pregnancy Corpus luteum Hemoperitoneum Ectopic pregnancy Sonographic diagnosis of ectopic pregnancy Endometrial findings Adnexal findings Use of color Doppler in diagnosis of ectopic pregnancy

& & & & & & &

Interstitial pregnancy Cervical ectopic pregnancy Scar pregnancy Ovarian and abdominal ectopic pregnancy Ovarian hyperstimulation Ovarian torsion Fibroids Urinary tract Gastrointestinal causes of pelvic pain Summary References

The noninvasive nature, safety, and reliability of ultrasonography make it the diagnostic method of choice for pregnant patients who have pelvic pain. Sonography provides information that allows for diagnosis of both pregnancy-related pain, such as a ruptured ectopic pregnancy, miscarriage, or threatened abortion; and may be useful in the diagnosis of pain unrelated to pregnancy, such as that seen in appendicitis and nephrolithiasis.

Normal pregnancy
Because of hormonal changes, rapid growth of the uterus, and increased blood flow, crampy pelvic

pain is common in early pregnancy. For the primapara, this pain can be quite worrisome. It is common for pregnant patients to present with pain in the first trimester and have normal findings on sonography. The first sonographic demonstration of early pregnancy is the intradecidual sign [Fig. 1] [13]. This is visualized as a discrete hypoechoic fluid collection with an echogenic rim that is eccentrically located in the endometrial cavity, and deviates the endometrial stripe. This is seen at 4.5 to 5 weeks of gestation [3]. Because small endometrial fluid collections can simulate the intradecidual sign, care should be taken to ensure that the collection has a well-defined echogenic rim, is just beneath the central endometrial echo, and has an

Department of Obstetrics and Gynecology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA b Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA * Corresponding author. E-mail address: dlevine@bidmc.harvard.edu (D. Levine).
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.001

ultrasound.theclinics.com

258

Eyvazzadeh & Levine

the end of the second trimester [10]. As in all early pregnancy assessments, demonstration of cardiac activity is crucial in determining prognosis.

Spontaneous abortion
First-trimester spontaneous abortion occurs in 10% to 12% of clinically recognized pregnancies [11]. Pain may be constant or intermittent and crampy over the uterus or lower back. Most women with spontaneous abortion experience vaginal bleeding. Up to 25% of all pregnant women bleed some time during pregnancy, with about half of them eventually undergoing miscarriage. The term threatened abortion is used to define bleeding in the first 20 weeks of pregnancy with a closed internal os. Ultrasound in the case of a threatened abortion is used to detect an intrauterine pregnancy and to determine if a live embryo or fetus is present. The landmarks for normal pregnancy help to distinguish between a normal early intrauterine pregnancy and a miscarriage. To ensure high specificity in our diagnosis of spontaneous abortion, the authors use generous thresholds: visualization of a yolk sac by the time the gestational sac has a mean sac diameter of 13 mm, visualization of an embryo by the time the mean sac diameter is 18 mm, and visualization of cardiac activity by the time the embryonic pole is 5 mm [12]. Between 6.5 to 10 weeks of gestation, the length of the amniotic cavity is similar to that of the embryo. At times a failed early pregnancy will present as an empty amnion sign [13] [Fig. 4]. In addition to the absolute criteria mentioned above, sonographic findings in spontaneous abortion include a thin decidual reaction (less than 2 mm), weak decidual amplitude, irregular contour

Fig. 1. Normal early pregnancy. Sagittal view of the uterus at 4 1/2 weeks gestational age shows an intradecidual sign with a small sac (arrow) eccentrically located in the endometrium.

unchanging appearance [1]. It is prudent to obtain follow-up in patients at high risk for ectopic pregnancy or patients who have symptoms in order to ensure that an intrauterine pregnancy is present. Slightly later the decidua capsularis and decidua vera are seen as two distinct hyperechoic layers surrounding the early gestational sac; this is known as the double decidual sac sign [4]. The yolk sac is the next structure to be visualized. It appears as a small hyperechoic ring within the gestational sac, and is present at 5.5 weeks [Fig. 2]. Finally, the embryo can be seen adjacent to the yolk sac. Cardiac activity can usually be observed whenever an embryonic pole is seen, but should be visualized by the time the embryonic pole is 5 mm [5,6].

Subchorionic hemorrhage
Subchorionic hemorrhage is seen on ultrasound in 4% to 22% of patients who have symptoms of pain and bleeding in early pregnancy [7]. It is caused by a partial detachment of the trophoblast from the uterine wall. On ultrasound the placental margin is displaced by anechoic or heterogeneous hypoechoic material [8]. Small echogenic structures can be found in such areas, likely due to blood clots. Because the hematoma can dissect in the potential space between the chorion and endometrial cavity, it may be visualized separate from the placenta. Because it typically conforms to the shape of the uterus, it usually has a falciform shape [Fig. 3]. A small collection likely has no clinical significance, whereas moderate or large subchorionic hematomas have a poorer prognosis [9]. Seventy percent of subchorionic hematomas resolve spontaneously by

Fig. 2. Normal early pregnancy. Sagittal view of the uterus at 5 1/2 weeks gestational age shows a yolk sac (arrow) within the intrauterine gestational sac.

Imaging of First-Trimester Pelvic Pain

259

Fig. 3. Subchorionic hematoma at 10 weeks gestational age. (A) Transabdominal sagittal image shows an intrauterine gestational sac (gs) with a subchorionic hematoma (H). (B) Transvaginal view with m-mode shows fetal pole with normal cardiac activity.

of the sac, absent double decidual sac sign, and low position of the sac.

Molar pregnancy
Molar pregnancy can be associated with pelvic pain because of either the rapid change in size of the

uterus, the size of the associated theca lutein cysts, or torsion of the ovaries caused by the theca lutein cysts [Fig. 5]. The classic sonographic appearance of a complete mole has multiple cystic spaces representing hydropic villi; however, the size of the villi is directly proportional to gestational age [14], and early molar pregnancies frequently do not have the typical sonographic appearance [15]. Other appearances that can be seen in the first trimester include an intrauterine anechoic fluid collection similar to a gestational sac, a fluid collection with a complex

Fig. 4. Incomplete miscarriage at 8 weeks by menstrual dates. A prior sonogram had shown a live embryo. Transvaginal image of the uterus shows an intrauterine gestational sac with mean sac diameter of 22 mm. An amnion (arrow) is present that measures 10 mm. A residual 1 mm embryonic pole is present (arrowhead). No yolk sac was visualized. Even without the history of a prior sonogram demonstrating a live pregnancy, a miscarriage can be diagnosed because the amnion is much larger than the residual embryonic pole.

Fig. 5. Molar pregnancy at 10 weeks gestational age in patient with pelvic pain. Sagittal transvaginal image shows the endometrial cavity (arrows) to be distended with echogenic material with multiple small cysts compatible with a molar pregnancy. Human chorionic gonadotropin level was 42,000.

260

Eyvazzadeh & Levine

Fig. 6. Transverse transvaginal image of a hemorrhagic cyst. Note the strands of internal density that have a cobweb appearance.

transmission because of the cystic composition. The internal echotexture varies, depending on the stage of hemorrhage and the amount of fluid within the cyst. This is best appreciated with transvaginal scanning. The diagnosis of a hemorrhagic cyst can be made with the presence of fibrin strands, a retracting clot, septations, and wall irregularity [16,17]. The wall of the cyst may appear thick or thin, ranging from 2 to 22 mm [Fig. 6]. The corpus luteum is a very vascular structure, and typically a ring of color flow can be demonstrated [Fig. 7] [18]. It is important to recognize that this flow is a normal finding, so as not to mistake a corpus luteum for an ectopic pregnancy. If a hemorrhagic corpus luteum cyst is the cause of the patients pain, it should be tender to direct pressure using the transvaginal probe. If it is painfree, another source for the patients pelvic pain should be sought.

echogenic mass similar to an edematous placenta, a heterogeneously thickened endometrium, and echogenic fluid-fluid levels within the endometrium [15].

Hemoperitoneum
Echogenic fluid suggests hemoperitoneum. When echogenic fluid is visualized in a patient who has positive -hCG results, this has a positive predictive value (86%93%) in the diagnosis of ectopic pregnancy [19], and may be the only endovaginal sonographic finding [20]; however, a ruptured hemorrhagic corpus luteum cyst can also result in hemoperitoneum [Fig. 8]. If the patient is clinically unstable, differentiating between a ruptured ectopic and a ruptured hemorrhagic corpus luteum is unimportant, because in either case a laparotomy is indicated. In unstable patients who have demon-

Corpus luteum
The corpus luteum is the most common adnexal mass in pregnancy, and is a common cause of pelvic pain. The pain is lateralized to the side of the cyst. Pain can be due to the size of the cyst, bleeding within the cyst, torsion, or rupture. The cyst is typically less than 6 cm in diameter, but may be larger. There is typically posterior through

Fig. 7. Ring of flow on hemorrhagic cyst. (A) Sagittal transvaginal color Doppler image of 2 cm thick-walled hemorrhagic cyst in a pregnant patient. Note the central fibrin stand mimicking a yolk sac (arrow). Note the ringof-fire appearance to the cyst. (B) Transverse, transvaginal image of the same patient in (A), showing that the mass (arrows) is located within the ovary (arrowheads). Additional images (not shown) demonstrated an intrauterine gestational sac with yolk sac. (From Swire MN, Castro-Aragon I, Levine D. Various sonographic appearances of the hemorrhagic corpus luteum cyst. Ultrasound Q 2004;20:49; with permission.)

Imaging of First-Trimester Pelvic Pain

261

Fig. 8. Ruptured hemorrhagic cyst in patient 4 weeks pregnant with pelvic pain. Sagittal view of the uterus shows hemorrhage (arrows) around the uterus (U). No intrauterine gestational sac was seen. Because of continued pain and bleeding, the patient underwent laparotomy. A ruptured hemorrhagic cyst was found. Follow-up sonogram demonstrated a live intrauterine pregnancy.

stration of hemoperitoneum, the sonographic examination may not demonstrate an ectopic pregnancy. In the clinically stable patient it is more important to carefully examine the adnexa to determine if an ectopic pregnancy is present. When free fluid is documented in the pelvis, it is helpful to obtain images of the kidneys to assess whether a large amount of hemoperitoneum is present [Fig. 9].

Ectopic pregnancy
Symptoms of an ectopic pregnancy are pelvic and abdominal pain and amenorrhea. Vaginal spotting or bleeding may be present. In a 5-year review of 98 cases who underwent surgery for ectopic, Aboud [21] showed that the most common presenting symptoms were pain (in 97%), followed by vaginal bleeding (in 79%), with the most frequent physical findings being abdominal tenderness (in 91%) and adnexal tenderness (in 54%). The combination of ultrasound and hCG level is the best way to diagnose an ectopic pregnancy. More than 1 in every 100 pregnancies in the United States is ectopic [22]. The incidence has increased fourfold from 1970 to 1992 [22]. Some causes include a higher incidence of salpingitis and an increased use of assisted reproductive techniques [23]. Patients typically present at about 5 to 6 weeks gestational age. Because menstrual dates are often inaccurate, however, an early gestational age by dates should not influence the diligence taken to diagnose an ectopic pregnancy.

The possibility of an ectopic pregnancy is low if a gestational sac is clearly documented within the uterine cavity. The incidence of heterotopic pregnancy (the occurrence of intrauterine and extrauterine pregnancy) ranges from 1/2,100 to 1/30,000 [24,25]. Of importance, the incidence is as high as 2.9% in the assisted fertilization population [26,27]. Therefore, although visualization of an intrauterine gestation is crucial, careful attention to the adnexa is always important. Ectopic pregnancy should be suspected in patients who present with a positive pregnancy test with absence of an intrauterine pregnancy on ultrasound. In general, an intrauterine gestational sac is expected to be visualized when -hCG is 1000 mIU/ml (Second International Standard,) or 2000 mIU/ml international reference preparation (IRP) [28,29]. It should be emphasized that the majority of studies of b-hCG in early pregnancy evaluated normal early pregnancy, and described an intrauterine gestational sac as any collection of fluid in the endometrial cavity. Small fluid collections of 2 mm without a decidual reaction were considered sufficient to describe an early gestational sac. It should be noted that this type of fluid collection can be caused by a decidual cyst or even a pseudosac, and therefore may not represent a normal intrauterine pregnancy; however, these values are helpful in triaging patients. When -hCG is below the discriminatory zone (2000 mIU/mL, IRP) and no intrauterine gestation is present, the diagnosis could be an early intrauterine pregnancy, a miscarriage, or an ectopic pregnancy, and therefore close follow-up is indicated [30]. When the -hCG value is above the discriminatory zone, one can expect to see an intrauterine

Fig. 9. Hemoperitoneum in patient with ectopic pregnancy. Oblique sagittal view of right upper quadrant in patient with pelvic pain in the first trimester shows fluid (arrows) around the liver and kidneys, consistent with a large amount of hemoperitoneum.

262

Eyvazzadeh & Levine

gestational sac; however, even without visualization of a sac there could still be a very early normal intrauterine pregnancy. Technical quality of the examination, presence of fibroids, intrauterine contraceptive devices, large hemorrhage, and multiple gestation may contribute to nonvisualization of an early sac [3032]; however, none of these factors may be present, and follow-up may still reveal a normal early pregnancy [30]. Because of this, and because stable patients can be watched rather than treated [3335], it is reasonable to follow stable patients who have a nonvisualized gestational sac with serial -hCG and ultrasound rather than immediately treating with methotrexate or laparotomy. A normal pregnancy shows a doubling time of the -hCG value of 2 days (range 1.22.2 days) [36]. This doubling time is increased in ectopic pregnancy. If the -hCG values rise abnormally (<60% increase over 48 hours and not steadily declining), the patient is presumed to have an ectopic pregnancy. The most common location for ectopic pregnancy is in the fallopian tubes, occurring in up to

97% of the cases. Of these, 75% to 80% are located in the ampullary region, 10% in the isthmic portion, 5% in the fimbrial portion, and 2% to 4% in the interstitial portion. Uncommon locations include the ovary, abdomen, cervix, and uterine scars [37,38]. Because most ectopic pregnancies are located within the tubes, it is important to scan above and below the ovaries and between the uterus and ovaries.

Sonographic diagnosis of ectopic pregnancy Endometrial findings


Small fluid collections without an echogenic rim can be present. These decidual cysts are typically located at the junction of the endometrium with the myometrium, and were originally reported as being highly specific for ectopic pregnancy [39], but are now known to be neither specific nor sensitive [40,41]. When fluid is seen centrally in the endometrial cavity, this is termed a pseudosac [Fig. 10]. This fluid collection represents blood in the endometrial cavity, which can be present in

Fig. 10. Pseudosac in patient with ectopic pregnancy at 5 weeks gestational age. Transvaginal sagittal (A) and transverse (B) images show fluid (arrows) centrally located within the endometrial cavity. Oblique image in the left adnexa (C ) shows a ringlike mass (arrowheads) with a faint yolk sac and some free fluid (f ). The mass was separate from the left ovary (not shown). A left-sided ectopic pregnancy was confirmed at laparotomy.

Imaging of First-Trimester Pelvic Pain

263

both intrauterine and ectopic pregnancies. The pseudosac has only one layer corresponding to the endometrial decidual reaction, compared with the double decidual sac sign seen in early intrauterine pregnancy [4].

Adnexal findings
The most specific finding for ectopic pregnancy is the presence of a live extrauterine pregnancy [Fig. 11]; however, this pathognomonic sign is present only in only 8% to 26% of ectopic pregnancies on transvaginal sonogram [42]. The next

most specific sign is an extrauterine gestational sac containing a yolk sac, with or without an embryo [see Fig. 10] [19]; however, care should be taken not to confuse a hemorrhagic cyst with debris mimicking a yolk sac or embryo [see Fig. 7]. An extra-ovarian tubal ring is 40% to 68% sensitive for ectopic pregnancy [see Fig. 10] [43,44]. Slightly less specific but most common is a complex adnexal mass separate from the ovary [19,20, 31,4355]. These should be distinguished from a hemorrhagic corpus luteum cyst arising from the ovary. The transvaginal transducer can be used real-time to determine if the echogenic ring

Fig. 11. Live ectopic pregnancy. (A) Transverse transabdominal image shows a left- sided gestational sac (arrow) adjacent to the uterus (UT ), clearly separate from the left ovary (LT O). (B) Transverse transvaginal image shows the ectopic pregnancy adjacent to the left ovary. (C ) M-Mode demonstrates cardiac activity.

264

Eyvazzadeh & Levine

moves with or is independent of, the ovary. Another sonographic finding that can help distinguish the corpus luteum from the adnexal ring of an ectopic pregnancy is the relative echogenicity of the wall of the corpus luteum compared with that of a tubal ectopic and of the endometrium. The wall of a corpus luteum is less echogenic when compared with the wall of the tubal ring associated with an ectopic pregnancy, and is less echogenic compared with the endometrium [56,57]. If the diagnosis of an adherent ectopic pregnancy or an exophytic ovarian cyst cannot be confirmed and the patient is stable, a follow-up examination is reasonable, because an intrauterine pregnancy may be seen on follow-up, and a hemorrhagic cyst is expected to undergo evolution. The least specific finding of ectopic pregnancy is the presence of any adnexal mass other than a simple cyst. Even a complex cyst in the ovary is more likely to be the corpus luteum than an ectopic pregnancy.

tween the ovarian and uterine arteries, rupture in this area may prove fatal [64]. The diagnosis is suggested when what appears to be an intrauterine pregnancy is visualized high in the fundus and is not surrounded in all planes by 5 mm of myometrium [Fig. 12] [44,65]. These can be treated with laparotomy, systemic methotrexate [66], or transvaginal, sonographically guided injection of potassium chloride [67].

Cervical ectopic pregnancy


Cervical ectopic pregnancy occurs in fewer than 1% of all ectopics [68,69]. The sonographic diagnosis is made when a gestational sac with peritrophoblastic flow or a live embryo is identified within the cervix. When a gestational sac with a yolk sac or embryo is seen within the cervix without a heartbeat, the differential diagnosis includes spontaneous abortion and cervical ectopic. Followup scanning allows for differentiation; in cases of ectopic pregnancy the sac does not change in position, whereas in spontaneous abortion, the sac shape and position will change. Patients who have cervical ectopics tend to bleed profusely because the cervix does not have contractile tissue. Therefore treatment by dilatation and curettage is more risky than treatment of an intrauterine pregnancy. Because of these risks, in the past cervical ectopics were often treated with hysterectomy. Newer conservative therapies include sonographically guided local potassium chloride injection [67,70,71], systemic or local methotrexate [7174], or preoperative uterine artery embolization before dilatation and evacuation [71,75].

Use of color Doppler in diagnosis of ectopic pregnancy


Using color Doppler flow, uterine or extrauterine sites of vascular color can be identified in a characteristic placental shape, the so-called ring-offire pattern, and a high-velocity, low-impedance flow pattern may also be identified that is compatible with placental perfusion [58]. A ring of fire has been described as characterizing the appearance of flow around an ectopic pregnancy; however, the corpus luteum is also very vascular and can have a similar appearance [see Fig. 7] [59,60]. Color Doppler is most helpful when an extra ovarian mass has not yet been found, because use of Doppler may allow for detection of an ectopic surrounded by loops of bowel. Luteal flow can be helpful in identifying an ectopic, because about 90% of ectopic pregnancies occur on the same side as luteal flow [61].

Scar pregnancy
Scars in the uterus can be sites for implantation of pregnancy. Cesarean section scar pregnancy is being increasingly reported [76]. There is complete embedding of the gestational sac in the myometrium. The myometrium between the bladder and the sac becomes thinner or disappears because of distension of the sac. Only the thin, serosal layer is apparent. Criteria used for diagnosis are an empty uterus, empty cervical canal, and development of the sac in the anterior part of the lower uterine segment [Fig. 13] [77]. Current non- and minimally invasive treatments include sonographically guided methotrexate or potassium chloride injection [67,78], or intramuscular methotrexate [79]. Definitive treatment of a cesarean scar pregnancy is by laparotomy and hysterotomy, with repair of the accompanying uterine scar dehiscence [80]. Other procedures that scar the uterus put the patient at increased risk for scar pregnancy. For example, a pregnancy can implant in a myomectomy scar [60].

Interstitial pregnancy
Interstitial pregnancies represent 2% to 4% of ectopic pregnancies [62]. These pregnancies are associated with a higher morbidity and mortality than other tubal pregnancies [63]. Although some term these cornual pregnancies, this term is best used if pregnancy occurs in a bicornuate uterus. The high morbidity from these pregnancies is caused by the fact that the interstitial portion of the tube dilates more freely and painlessly than the rest of the tube, leading to later clinical presentation than the typical ectopic pregnancy, and the potential for massive hemorrhage. Rupture occurs later in interstitial ectopics, usually between 8 and 16 weeks. Because the implantation site may be located be-

Imaging of First-Trimester Pelvic Pain

265

Fig. 12. Ruptured isthmic pregnancy at 11 weeks gestational age. (A) Sagittal transabdominal image shows a gestational sac (arrowheads) located high in the uterus, with the superior portion of the sac (thin arrows) bulging beyond the confines of the uterus. (B,C ) Sagittal transvaginal images show blood (B) surrounding the uterus (arrows). The gestational sac (arrowheads) is again noted to be high in the uterus, without myometrium around the superior portion of the sac. At surgery a ruptured isthmic pregnancy was found.

Fig. 13. Twin gestation in cesarean section scar. (A) Transabdominal view of a retroflexed uterus shows two gestational sacs (A,B) in the region of a prior cesarean section scar. (B) Transvaginal image shows embryos within the gestational sacs. These are in the anterior myometrium, separate from the endometrial cavity. The patient was given systemic methotrexate and the embryos were injected with potassium chloride.

266

Eyvazzadeh & Levine

Ovarian and abdominal ectopic pregnancy


Ovarian pregnancies usually appear as an ovarian cyst with a wide, echogenic outside ring. A yolk sac or embryo is less commonly seen, with the appearance of the contents lagging in comparison with the gestational age. Abdominal pain before 7 weeks gestational age is typically present [81]. Abdominal pregnancies are rare. The pregnancy typically develops in the ligaments of the ovary, usually the broad ligament. It can then obtain blood supply from the omentum and abdominal organs. Sonographically, the pregnancy is seen separate from the uterus, adnexa, and ovaries. Treatment is by laparotomy or laparoscopy [82]. Abdominal pregnancy can result in a life-threatening emergency. However, if diagnosed late in gestation, a viable pregnancy can result.

enzymes, dyspnea, anasarca, or acute renal failure [83]. These patients may benefit by sonographically guided drainage of hyperstimulated ovaries to relieve the abdominal pain and distension they experience. One problem in the diagnosis of ovarian hyperstimulation is that if the patient is pregnant, ectopic pregnancy is still a possibility. If the pain is severe, torsion may also be present [Fig. 14].

Ovarian torsion
Ovarian torsion is the most frequent and most serious complication of benign ovarian cysts during pregnancy. Torsion is most common in the first trimester, and may result in cyst rupture into the peritoneal cavity. Symptoms include abdominal pain and tenderness that are usually sudden in onset, and localized to the torsed ovary. Ultrasound frequently demonstrates an adnexal mass, and may show altered blood flow on Doppler studies. Doppler of ovarian torsion can be difficult because the ovaries have a dual blood supply, from the ovarian artery laterally and from the ovarian branch of the uterine artery medially. Presence of venous flow is predictive of ovarian viability [84]. In difficult cases, the authors have found MRI to

Ovarian hyperstimulation
Ovarian hyperstimulation is diagnosed by the presence of abdominal pain, enlargement of the ovary greater than 5 cm, and ascites or hydrothorax [83]. In addition, one of the following criteria has to be met: hematocrit 45% or more, white blood cells greater than 15,000/ml, oliguria, elevated liver

Fig. 14. Hyperstimulated torsed ovary in patient 7 weeks pregnant with severe pain. (A) Transverse sonogram demonstrates an enlarged left ovary measuring 11 cm with multiple cysts consistent with the patients history of hyperstimulation. (B) Color Doppler shows flow in the ovary. Pulsed Doppler (not shown) demonstrated both arterial and venous flow. (C ) Image at surgery shows torsion of the hyperstimulated ovary.

Imaging of First-Trimester Pelvic Pain

267

Fig. 15. Ovarian torsion in patient with twins after in vitro fertilization with severe intermittent right lower quadrant pain (11 weeks pregnant). Sonogram (not shown) had demonstrated enlarged ovaries with flow. Due to severe pain, an MR was performed. Axial fat saturated, T2-weighted, single-shot, fast-spin echo image shows large ovaries, right (arrows) greater than left (arrowheads), with multiple follicles, consistent with history of hyperstimulation. The stroma of the right ovary is brighter than the left, consistent with edema caused by torsion. At surgery the ovary was edematous with 360 of torsion. (From Levine D, Pedrosa I. MR imaging of the maternal abdomen and pelvis in pregnancy. In: Levine D, editor. Atlas of fetal MRI. Boca Raton (FL): Taylor & Francis Group; 2005. p. 216; 2005. Reproduced by permission of Routledge/Taylor & Francis Group, LLC.)

on the ureters, which can partially obstruct the normal downward flow of urine. Pregnancy also increases the risk of reflux of urine by causing the ureters to dilate and reducing the muscle contractions that propel urine downwards into the bladder. These changes make urinary tract infections very common. Many women who have bacteriuria will develop pyelonephritis during pregnancy. Both cystitis and pyelonephritis can be a cause of pelvic pain. Although hydronephrosis of pregnancy can cause flank pain, is not a typical cause of pelvic pain. The appearance of dilated tracts can be confusing in pregnancy, however, because hydronephrosis can be caused by physiologic dilation of pregnancy, nephrolithiasis, or structural abnormalities. Nephrolithiasis is an uncommon but important condition in pregnant women. The most common presenting complaint is flank pain.; however, when the stone is at the ureterovescicle junction, the patient may present with pelvic pain [Fig. 17]. The incidence of nephrolithiasis in pregnancy is about 1 per 2000 pregnancies [88]. If the ureter is dilated and a stone is not visualized, it can be helpful to assess for urinary jets in the bladder; however, these jets can be absent in cases without stones, and present with nonobstructing stones [89,90].

be helpful in confirming the diagnosis of torsion [Fig. 15] [85].

Gastrointestinal causes of pelvic pain


Acute appendicitis is the most common nonobstetrical surgical condition of the abdomen complicating pregnancy. Although the incidence of appendicitis occurring in pregnant women is considered to be the same as in nonpregnant women, the signs and symptoms and the laboratory find-

Fibroids
Uterine fibroids are commonly found during pregnancy. One in 500 pregnant women is admitted for a complication related to a fibroid [86]. Inconsistency of uterine size and gestational dates in a pregnant patient who has acute abdominal pain may be the first sign of leiomyoma. Fibroids during pregnancy occasionally undergo red degeneration that is caused by hemorrhagic infarction. The symptoms and signs are focal pain, with tenderness on palpation and sometimes low-grade fever. Moderate leukocytosis is common. The greatest increase in volume of myomas occurs before the 10th week of gestation. Fibroids either remain unchanged or increase in size in the first trimester as a response to increased estrogen [87]. The sonographic diagnosis of a degenerating fibroid is made when the patient experiences pain when the probe is placed over the fibroid. At times a lucent center will be visualized [Fig. 16].

Urinary tract
The urinary system undergoes many changes during pregnancy. The enlarging uterus puts pressure

Fig. 16. Degenerating fibroid in patient 10 weeks pregnant. Transabdominal view of the uterus shows a gestational sac (GS) and an anterior fibroid (arrowhead) with a small lucency centrally (thin arrow). The patient was focally tender over the fibroid.

268

Eyvazzadeh & Levine

Fig. 17. Obstructing stone at 13 weeks gestational age. (A) Sagittal view of the right kidney (arrowheads) demonstrates hydronephrosis. (B) Sagittal view of the uteropelvic junction demonstrates dilation of the proximal right ureter (U, long arrows). (C ) Transverse view of the bladder with color shows a left ureteral jet but no right jet was demonstrated. (D) View of the right ureterovescicle junction demonstrates a small stone (small arrow) without a shadow.

ings usually associated with appendicitis in the nonpregnant condition are frequently unreliable during pregnancy [91]. On ultrasound, the abnormal appendix is visualized as a noncompressible tubular structure measuring 6 mm or greater in the region of the patients pain [Fig. 18]. An appendicolith or periappendiceal fluid may be visualized. If ultrasound diagnosis is inadequate, MRI can be helpful in assessing the etiology of right-sided pain in pregnancy [92,93]. Crohns disease can also be a cause of pelvic pain in pregnancy. Most pregnant women who have a history of inflammatory bowel disease have uneventful pregnancies, and exacerbations of disease can be controlled with medical therapy. Although it is rare for the new onset of inflammatory bowel disease to be diagnosed during pregnancy [94], when a relapse of Crohns disease occurs during pregnancy, it typically will occur during the first

Fig. 18. Appendicitis in pregnancy. Oblique view in the right lower quadrant demonstrates the dilated appendix (arrows).

Imaging of First-Trimester Pelvic Pain

269

trimester [95]. Imaging can start with ultrasound, but frequently another modality is needed, such as MRI or CT.

[15]

Summary
Pelvic pain during the first trimester of pregnancy can pose a challenge to the clinician. Ultrasound is a very important imaging modality in evaluating these patients.

[16]

[17]

References
[1] Chiang G, Levine D, Swire M, et al. The intradecidual sign: is it reliable for diagnosis of early intrauterine pregnancy? AJR Am J Roentgenol 2004;183:72531. [2] Yeh HC. Efficacy of the intradecidual sign and fallacy of the double decidual sac sign in the diagnosis of early intrauterine pregnancy. Radiology 1999;210:57982. [3] Yeh HC, Goodman JD, Carr L, et al. Intradecidual sign: a US criterion of early intrauterine pregnancy. Radiology 1986;161:4637. [4] Bradley WG, Fiske CE, Filly RA. The double sac sign of early intrauterine pregnancy: use in exclusion of ectopic pregnancy. Radiology 1982;143: 2236. [5] Levi CS, Lyons EA, Lindsay DJ. Ultrasound in the first trimester of pregnancy. Radiol Clin North Am 1990;28:1938. [6] Levi CS, Lyons EA, Zheng XH, et al. Endovaginal US: demonstration of cardiac activity in embryos of less than 5.0 mm in crown-rump length. Radiology 1990;176:714. [7] Pearlstone M, Baxi L. Subchorionic hematoma: a review. Obstet Gynecol Surv 1993;48:658. [8] Mantoni M, Pedersen JF. Intrauterine haematoma. An ultrasonic study of threatened abortion. Br J Obstet Gynaecol 1981;88:4751. [9] Ball RH, Ade CM, Schoenborn JA, et al. The clinical significance of ultransonographically detected subchorionic hemorrhages. Am J Obstet Gynecol 1996;174:9961002. [10] Nagy S, Bush M, Stone J, et al. Clinical significance of subchorionic and retroplacental hematomas detected in the first trimester of pregnancy. Obstet Gynecol 2003;102:94100. [11] Simpson J, Carson S. Genetic and non-genetic casues of spontaneous abortions. In: Sciarra J, editor. Gynecology and obstetrics. Philadelphia: JB Lippencott; 1995. p. 20. [12] Filly RA. Ultrasound evaluation during the first trimester. In: Callen PW, editor. Ultrasonography in obstetrics and gynecology. Philadelphia: WB Saunders; 1998. p. 6385. [13] McKenna KM, Feldstein VA, Goldstein RB, et al. The empty amnion: a sign of early pregnancy failure. J Ultrasound Med 1995;14:11721. [14] Szulman AE, Surti U. The syndromes of hydatid-

[18] [19]

[20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

iform mole. II. Morphologic evolution of the complete and partial mole. Am J Obstet Gynecol 1978;132:207. Lazarus E, Hulka C, Siewert B, et al. Sonographic appearance of early complete molar pregnancies. J Ultrasound Med 1999;18:58994. Chiang G, Levine D. Imaging of adnexal masses in pregnancy. J Ultrasound Med 2004;23: 80519. Patel MD, Feldstein VA, Filly RA. The likelihood ratio of sonographic findings for the diagnosis of hemorrhagic ovarian cysts. J Ultrasound Med 2005;24:60715. Jain KA. Sonographic spectrum of hemorrhagic ovarian cysts. J Ultrasound Med 2002;21:87986. Russell SA, Filly RA, Damato N. Sonographic diagnosis of ectopic pregnancy with endovaginal probes: what really has changed? J Ultrasound Med 1993;12:14551. Nyberg DA, Hughes MP, Mack LA, et al. Extrauterine findings of ectopic pregnancy of transvaginal US: importance of echogenic fluid. Radiology 1991;178:8236. Aboud E. A five-year review of ectopic pregnancy. Clin Exp Obstet Gynecol 1997;24:1279. From the Centers for Disease Control and Prevention. Ectopic pregnancyUnited States, 1990 1992. JAMA 1995;273:533. Chow WH, Daling JR, Cates Jr W, et al. Epidemiology of ectopic pregnancy. Epidemiol Rev 1987; 9:7094. DeVoe RW, Pratt JH. Simultaneous intra- and extrauterine pregnancy. Am J Obstet Gynecol 1948;56:1119. Richards SR, Stempel LE, Carlton BD. Heterotopic pregnancy: reappraisal of incidence. Am J Obstet Gynecol 1982;142:92830. Bello GV, Schonholz D, Moshirpur J, et al. Combined pregnancy: the Mount Sinai experience. Obstet Gynecol Surv 1986;41:60313. Berger MJ, Taymor ML. Simultaneous intrauterine and tubal pregnancies following ovulation induction. Am J Obstet Gynecol 1972;113: 8123. Cacciatore B, Ulf-hakan S, Ylostalo P. Diagnosis of ectopic pregnancy by vaginal ultrasonography in combination with a discriminatory serum hCG level of 1000 IU/l (IRP). Br J Obstet Gynaecol 1990;97:9048. Barnhart K, Mennuti MT, Benjamin I, et al. Prompt diagnosis of ectopic pregnancy in an emergency department setting. Obstet Gynecol 1994;84:10105. Mehta TS, Levine D, Beckwith B. Treatment of ectopic pregnancy: is a human chorionic gonadotropin level of 2,000 mIU/mL a reasonable threshold? Radiology 1997;205:56973. Bateman BG, Nunley WC, Kolp LA, et al. Vaginal sonography findings and hCG dynamics of early intrauterine and tubal pregnancies. Obstet Gynecol 1990;75:4217. Goldstein SR, Snyder JR, Watson C, et al. Very

270

Eyvazzadeh & Levine

[33]

[34]

[35]

[36] [37] [38] [39]

[40] [41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

early pregnancy detection with endovaginal ultrasound. Obstet Gynecol 1988;72:2004. Sauer MV, Gorrill MJ, Rodi IA, et al. Nonsurgical management of unruptured ectopic pregnancy: an extended clinical trial. Fertil Steril 1987;48: 7525. Fernandez H, Rainhorn JD, Papiernik E, et al. Spontaneous resolution of ectopic pregnancy. Obstet Gynecol 1988;71:1714. Atri M, Bret PM, Tulandi T. Spontaneous resolution of ectopic pregnancy: initial appearance and evolution at transvaginal US. Radiology 1993;186:836. Batzer R. Guidelines for choosing a pregnancy test. Contemp Ob Gyn 1985;30:57. Breen JL. A 21 year survey of 654 ectopic pregnancies. Am J Obstet Gynecol 1970;106:100419. Dialani V, Levine D. Ectopic pregnancy: a review. Ultrasound Q 2004;20:10517. Ackerman TE, Levi CS, Dashefsky SM, et al. Interstitial line: sonographic finding in interstitial (cornual) ectopic pregnancy. Radiology 1993;189:837. Yeh HC. Some misconceptions and pitfalls in ultrasonography. Ultrasound Q 2001;17:12955. Frates MC, Laing FC. Sonographic evaluation of ectopic pregnancy: an update. AJR Am J Roentgenol 1995;165:2519. Nyberg DA, Mack LA, Jeffrey Jr RB, et al. Endovaginal sonographic evaluation of ectopic pregnancy: a prospective study. AJR Am J Roentgenol 1987;149:11816. Atri M, de Stempel J, Bret PM. Accuracy of transvaginal ultrasonography for detection of hematosalpinx in ectopic pregnancy. J Clin Ultrasound 1992;20:25561. Fleischer AC, Pennell RG, McKee MS, et al. Ectopic pregnancy: features at transvaginal sonography. Radiology 1990;174:3758. Cacciatore B. Can the status of tubal pregnancy be predicted with transvaginal sonography? A prospective comparison of sonographic, surgical, and serum hCG findings. Radiology 1990;177: 4814. Nyberg DA, Mack LA, Laing FC, et al. Early pregnancy complications: endovaginal sonographic findings correlated with human chorionic gonadotropin levels. Radiology 1988;167:61922. Cacciatore B, Stenman UH, Ylostalo P. Early screening for ectopic pregnancy in high-risk symptom-free women. Lancet 1994;343:5178. Cacciatore B, Stenman U-H, Ylostalo P. Comparison of abdominal and vaginal sonography in suspected ectopic pregnancy. Obstet Gynecol 1989; 73:7704. Dashefsky SM, Lyons EA, Levi CS, et al. Suspected ectopic pregnancy: endovaginal and transvesical US. Radiology 1988;169:1814. Thorsen MK, Lawson TL, Aiman EJ, et al. Diagnosis of ectopic pregnancy: endovaginal vs transabdominal sonography. AJR Am J Roentgenol 1990;155:30710.

[51] Kivikoski AI, Martin CM, Smeltzer JS. Transabdominal and transvaginal ultrasonography in the diagnosis of ectopic pregnancy: a comparative study. Am J Obstet Gynecol 1990;163:1238. [52] Frates MC, Brown DL, Doubilet PM, et al. Tubal rupture in patients with ectopic pregnancy: diagnosis with transvaginal US. Radiology 1994;191: 76972. [53] Brown DL, Doubilet PM. Transvaginal sonography for diagnosing ectopic pregnancy: positivity criteria and performance characteristics. J Ultrasound Med 1994;13:25966. [54] Stiller RJ, Haynes de Regt R, Blair E. Transvaginal ultrasonography in patients at risk for ectopic pregnancy. Am J Obstet Gynecol 1989;161: 9303. [55] Filly RA. Ectopic pregnancy: the role of sonography. Radiology 1987;162:6618. [56] Frates MC, Visweswaran A, Laing FC. Comparison of tubal ring and corpus luteum echogenicities: a useful differentiating characteristic. J Ultrasound Med 2001;20:2731. [57] Stein MW, Ricci ZJ, Novak L, et al. Sonographic comparison of the tubal ring of ectopic pregnancy with the corpus luteum. J Ultrasound Med 2004;23:5762. [58] Emerson DS, Cartier MS, Altieri LA, et al. Diagnostic efficacy of endovaginal color Doppler flow imaging in an ectopic pregnancy screening program. Radiology 1992;183:41320. [59] Levine D. Ectopic pregnancy. In: Callen PW, editor. Ultrasonography in obstetrics and gynecology. Pennsylvania: WB Saunders Co.; 2000. p. 91234. [60] Swire MN, Castro-Aragon I, Levine D. Various sonographic appearances of the hemorrhagic corpus luteum cyst. Ultrasound Q 2004;20: 4558. [61] Taylor KJ, Meyer WR. New techniques in the diagnosis of ectopic pregnancy. Obstet Gynecol Clin North Am 1991;18:3954. [62] Bouyer J, Coste J, Fernandez H, et al. Sites of ectopic pregnancy: a 10 year population-based study of 1800 cases. Hum Reprod 2002;17: 322430. [63] Jafri SZ, Loginsky SJ, Bouffard JA, et al. Sonographic detection of interstitial pregnancy. J Clin Ultrasound 1987;15:2537. [64] Lee GS, Hur SY, Kown I, et al. Diagnosis of early intramural ectopic pregnancy. J Clin Ultrasound 2005;33:1902. [65] Chen GD, Lin MT, Lee MS. Diagnosis of interstitial pregnancy with sonography. J Clin Ultrasound 1994;22:43942. [66] Fernandez H, Benifla JL, Lelaidier C, et al. Methotrexate treatment of ectopic pregnancy: 100 cases treated by primary transvaginal injection under sonographic control. Fertil Steril 1993;59:7737. [67] Doubilet PM, Benson CB, Frates MC, et al. Sonographically guided minimally invasive treatment of unusual ectopic pregnancies. J Ultrasound Med 2004;23:35970.

Imaging of First-Trimester Pelvic Pain

271

[68] Celik C, Bala A, Acar A, et al. Methotrexate for cervical pregnancy. A case report. J Reprod Med 2003;48:1302. [69] Ushakov FB, Elchalal U, Aceman PJ, et al. Cervical pregnancy: past and future. Obstet Gynecol Surv 1997;52:4559. [70] Monteagudo A, Tarricone NJ, Timor-Tritsch IE, et al. Successful transvaginal ultrasound-guided puncture and injection of a cervical pregnancy in a patient with simultaneous intrauterine pregnancy and a history of a previous cervical pregnancy. Ultrasound Obstet Gynecol 1996;8: 3816. [71] Frates MC, Benson CB, Doubilet PM, et al. Cervical ectopic pregnancy: results of conservative treatment. Radiology 1994;191:7735. [72] Jurkovic D, Hacket E, Campbell S. Diagnosis and treatment of early cervical pregnancy: a review and a report of two cases treated conservatively. Ultrasound Obstet Gynecol 1996;8:37380. [73] Stovall TG, Ling FW. Ectopic pregnancy. Diagnostic and therapeutic algorithms minimizing surgical intervention. J Reprod Med 1993;38: 80712. [74] Sherer DM, Abramowicz JS, Thompson HO, et al. Comparison of transabdominal and endovaginal sonographic approaches in the diagnosis of a case of cervical pregnancy successfully treated with methotrexate. J Ultrasound Med 1991;10: 40911. [75] Meyerovitz MF, Lobel SM, Harrington DP, et al. Preoperative uterine artery embolization in cervical pregnancy. J Vasc Interv Radiol 1991;2:957. [76] Jurkovic D, Hillaby K, Woelfer B, et al. Firsttrimester diagnosis and management of pregnancies implanted into the lower uterine segment Cesarean section scar. Ultrasound Obstet Gynecol 2003;21:2207. [77] Li SP, Wang W, Tang XL, et al. Cesarean scar pregnancy: a case report. Chin Med J (Engl) 2004;117: 3167. [78] Seow KM, Huang LW, Lin YH, et al. Cesarean scar pregnancy: issues in management. Ultrasound Obstet Gynecol 2004;23:24753. [79] Haimov-Kochman R, Sciaky-Tamir Y, Yanai N, et al. Conservative management of two ectopic pregnancies implanted in previous uterine scars. Ultrasound Obstet Gynecol 2002;19:6169. [80] Fylstra DL. Ectopic pregnancy within a cesarean scar: a review. Obstet Gynecol Surv 2002;57: 53743.

[81] Comstock C, Huston K, Lee W. The ultrasonographic appearance of ovarian ectopic pregnancies. Obstet Gynecol 2005;105:425. [82] Siow A, Chern B, Soong Y. Successful laparoscopic treatment of an abdominal pregnancy in the broad ligament. Singapore Med J 2004;45: 889. [83] Practice Committee of the American Society of Reproductive Medicine. Ovarian hyperstimulation syndrome. Fertil Steril 2004;82(Suppl 1): S816. [84] Fleischer AC, Stein SM, Cullinan JA, et al. Color Doppler sonography of adnexal torsion. J Ultrasound Med 1995;14:5238. [85] Levine D, Pedrosa I. MR imaging of the maternal abdomen and pelvis in pregnancy. In: Levine D, editor. Atlas of fetal MRI. Boca Raton (FL): Taylor & Francis Group; 2005. p. 17592. [86] Katz VL, Dotters DJ, Droegemeuller W. Complications of uterine leiomyomas in pregnancy. Obstet Gynecol 1989;73:5936. [87] Lev-Toaff AS, Coleman BG, Arger PH, et al. Leiomyomas in pregnancy: sonographic study. Radiology 1987;164:37580. [88] Hendricks SK, Ross SO, Krieger JN. An algorithm for diagnosis and therapy of management and complications of urolithiasis during pregnancy. Surg Gynecol Obstet 1991;172:4954. [89] Deyoe LA, Cronan JJ, Breslaw BH, et al. New techniques of ultrasound and color Doppler in the prospective evaluation of acute renal obstruction. Do they replace the intravenous urogram? Abdom Imaging 1995;20:5863. [90] Geavlete P, Georgescu D, Cauni V, et al. Value of duplex Doppler ultrasonography in renal colic. Eur Urol 2002;41:718. [91] Tamir IL, Bongard FS, Klein SR. Acute appendicitis in the pregnant patient. Am J Surg 1990; 160:5715 [discussion: 5756]. [92] Eyvazzadeh AD, Pedrosa I, Rofsky NM, et al. MRI of right-sided abdominal pain in pregnancy. AJR Am J Roentgenol 2004;183:90714. [93] Pedrosa I, Levine D, Eyvazzadeh AD, et al. MRI evaluation of suspected acute appendicitis in pregnancy. Radiology, in press. [94] Goettler CE, Stellato TA. Initial presentation of Crohns disease in pregnancy: report of a case. Dis Colon Rectum 2003;46:40610. [95] Hill J, Clark A, Scott NA. Surgical treatment of acute manifestations of Crohns disease during pregnancy. J R Soc Med 1997;90:646.

273

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 273291

Prenatal Diagnosis of Congenital Heart Disease: Where Are We Now?


Wesley Lee,
& &

MD

a,b,*

, Christine H. Comstock,
&

MD

a,c

&

Fetal cardiac screening What influences the prenatal detection of congenital heart disease? Factors influencing adequate examination of the heart Unrecognized abnormalities Evolution of cardiac abnormalities Standardized images may not demonstrate the anomalous heart Fetal cardiac screening guidelines Basic examination Extended basic examination

& &

Sonographic detection of selected cardiac anomalies Ventricular septal defects Atrioventricular septal defects Hypoplastic left heart syndrome Tricuspid valve abnormalities Conotruncal abnormalities Aortic coarctation Total anomalous pulmonary venous return Summary References

The antenatal detection of birth defects is an important public health concern with significant clinical ramifications. In 2002, more than 28,000 infants died within the first year of birth, with an overall rate of 7.0 deaths per 1000 live births in the United States [1]. This mortality rate was largely attributed to birth defects, of which congenital heart disease has been a leading cause of related adverse outcomes [2]. Furthermore, data from the World Health Organization indicates that 42% of infant deaths were attributed to heart defects [3]. Prenatal sonography has played an important role for the timely detection of congenital heart disease (CHD). Despite promising results of early studies, however, the efficacy of fetal cardiac screening programs has been variably successful. This article provides an update regarding various factors affecting the prenatal identification of cardiac de-

fects and summarizes current guidelines for fetal heart screening during the second trimester of pregnancy. The diagnostic imaging characteristics and clinical significance of selected fetal cardiac abnormalities are also reviewed.

Fetal cardiac screening


Cardiac abnormalities occur with an estimated incidence of approximately 413 per 1000 live births [46]. Most of the affected children will be born to mothers with no identifiable risk factors for CHD. Consequently, standardized approaches are needed to screen low-risk populations for cardiac abnormalities. Prenatal cardiac screening was introduced in the mid-1980s when the four-chamber view of the heart was incorporated into a routine obstetric scan

Division of Fetal Imaging, 3601 W. Thirteen Mile Road,William Beaumont Hospital, Royal Oak, MI 48073, USA Department of Obstetrics and Gynecology, Wayne State University, Detroit, MI, USA c Department of Obstetrics and Gynecology, University of Michigan, Ann Arbor, MI, USA * Corresponding author. Department of Obstetrics and Gynecology, Division of Fetal Imaging, William Beaumont Hospital, 3601 W. Thirteen Mile Road, Royal Oak, MI 48073. E-mail address: wlee@beaumont.edu (W. Lee).
b

1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.cult.2006.01.005

ultrasound.theclinics.com

274

Lee & Comstock

between 18 and 22 weeks, menstrual age [7]. After a training period, several centers in the United Kingdom were able to detect 77% of all cardiac anomalies over a 2.5-year period [8]. Copel and associates [9] also examined fetuses in 1022 pregnancies and found 74 abnormal heart cases using the four-chamber view. They reported 92% sensitivity and 99.7% specificity for the detection of CHD. Unfortunately, others have found a wide range of detection rates for the four-chamber view of the heart in unselected patient populations [1015]. Additional diagnostic benefit has been subsequently demonstrated for the inclusion of cardiac outflow tracts into routine screening examination of the heart [11,16,17]. One of the most comprehensive studies regarding fetal heart screening in a nonselected population has been recently completed. Tegnander and colleagues [18] analyzed results of a fetal heart screening program over a 10-year period in Norway. Their study population included 29,460 gravidas, representing 98% of their deliveries at a single institution. Routine fetal examinations were performed at approximately 18 weeks, and included the fourchamber view and outflow tracts. Beginning in 1995, patients were also asked to return for another scan if a satisfactory four-chamber view had not been obtained during the initial visit. Heart defects were retrospectively classified after delivery as critical when surgery was likely to be required (eg, transposition of the great arteries, hypoplastic left heart syndrome, atrioventricular septal defect, aortic coarctation, and large ventricular septal defects) [19]. This investigation identified 97 critical cases of CHD, of which 55 (57%) had been detected before birth. Forty-four percent of affected fetuses had isolated CHD and 38% had an abnormal karyotype. About one-half (48%) of the abnormal fetuses with ductal-dependent lesions were detected. Only 27% of infants born with critical CHD were alive at 2 years after delivery.

Factors influencing adequate examination of the heart


Inadequate examination of the fetal heart can be related to timing of scan, fetal position, image quality, maternal wall thickness, and a history of prior maternal surgical procedures.

Timing of second trimester fetal heart scans Satisfactory views of the fetal heart are typically obtained between 18 and 22 weeks, menstrual age. A prospective and randomized study of 1206 women found an incomplete four-chamber view more frequently between 18 to 18.9 weeks (18.4%) as compared with scans occurring between 20 to 20.9 weeks (2.9%) (P<.001) [21]. Many heart structures can still be satisfactorily visualized beyond this time if the fetal position is favorable. Many patients, however, prefer to know about major cardiac defects at an earlier stage of pregnancy. Fetal position Optimal views of the fetal heart are often obtained when the cardiac apex is pointing toward the transducer [Fig. 1] [22]. Suboptimal views occur when the spine causes acoustic shadowing over cardiac structures from a prone position. This situation can be worsened in the presence of oligohydramnios. Image quality Technical specifications of the ultrasound system (eg, beam former, transducer, display screen) can also affect satisfactory visualization of the fetal heart. One should always adjust the lowest acoustic power intensity settings (ie, thermal index and mechanical index) that provide satisfactory diagnostic images from using output display standards and the ALARA (As Low As Reasonably Achievable) principle [23]. Image quality relies on the examiners efforts to position the transducer in a manner that is most suitable for acquiring this information. The degree of screen magnification, gain, and acoustic focus should be optimized for the region of interest. The transducers field of view should be minimized to improve frame rate (ie, temporal resolution). A color filter can be applied to improve contrast between soft tissue borders. Frame persistence should not be set too high. Images should be zoomed to fill at least one-half the display screen with the fetal heart. Harmonic imaging often improves the image display [24]. Wavelength (sound velocity divided by frequency) is an important concept that is used to explain image resolution [25]. Contemporary ultrasound systems usually produce images with axial resolution between 2 to 4 wavelengths and lateral

What influences the prenatal detection of congenital heart disease?


Several factors affect the quality of a successful fetal cardiac screening program. These factors help to explain why prenatal detection rates have varied so widely in the medical literature. Chaoui [20] has summarized key reasons as to why the four-chamber view does not always provide an optimal detection rate among various centers. The reasons include inadequate examination, unrecognized abnormalities, evolution of cardiac lesions, and the inability for this view to detect specific anomalies that are best identified from other scanning planes.

Prenatal Diagnosis of Congenital Heart Disease

275

Fig. 1. Fetal position affects the cardiac screening examination. Satisfactory visualization of the fetal heart depends on orientation of cardiac structures in relation to the maternal abdomen. The best views are often obtained with the cardiac apex pointing toward the anterior maternal abdominal wall (AC ). Anatomicstructures are usually not well visualized when the cardiac apex is oriented toward the posterior maternal abdominal wall (DF ). (Adapted from Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998;17:6017; with permission.)

resolution between 3 to 10 wavelengths. Depending on the imaging system, the axial resolution would range between 0.6 and 1.2 mm at 5 MHz. At the same transducer frequency, lateral resolution ranges between 0.9 and 3.0 mm. Therefore, the ultrasound beam resolution and angle of beam insonation can have important ramifications for identifying small structures such as 2-mm ventricular septal defects. The highest transducer frequencies, typically 5 MHz and above, often provide the image resolution necessary to resolve subtle cardiac defects. This has to be balanced as a trade-off between image resolution and acoustic penetration. Maternal wall characteristics DeVore and colleagues [26] examined technical factors that influence imaging of the fetal heart during the second trimester of pregnancy. More than 700 trimester pregnancies were analyzed to identify independent risk factors that contribute to difficult cardiac screening examinations. Gestational age, maternal adipose tissue thickness, and prior lower abdominal surgery were found as the most significant factors that were associated with poor visualization of the fetal heart.

now indicates that the embryonic heart results from a modular process with initial development of a primary cardiac crescent. Subsequent development of a second, more anterior heart field is responsible for the appearance of the right ventricular outflow tract [27]. Cook and associates [28] have nicely reviewed cardiac development in the human fetus as it relates to the prenatal diagnosis of CHD. Education and training of health care professionals can improve the prenatal recognition of CHD. As a minimal goal, examiners must understand how to acquire images from standardized cardiac scanning planes to classify them into normal or abnormal categories. Cardiac screening programs can be further improved after continuous training of health care professionals based on feedback, a low threshold for echocardiography referrals, and convenient access to fetal heart specialists [29,30]. As one example, Hunter and colleagues [31] reported a twofold increase in the detection rate (17% to 36%) of major cardiac defects after implementing a targeted training program for fetal heart screening at 16 ultrasound centers in the United Kingdom.

Evolution of cardiac abnormalities Unrecognized abnormalities


Unrecognized abnormalities are another reason why the prenatal detection rates of CHD have varied so widely. Examiners should be familiar with development of the human heart and how to translate this information to clinical practice. Anatomic and molecular methods are clarifying new aspects of cardiac development. For example, traditional teaching has suggested that the heart initially forms from paired linear tubes that fuse and contain all major cardiac segments. More recent work Evolution of cardiac lesions is another important reason why these abnormalities are not always detected at the time of an ultrasound scan. An observational investigation of 22,050 pregnant women (77.5% low-risk patients) was undertaken by dividing them into two groups: Group A - 6924 with initial vaginal sonography at 1316 weeks gestation that were followed by abdominal scans at 2022 weeks; and Group B - 15,126 women who only had initial transabdominal scans at 2022 weeks [32]. Both groups were scanned dur-

276

Lee & Comstock

ing the third trimester and diagnoses were confirmed after birth. Two experienced examiners conducted all ultrasound examinations. Congenital heart disease occurred in 168 infants (Group A - 66 infants; Group B - 102 infants). Of the Group A fetuses, 42 (64%) heart anomalies were detected at the first vaginal scan, and 11 (17%) were subsequently identified during the abdominal study. Three additional anomalies (4%) were found during the third trimester exam and 10 more (15%) were only detected after delivery. Group B fetuses had 80 (78%) cardiac malformations identified at the time of their first abdominal scan at 2022 weeks gestation. An additional 7 (7%) and 15 (15%) cases were identified during the third trimester and after delivery, respectively. Ten heart anomalies that were discovered during the third trimester included aortic stenosis (n=2), cardiac rhabdomyoma (n=2), subaortic stenosis (n=1), tetralogy of Fallot (TOF) (n=1), aortic coarctation (n=1), sealed foramen ovale (n=1), ventricular septal defect (n=1), and hypertrophic cardiomyopathy (n=1). Their results indicated that fetal heart anomalies can vary in appearance throughout pregnancy. In this study, two experienced examiners, using both early vaginal and second-trimester abdominal scans, were unable to identify 20% of CHD cases. Of note, early vaginal scans of the fetal heart were able to detect nearly two-thirds (64%) of abnormal hearts. This observation is consistent with the finding of others who have described early detection of heart anomalies, especially when increased nuchal translucency is present [3339]. Although second trimester fetal heart screening can often be completed between 18 and 22 weeks gestation, many anomalies can still be identified before this stage of pregnancy.

Box 1: Common indications for fetal echocardiography Maternal indications Family history 1st degree relative of proband Pre-existing metabolic diseases diabetes phenylketonuria Maternal infections parvovirus B19 rubella coxackie Cardiac teratogen exposure retinoids phenyltoin carbamazepine lithium carbonate valproic acid Maternal antibodies anti-Ro (SSA) anti-La (SSB) Fetal indications Suspected fetal heart anomaly Abnormal fetal karyotype Major extra-cardiac anomaly Abnormal nuchal translucency 3.5 mm before 14 weeks, menstrual age Abnormal nuchal fold 6.0 mm: 1520 weeks, menstrual age Abnormal cardiac rate or rhythm persistent bradycardia persistent tachycardia persistent irregular heart rhythm

Standardized images may not demonstrate the anomalous heart


The inability of specific scanning planes for detecting all forms of CHD can be explained by considering fetal cardiac anatomy. Although the four-chamber view across the fetal thorax can be quite informative, this two-dimensional scanning plane may not provide satisfactory views of a small ventricular septal defect or conotruncal anomalies involving the more anterior ventricular outflow tracts.

technically feasible, an extended basic examination of the left and right ventricular outflow tracts is also recommended [4042]. Fetuses with suspected anomalies should be referred for fetal echocardiography to assess the seriousness of the anomaly and the likelihood of a ductal dependent lesion at birth. Common indications for fetal echocardiography are summarized in Box 1 [43].

Basic examination
General considerations The basic examination requires specific sonographic criteria using an adequately visualized four-chamber view of the heart [Table 1; Fig. 2]. This approach must not be mistaken for a simple count of cardiac chambers. Cardiac rate (120 to 160 beats/minute) and regular rhythm should be confirmed, although mild fetal bradycardia can transiently occur during the second trimester. The heart normally fills no more than a third of the thoracic area at the level of the four-chamber view. A small layer of fluid (<2 mm) can appear around the normal fetal heart, although this finding may

Fetal cardiac screening guidelines


The primary goal of cardiac screening is to identify which fetuses are likely to have CHD. Current guidelines emphasize a basic examination using a satisfactory four-chamber view of the heart. If

Prenatal Diagnosis of Congenital Heart Disease

277

Table 1: Basic cardiac screening examination General Normal cardiac situs, axis, and position Heart occupies a third of thoracic area Majority of heart in left chest Four cardiac chambers present No pericardial effusion or hypertrophy Atria approximately equal in size Foramen ovale flap in left atrium Atrial septum primum present Ventricles about equal in size No cardiac wall hypertrophy Moderator band at right ventricular apex Ventricular septum intact (apex to crux) Both atrioventricular valves open and move freely Tricuspid valve leaflet inserts on septum closer to the cardiac apex as compared to the mitral valve

bar emphysema), diaphragmatic hernia, or situs anomalies [4648]. Cardiac chambers Both atrial chambers should appear similar in size with an intact atrial septum primum. The foramen ovale flap should move freely toward the left atrium. In the human fetus, more than one-fifth of combined ventricular output is directed to the lungs and eventually drained back to the heart through the pulmonary veins [49,50]. At least one pulmonary vein should always be seen entering the left atrium. The ventricular chambers also appear similar in size with an intact intervening septum. Relatively thin ventricular walls usually have no greater than 2 mm of surrounding pericardial fluid. The right ventricle has a moderator band at the cardiac apex and normally resides in the anterior chest on the side opposite the fetal stomach. It usually appears more triangular in shape as compared with the left ventricular chamber. Atrioventricular valves Both atrioventricular valves should move freely and not appear thickened. The tricuspid valve leaflet normally inserts on the ventricular septum at a position that is closer to the cardiac apex as compared with the mitral valve septal leaflet insertion [51]. This describes the normal offsetting of both atrioventricular valves.

Atria

Ventricles

AV Valves

Adapted from [22]; with permission.

be clinically significant in the presence of cardiac failure, other structural anomalies, or hydrops [44,45]. Cardiac axis and position should be normal [Fig. 3] [46]. The cardiac axis can be shifted as a normal variant, although a detailed examination should be considered for the possibility of associated abnormalities. Abnormal heart deviation into the right chest (dextroposition) may be caused by a chest mass (eg, cystic adenomatoid malformation, pulmonary sequestration, or congenital lo-

Extended basic examination


If technically feasible, routine views of the outflow tracts should also be included as part of an ex-

Fig. 2. Four-chamber view of the heart. Key components of a normal four-chamber view include an intact interventricular septum and atrial septum primum. There is no disproportion between the left (LV) and right (RV) ventricles. A moderator band helps to identify the morphologic right ventricle. Note how the offset atrioventricular septal valve leaflets insert into the crux. (From Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998;17:6017; with permission.)

278

Lee & Comstock

Fig. 3. Fetal cardiac axis and position The cardiac axis can be measured from a four-chamber view of the fetal heart. A line through the interventricular axis is extended to the posterior border of the heart to produce point P, the location of which can be used to define fetal cardiac position. (Adapted from Comstock CH. Normal fetal heart axis and position. Obstet Gynecol 1987;70:2559; with permission.)

tended basic cardiac examination [Figs. 4 and 5]. A four-chamber view of the heart can appear as normal in the presence of a ventricular septal defect or conotruncal anomaly that would otherwise be seen from the extended basic examination. Scanning planes for both the basic and extended basic examinations are illustrated in Fig. 6 [22]. Both outflow tracts are examined as the transducer is angled from the four-chamber view toward the fetal head. Another method for evaluating the outflow tracts has also been described when the fetal interventricular cardiac septum is perpendicular to the ultrasound beam [52]. This approach begins with a four-chamber view of the heart and involves probe rotation until a left ventricular outflow tract is visualized. The probe can then be rocked cephalad until the pulmonary

arterial outflow tract is observed in a plane that appears perpendicular to the aortic outflow tract. Others have reported the use of a three-vessel view to describe relative sizes and relationships between the pulmonary artery, ascending aorta, and right superior vena cava [5355]. Most second trimester cardiac screening examinations will permit satisfactory visualization of the four-chamber view and outflow tracts. Over 18,000 second trimester fetuses underwent cardiac screening of the outflow tracts to evaluate the standardized practice of incorporating a basic fetal cardiac exam into a 30-minute scan [56]. When technically feasible, an extended basic evaluation of the outflow tracts was also attempted. Of the studies that included an adequate four-chamber view, the majority (93%) of scans had satisfactory

Fig. 4. Left ventricular outflow tract. A left ventricular outflow tract view (LVOT) emphasizes that a great vessel can be seen exiting the left ventricle. The aortic valve leaflets should be freely moving and not thickened. (From Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998;17:6017; with permission.)

Prenatal Diagnosis of Congenital Heart Disease

279

Fig. 5. Right ventricular outflow tract. Cardiac axis and position are identical to Fig. 4. A right ventricular outflow tract (RVOT) view emphasizes that a great vessel can be seen exiting the morphologic right ventricle (RV). The bifurcation of the pulmonary artery is not always seen in this scanning plane. Note that the RVOT exits the ventricle at about 90 to the aortic outflow tract. The pulmonary valve leaflets should be freely moving and not thickened. Sometimes, the right superior vena cava (SVC) can be seen. (Adapted from Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998;17:6017; with permission.)

views of the outflow tracts. Nonvisualization rates were: left ventricular outflow tract (4.2%); right ventricular outflow tract (1.6%); both outflow tracts (1.3%).

Ventricular septal defects


Ventricular septal defects are the most common type of CHD [Fig. 7]. They are caused by incomplete closure of the ventricular septum during fetal development and can occur at various locations [Fig. 8]. Although isolated lesions may not be clinically significant, their presence should raise the possibility of abnormal karyotype and associated structural findings. Paladini and colleagues [57] summarized their experience with isolated lesions that are detected using prenatal sonography. Of 26 fetuses that reached

Sonographic detection of selected cardiac anomalies


This section summarizes the clinical significance and sonographic findings associated with selected examples of CHD that may be detected by basic and extended basic cardiac screening exam.

Fig. 6. Scanning planes for the basic and extended basic cardiac examinations. A four-chamber view of the heart is obtained from an axial scanning plane across the fetal thorax. Corresponding views of the left (LVOT) and right (RVOT) ventricular outflow tracts are found by angling the transducer toward the fetal head. (Adapted from Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998;17:6017; with permission.)

280

Lee & Comstock

Fig. 7. Perimembranous ventricular septal defect. A ventricular septal defect (VSD) is circled on the heart specimen in an abortus at 20 weeks, menstrual age. This small perimembranous lesion can be compared with President Thomas Jefferson's nose and can be easily missed if technical factors do not favor satisfactory visualization.

1 year of age, 46.1% (12 cases) of all defects closed in utero. Trisomy 21 was found in 50% of fetuses with inlet-type (nine cases) or large (two cases) defects. Trisomy 18 occurred in 56.3% of ventricular septal defects (VSD) lesions (nine cases) that were associated with aortic-septal override. None of the malalignment VSDs closed after birth. By comparison, 69% of the perimembranous and 60% of the muscular defects closed within 1 year of delivery. In a related pediatric study, Turner

Fig. 8. Variable locations of ventricular septal defects. The location of a ventricular septal defect (VSD) can be used to estimate the likelihood of spontaneous closure and for counseling about therapeutic intervention after birth. An inlet VSD of the posterior septum, near either the tricuspid or mitral valves, will rarely close. Infants with this type of lesion are not optimal candidates for repair using a percutaneous septal occlusion device. (From Fetal ultrasound simulator [CDROM], Washington, D.C: American College of Obstetricians and Gynecologists; 1998; with permission.)

and colleagues [58] found that the location of the lesion was relevant to its natural history in 68 infants with VSD. Perimembranous defects accounted for most of the moderate and large defects that required surgical correction. After more than 6 years, almost a third of all perimembranous and just over two-thirds of all muscular defects closed spontaneously. Tegnander and colleagues [18] studied a nonselected population of 30,149 fetuses: an isolated VSD occurred in 57% (188/333) of noncritical heart defects. Of these, 162 (86%) were muscular and 26 (14%) were perimembranous. During the first year after birth, 77/162 (48%) of the isolated muscular and 3/26 (12%) of the isolated perimembranous VSDs closed spontaneously. Prenatal detection of a small VSD <2 mm can be particularly difficult and depends on several factors such as the lateral resolution of the ultrasound system, gestational age, fetal size, maternal wall thickness, and a history of prior abdominal or uterine surgeries. Aside from optimizing image settings (eg, depth, gain, focus), the entire ventricular septum should be systematically examined during both the basic and extended basic cardiac scans. Other diagnostic tools include the use of digital cineloop technology and Doppler flow studies. Direct measurements of intracardiac pressures in human fetuses indicate that no measurable differences occur between the left and right ventricles [59]. Therefore, color Doppler ultrasonography may not detect flow across a VSD as easily as in the case of adult patients. However, VSD lesions can be identified using frame-by-frame analysis of the ventricular septum (digital cine-loop), sometimes in conjunction with Doppler flow studies as well [Figs. 9 and 10]. Some ultrasound systems provide a write priority control that allows the

Prenatal Diagnosis of Congenital Heart Disease

281

Fig. 9. Intramuscular ventricular septal defect. This 2.6-mm VSD can be easily visualized using gray-scale sonography. Simultaneous use of digital cineloop and color Doppler techniques allow visualization of blood flow across this defect. This abnormality would be much easier to visualize after birth when a high-pressure gradient is present between the left and right ventricles.

user to balance the color or power Doppler display so it doesnt bleed over surrounding intra-cardiac structures that are visualized with gray scale. Finally, one should remember to verify suspected VSD lesions from more than one view to avoid diagnostic errors due to sonographic artifacts.

Atrioventricular septal defects


Complete atrioventricular septal defects (AVSD) consist of a common atrioventricular junction instead of separate mitral and tricuspid valve orifices [Fig. 11]. In the past, this lesion has also been known as an endocardial cushion defect or AV canal defect. Milder forms of this anomaly (ie, incomplete AVSD) have only a defect in the inferior part of the atrial septum, just above the atrioven-

tricular valves. In this case, two separate valve orifices will be visualized. Allan [60] reviewed the prenatal sonographic findings of 49 fetuses with AVSD; 18 cases appeared to only involve AVSD only, of which 13 were subsequently found to have trisomy 21. Other abnormalities included heterotaxy syndrome (22 cases), left ventricular malformations (8 cases), and TOF (1 case). Increased nuchal translucency has also been found in embryos with AVSD at 1014 weeks, menstrual age [61]. There were 22 cases of heterotaxy syndrome involving isomerism of the atrial appendages. Sixteen fetuses were suspected to have right atrial isomerism (eg, anomalies of pulmonary venous drainage, double outlet right ventricle with pulmonary valve

Fig. 10. Perimembranous ventricular septal defect. A small VSD may not be identified on gray-scale sonography (left), although the addition of a color Doppler study can be used demonstrate this lesion in the same fetus (right).

282

Lee & Comstock

Fig. 11. Atrioventricular septal defect. A complete form of atrioventricular septal defect is characterized by incomplete formation of separate mitral and tricuspid valves. A cross-section demonstrates a common valve apparatus (inset). RA, right atrium; LA, left atrium; RV, right ventricle; LV, left ventricle. (Adapted from Fetal ultrasound simulator [CD-ROM]. Washington, D.C: American College of Obstetricians and Gynecologists; 1998; with permission.)

atresia, bilateral right bronchi and lung lobation, and asplenia). Six fetuses had sonographic evidence of left atrial isomerism, where typical findings may include an interrupted inferior vena cava with azygous continuation, complete heart block, bilateral left bronchi and lung lobation, and polysplenia. This series emphasized the high rate of intrauterine death for fetuses with left atrial isomerism because only one of six abnormal fetuses survived to delivery, but died soon after birth. Although the prognosis for fetuses with AVSD and right atrial isomerism typically is to survive pregnancy, there is a high postnatal mortality rate found in infancy and early childhood [62]. The sonographic detection of complete AVSD is usually straightforward because the normal offsetting of the atrioventricular valves is not present. Instead, the common atrioventricular valve appears as a straight line [Fig. 12]. Occasionally, a dilated coronary sinus can simulate an AVSD lesion [63]. Incomplete forms of AVSD also may obscure the diagnosis because two separate inlet valves are still seen, in addition to both atrial and ventrticular septal defects. As a final consideration, ventricular disproportion may appear as an unbalanced form of AVSD. Complete AVSD can be typically repaired with low mortality and good intermediate to longterm results [64].

examination, is extremely important because of improved surgical outcome in fetuses in whom the lesion was detected prenatally [65] [Fig. 13]. Severe ventricular disproportion is the hallmark of this abnormality where the left ventricle can appear very small [Fig. 14]. This condition refers to an underdeveloped left ventricle from abnormal development of the mitral or aortic valve. Valve obstruction causes a shift of blood flow back over the foramen ovale to the right

Hypoplastic left heart syndrome


Prenatal identification of hypoplastic left heart syndrome (HLHS), using the basic cardiac screening

Fig. 12. Atrioventricular septal defect. Sonographic findings for a complete atrioventricular septal defect consist of atrial septal defect (asd), ventricular septal defect (vsd), and lack of the normal offset atrioventricular valve insertion sites. The common valve appears as a straight echogenic line.

Prenatal Diagnosis of Congenital Heart Disease

283

sion will keep these fetal shunts open temporarily, but eventually surgery will be necessary. The conventional surgical repair is a three-stage procedure, the first of which is a Norwood procedure that involves connecting the aorta to the proximal pulmonary artery, thus allowing the right ventricle to pump blood to the body [67]. Two lower risk operations, the hemi-Fontan (4 to 6 months of age) and Fontan (18 months to 2 years of age) are subsequently required [68]. Although these children are now reaching young adulthood and are doing remarkably well, recent studies suggest increased risk for cognitive, neuromotor, and psychosocial problems [69].
Fig. 13. Hypoplastic left heart syndrome Hypoplastic left heart syndrome consists of a small left ventricle (LV) with abnormal mitral or aortic valve development. The LV will initially appear slightly small during early pregnancy, with subsequent development as a rudimentary slit-like cavity as the pregnancy progresses. The dominant chamber is the right ventricle (RV). Blood flow across the foramen ovale is reversed from the right (RA) to left (LA) atrium. (Adapted from Fetal ultrasound simulator [CD-ROM]. Washington, D.C: American College of Obstetricians and Gynecologists; 1998; with permission.)

Tricuspid valve abnormalities


Tricuspid valve atresia is caused by marked dysplasia of the leaflets and cords where a connection fails to develop between the right atrium and ventricle. Therefore, systemic venous blood return bypasses the right heart and traverses a patent foramen ovale (secundum atrial septal defect, ASD) into the left atrium and ventricle. An inlet VSD typically allows some blood flow from the left ventricle into an underdeveloped right ventricle. Tongsong and colleagues [70] published prenatal sonographic findings for isolated tricuspid valve atresia. The four chamber view fails to demonstrate a patent tricuspid valve and this area appears echogenic [Fig. 15]. A small right ventricle is seen, de-

atrium. The left ventricle stays small due to decreased blood flow. A truly anatomic univentricular heart is very uncommon since a small slit can usuallly be seen to the left of the right ventricle. In severe mitral valve dysplasia, the left ventricle is usually quite small at the time of the screening exam. Milder involvement of the valves causes disparity of the size of the right and left ventricles, but both are still visible. Another presentation is that of an enlarged left atrium with a large noncontracting left ventricle. Although the ventricle is large, blood does not flow into or out of it. The enlarged ventricle can shrink with advancing pregnancy. Associated findings may include aortic valve atresia, coarctation of the aorta, and echogenic thickening of the ventricular walls due to endocardial fibroelastosis [66]. Hypoplastic left ventricle can also occur with other defects such as atrioventricular septal defect. It is rare that chromosomes are abnormal, but if they are, trisomy 18 is the most frequent aneuploidy. Hypoplastic left ventricle can also occur in left-sided diaphragmatic hernia due to pressure on the left ventricle. The pregnancy course is usually uneventful in the absence of heart failure (eg, atrioventricular valve insufficiency, pericardial effusion, cardiomegaly). After birth the newborn will depend on a patent foramen ovale and ductus arteriosus to get blood to the aorta and neck vessels. Prostaglandin infu-

Fig. 14. Hypoplastic left heart syndrome. Sonographic findings demonstrate marked underdevelopment of the left ventricular cavity that gives it a slitlike appearance. This degree of ventricular disproportion may not be as obvious at an earlier stage of pregnancy.

284

Lee & Comstock

poor perinatal prognosis with a mortality rate as high as 85% [73].

Conotruncal abnormalities
A primitive endocardial heart tube, superior to the primitive ventricles, becomes divided in half by a spiral septum to form great arteries early in gestation. This process may be interrupted for unknown reasons, leading to a spectrum of developmental cardiac anomalies such as truncus arteriosus, transposition of the great arteries, double-outlet right ventricle, and TOF. The fundamental lesion depends on where formation of the spiral septum is disrupted and where the great arteries are positioned in relation to each other at that time. Sonographic distinction between these specific abnormalities may be very difficult before birth [74,75]. Some conotruncal abnormalities, such as TOF and double outlet right ventricle, are also at increased risk for chromosomal abnormalities [76]. In a recent population-based study of 255,849 births, 43 children were found to have 22q11.2 deletion with an overall prevalence of 1 in 5950 births (95% CI, 1 in 4417 1 in 8224 births) [77]. Most affected children (81%) had a heart defect that included conotruncal anomalies (46%), interrupted aortic arch (19%), ventricular septal defects (16%), and other assorted extra-cardiac vascular anomalies (51%). Conotruncal aberrations are more frequent in diabetics and in those women who have a congenital abnormality themselves or have had a child with a genetic disorder. The basic cardiac examination, using a fourchamber view alone, is notoriously unreliable for detecting these anomalies. However, an extended basic examination of the outflow tracts is the key for their effective prenatal detection. The cardiac axis may provide an initial sonographic indication of a conotruncal anomaly during the basic examination [78,79]. DeVore [80] has also described the use of color Doppler sonography to visualize great vessel relationships for second and third trimester pregnancies. Truncus arteriosus Truncus arteriosus occurs when one great artery arises from the base of the heart and gives rise to the coronary, pulmonary, and aortic circulations. The truncus usually overrides a VSD. The pulmonary arteries arise from the truncal root as a common trunk (Type I, most frequent) [Fig. 16A, B], close but separate (Type II), or widely separate (Type III) [81]. Fortunately, they are not ductaldependent lesions. Fetal echocardiography should focus on the ventricular origin of the truncus, truncal valve annular

Fig. 15. Tricuspid valve atresia. Key findings consist of an echogenic mass of non-mobile tissue between the right ventricle (RV) and atrium (RA), small right ventricle (RV), and the presence of a ventricular septal defect (*). Increased systemic venous return flows from the RA through the foramen ovale (FO).

pending on size of the VSD. Right ventricular outflow tract obstruction may occur as subvalvular pulmonary stenosis or pulmonary valve stenosisa potential cause of ductal dependency. Approximately 20% of cases will be associated with transposition of the great arteries. Aortic arch abnormalities have also been reported. Chromosomes are usually normal and it is rare to have extracardiac defects. Increased nuchal translucency, however, has been reported in embryos with triscuspid atresia as early as 11 weeks, menstrual age [61]. The surgical outcome of infants born with this lesion depends on the presence of associated anomalies and relies on use of the Fontan procedure. The Hospital for Sick Children recently summarized a total of 137 infants who underwent a Fontan procedure [71]. Cohort survival was 90% at the age of 1 month, 81% at 1 year, 70% at 10 years, and 60% at 20 years. The tricuspid valve septal leaflet is normally not formed before 12 weeks, menstrual age [28]. However, Ebsteins malformation is another tricuspid valve anomaly that is caused by failure of the inferior atrioventricular cushion to properly develop this leaflet from the right side of the ventricular septum. This anomaly is characterized by increased offsetting of the atrioventricular valves and tricuspid valve insufficiency that leads to an enlarged right atrium. A nomogram of the mitral to tricuspid valve distance is helpful for confirming this diagnosis [51]. Melendres and colleagues [72] recently reported a missed case of Ebsteins anomaly that was obscured by misinterpretation of an atrioventricular groove. Ebsteins anomaly has a

Prenatal Diagnosis of Congenital Heart Disease

285

Fig. 16. Truncus arteriosus. (A) Axial view of the fetal thorax reveals a small pulmonary artery (arrow) that directly originates from a large truncal root (Tr). The fetal spine (Sp) is seen as a point of reference. (B) Oblique coronal view of the left hemithorax demonstrates a large truncal vessel, with ventricular septal defect, that exits the right (RV) and left (LV) ventricles. This truncal vessel overrides the ventricular septum and gives off a small pulmonary vessel (arrow).

diameter, and evidence for valvular dysplasia. Doppler studies may reveal truncal valve stenosis or insufficiency [82]. In a pathology study of 28 Type I and Type II cases, other cardiovascular lesions included anomalous position of the left coronary artery (18.5%), right aortic arch (36%), and interrupted aortic arch (11%) [83]. The distinction between truncus arteriosus and pulmonary valve atresia with VSD can be very difficult [84]. An investigation from the University of Michigan described 46 infants with truncus arteriosus undergoing early primary repair and found an actuarial survival rate of 81 6% at 90 days and beyond [85]. More recently, an observational study of 23 affected fetuses indicates the following outcomes: termination of pregnancies (34.8%); intrauterine death (8.7%); postnatal deaths (21.7%) [84]. The eight remaining neonates (34.8%) were alive and doing well after surgery (n=6) or awaiting repair (n=2). Major cardiac surgical centers currently favor a primary repair of truncus arteriosus during the neonatal period. Transposition of the great arteries Transposition of the great arteries (TGA) occurs when the aorta arises from the right ventricle and the pulmonary artery arises from the left ventricle. The usual spiral relationship of the great arteries is lost so that the outflow tracts are parallel to each other [Fig. 17] [86]. In about half of cases there is no VSD. Affected newborns without a large VSD may not have adequate mixing of oxygenated blood and can experience rapid hemodynamic decompensation as the ductus arteriosus closes. This abnormality can be easily missed from the four-chamber view unless a very large VSD is pre-

sent. In this case, an extended basic examination of the outflow tracts is likely to identify TGA. Congenitally corrected transposition of the great arteries can also rarely occur where parallel vessels are also seen exiting the heart. In this situation, the atria connect with anatomically discordant ventricles and the ventricles connect with discordant and transposed great arteries [87]. Careful attention to the chamber morphology, presence of moderator band, and papillary muscle relationships can pro-

Fig. 17. Transposition of the great arteries. The maindiagnostic findings is the manner by which the great arteries exit the ventricles in a parallel manner. Fetal echocardiography would demonstrate that the aortic root originates from the right ventricle (RV), whereas the pulmonary artery would exit the left ventricle (LV). A small ventricular septal defect is also present (arrow).

286

Lee & Comstock

vide important clues to help the examiner accurately distinguish between right and left ventricles. The potential impact of early prenatal diagnosis is illustrated by a study that examined clinical outcome in 68 affected neonates in whom a prenatal diagnosis of TGA was suspected [88]. Results were compared with 250 affected neonates who were first identified with TGA after delivery. Newborns with late diagnosis had increased delay between birth and admission for special care and were more likely to present with metabolic acidosis and multi-organ failure. Preoperative mortality was 6% for the neonatal group as compared with no deaths in the prenatal group. Postoperative morbidity was not different between groups, although the hospital stay was slightly longer in babies diagnosed after birth. Postoperative death was significantly higher in the neonatal group (20 of 235 versus 0 of 68, P<.01. Their results suggest that prenatal diagnosis of TGA reduces mortality and morbidity. Timely detection of TGA before delivery provides adequate planning for monitoring and antepartum care. This information also facilitates in utero transfer of fetuses to a tertiary care facility that can appropriately treat newborns with ductal dependent lesions. Double-outlet right ventricle Double-outlet right ventricle (DORV) describes a condition in which most of the aorta and pulmonary artery arise from the right ventricle [89]. The sonographic findings may closely resemble TOF or transposition of the great arteries with VSD [Fig. 18]. The outcome of infants with DORV depends largely on the associated anomalies.

Fig. 19. Tetralogy of Fallot. The initial diagnostic clue for tetralogy of Fallot is the presence of an overriding aorta (ao) over a ventricular septal defect (*) during the extended basic examination of the heart. Sonographic evidence for a small pulmonary artery (not seen in this image) may not occur until a much later time in pregnancy. s, ventricular septum; sp, spine.

As in TOF there is always an accompanying VSD but, unlike TOF the relationship of the great vessels is often abnormal; the aorta is commonly transposed anterior and to the right of the pulmonary artery. Other defects are not infrequent such as pulmonary valve stenosis, right cardiac axis deviation, right aortic arch, atrial septal defect, and total anomalous pulmonary venous return. DORV with transposition is known as a Taussig-Bing defect and is commonly associated with coarctation. The combination of DORV and AVSD is difficult to repair surgically. Tetralogy of Fallot TOF is the only conotruncal anomaly in which the usual spiral relationship of the great vessels is maintained. The aorta overrides a ventricular sepal defect [Fig. 19]. The pulmonary artery may be small due to uneven division of the primitive truncus. In the fetus, unlike in children, the right ventricle is not hypertrophied because of shunting across the foramen ovale and the VSD reduces the load on the right heart. Early fetal TOF may simply present as a VSD with aortic septal override. This anomaly can be missed on a screening four-chamber view if the VSD is small [Fig. 20]. Left cardiac axis deviation may provide an initial diagnostic clue for this lesion [79]. However, the extended basic cardiac examination is most likely to demonstrate a VSD with aortic override. The aortic root itself can also be enlarged, although pulmonary valve stenosis may be become

Fig. 18. Double-outlet right ventricle. Both great vessels exit the right ventricle (RV). The most lateral one represents the aortic arch (Ao) because neck vessels are present (arrows). No arteries are seen exiting the left ventricle (LV). Sp = spine.

Prenatal Diagnosis of Congenital Heart Disease

287

These newborns may experience hemodynamic decompensation if the pulmonary valve diameter is small ( 5 mm at term) or if there is retrograde flow across the ductus arteriosus. Immediate surgery may not be required for affected infants with sufficient pulmonary valve flow and absence of ductal dependency. One variation, in which the pulmonary valve is atretic and the pulmonary artery is not visible, is known as pseudotruncus as only a solitary large vessel is seen straddling the VSD. Another variation of TOF involves absence of the pulmonary valve so that there is regurgitation back and forth. The pulmonary artery may enlarge to massive proportions and hydrops may occur.
Fig. 20. Tetralogy of Fallot with normal four-chamber view The same fetus with tetralogy of Fallot appears in Fig. 19. This four-chamber view appears normal despite the presence of this conotruncal anomaly. lv, left ventricle; rv, right ventricle; ra, right atrium; la, left atrium.

Aortic coarctation
Cardiac ventricular disproportion can occur as a normal variant, a consequence of fetal growth restriction, or an indirect sign of aortic coarctation [Fig. 21]. Aortic coarctation is usually difficult to directly visualize, especially in the context of a cardiac screening examination. However, the prenatal diagnosis of aortic coarctation has been reported to improve survival and reduce morbidity [94]. Allan and colleagues [95] found that 24 fetuses had dilatation of the right heart as compared with the left side. In 18 of these cases, the diagnosis of aortic coarctation or interruption was correctly inferred from indirect sonographic findings. Hornberger and associates [96] summarized a multicenter investigation for 20 infants with coarctation. They found quantitative hypoplasia of the aortic isthmus and transverse arch as the best predictors

apparent until later pregnancy [90,91]. The aortic to pulmonary size ratio will be high despite normal valve diameter measurements. In TOF, this ratio increases as gestational age advances because there is less than the usual growth of the pulmonary diameter [92]. The 90% confidence interval for the pulmonary artery to aortic diameter (Pa/Ao) ratio ranges from 0.84 to 1.41 and remains constant throughout pregnancy [93]. The right ventricular outflow tract should be re-evaluated before delivery to identify affected fetuses at greatest risk for ductal dependency.

Fig. 21. Ventricular disproportion A disparity in ventricular dimensions can represent a normal variant. However, one should also consider that this could be caused by several other conditions that include intrauterine growth restriction, aortic coarctation, or an evolving hypoplastic left heart syndome.

288

Lee & Comstock

for coarctation and emphasized the importance of conducting serial studies for suspected cases. Ventricular disproportion has only a moderate degree of sensitivity (62%) for the detection of coarctation that occurs with a high false-positive rate after 34 weeks [97]. The 90% confidence interval for the right ventricular to left ventricular diameter ratio is constant throughout pregnancy and ranges from 0.79 to 1.24 [93].

Total anomalous pulmonary venous return


The left atrium is normally positioned near the descending aorta. In the space between these two structures are the main left and right pulmonary veins. They can be followed to their drainage point in the back of the left atrium. In total anomalous pulmonary venous return (TAPVR), both main pulmonary veins drain into the right atrium via a vertical vein which then drains into a coronary sinus, persistent left superior vena cava, innominate vein, the hepatic veins, or even below the diaphragm (20%) into the inferior vena cava. Thus oxygenated blood from the lungs never reaches the body and brain but rather circulates around and around in the lungs with the only mixing available across a patent foramen ovale. TAPVR can lead to cardiorespiratory decompensation of the newborn that is unresponsive to prostaglandin infusion. This circulatory abnormality should be considered when a dilated coronary sinus, cardiac chamber disproportion, or size dis-

parity between the great arteries is seen. It is particularly common in fetuses with heterotaxic abnormalities. Direct documentation of pulmonary venous flow into the left atrium is the most reliable way to exclude TAPVR [98] [Fig. 22]. A nonlethal variation is partial anomalous venous return in which the right vein drains to the right atrium but the left still drains to the left atrium. As long as some return goes to the left atrium, oxygenated blood is available to the body and brain.

Summary
CHD is a leading cause of infant morbidity and mortality that results from birth defects. Diagnosticians who use ultrasonography to evaluate the fetal heart must be familiar with key factors that can impact the success of their cardiac screening programs. A good understanding of practice guidelines for the basic and extended basic cardiac examination is essential. Efforts should also be made to standardize diagnostic training of these who perform these examinations in an ongoing manner. The primary goal is for these individuals to accurately identify who should be referred for a more detailed evaluation of the fetus.

References
[1] Kochanek KD, Murphy SL, Anderson RN, Scott C. Deaths: final data for 2002. Natl Vital Stat Rep 2004;53:1115. [2] Centers for Disease Control and Prevention. Trends in infant mortality attributable to birth defects - United States, 19801995. MMWR Morb Mortal Wkly Rep 2005;47:7738. [3] Rosano A, Botto LD, Botting B, Mastroiacovo P. Infant mortality and congenital anomalies from 1950 to 1994: an international perspective. J Epidemiol Community Health 2000;54:6606. [4] Ferencz C, Rubin JD, McCarter RJ, et al. Congenital heart disease: prevalence at livebirth. The Baltimore-Washington infant study. Am J Epidemiol 1985;121:316. [5] Meberg A, Otterstad JE, Froland G, et al. Outcome of congenital heart defectsa populationbased study. Acta Paediatr 2000;89:134451. [6] Cuneo BF, Curran LF, Davis N, Elrad H. Trends in prenatal diagnosis of critical cardiac defects in an integrated obstetric and pediatric cardiac imaging center. J Perinatol 2004;24:6748. [7] Allan LD, Crawford DC, Chita SK, Tynan MJ. Prenatal screening for congenital heart disease. BMJ 1986;292:7179. [8] Sharland GK, Allan LD. Screening for congenital heart disease prenatally. Results of a 2 1/2 year study in the South East Thames Region. Br J Obstet Gynaecol 1992;99:2205.

Fig. 22. Total anomalous pulmonary venous return At least two pulmonary veins (arrows) drain into the right atrium instead of the left. A ventricular septal defect is also present between the two markers. LV, left ventricle. Increased volume loading of the right heart and cyanosis are the key clinical findings for this condition. An atrial septal defect is desirable because this defect will allow blood to flow from the right atrium into the left heart.

Prenatal Diagnosis of Congenital Heart Disease

289

[9] Copel JA, Pilu G, Green J, Hobbins JC, Kleinman CS. Fetal echocardiographic screening for congenital heart disease: the importance of the fourchamber view. Am J Obstet Gynecol 1987;157: 64855. [10] Crane JP, LeFevre ML, Winborn RC, et al. A randomized trial of prenatal ultrasonographic screening: impact on the detection, management, and outcome of anomalous fetuses. The RADIUS Study Group. Am J Obstet Gynecol 1994;171: 3929. [11] Kirk JS, Riggs TW, Comstock CH, et al. Prenatal screening for cardiac anomalies: the value of routine addition of the aortic root to the fourchamber view. Obstet Gynecol 1994;84:42731. [12] Ott WJ. The accuracy of antenatal fetal echocardiography screening in high and low risk patients. Am J Obstet Gynecol 1995;172:17417. [13] Tegnander E, Eik-Nes SH, Johansen OJ, Linker DT. Prenatal detection of heart defects at the routine fetal examination at 18 weeks in a nonselected population. Ultrasound Obstet Gynecol 1995;5:37280. [14] Stumpflen I, Stumpflen A, Wimmer M, Bernaschek G. Effect of detailed fetal echocardiography as part of routine prenatal ultrasound screening on detection of congenital heart disease. Lancet 1996;348:8547. [15] Buskens E, Grobbee DE, Frohn-Mulder IME. Efficacy of routine ultrasound screening for congenital heart disease in normal pregnancy. Circulation 1996;94:6772. [16] Bromley B, Estroff JA, Sanders SP, et al. Fetal echocardiography: accuracy and limitations in a population at high and low risk for heart defects. Am J Obstet Gynecol 1992;166:147381. [17] Wigton TR, Sabbagha RE, Tamura RK, et al. Sonographic diagnosis of congenital heart disease: comparison between the four-chamber view and multiple cardiac views. Obstet Gynecol 1993;82:21924. [18] Tegnander E, Williams W, Johansen OJ, et al. Prenatal detection of heart defects in a nonselected population of 30,149 fetusesdetection rates and outcome. Ultrasound Obstet Gynecol 2006;27:25265. [19] Mitchell SC, Korones SB, Berendes HW. Congenital heart disease in 56,109 births. Incidence and natural history. Circulation 1971;43:32332. [20] Chaoui R. The four-chamber view: four reasons why it seems to fail in screening for cardiac abnormalities and suggestions to improve detection rate. Ultrasound Obstet Gynecol 2003;22:310. [21] Senat MV, Holden D, Bernard JP, et al. Feasibility of the second-trimester fetal ultrasound examination in an unselected population at 18, 20 or 22 weeks of pregnancy: a randomized trial. Ultrasound Obstet Gynecol 1999;14:927. [22] Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med 1998; 17:6017.

[23] Barnett SB, Ter Haar GR, Ziskin MC, et al. International recommendations and guidelines for the safe use of diagnostic ultrasound in medicine. Ultrasound Med Biol 2000;26:35566. [24] Paladini D, Vassallo M, Tartaglione A, et al. The role of tissue harmonic imaging in fetal echocardiography. Ultrasound Obstet Gynecol 2004; 23:15964. [25] Kossof G. Basic physics and imaging characteristics of ultrasound. World J Surg 2000;24: 13442. [26] DeVore GR, Medearis AL, Bear MB, et al. Fetal echocardiography: factors that influence imaging of the fetal heart during the second trimester of pregnancy. J Ultrasound Med 1993;12:65963. [27] Moorman A, Webb S, Brown NA, et al. Development of the heart: (1) formation of the cardiac chambers and arterial trunks. Heart 2003;89: 80614. [28] Cook AC, Yates RW, Anderson RH. Normal and abnormal fetal cardiac anatomy. Prenat Diagn 2004;24:103248. [29] Carvalho JS, Mavrides E, Shinebourne EA, et al. Improving the effectiveness of routine prenatal screening for major congenital heart defects. Heart 2002;88:38791. [30] Cuneo BF, Curran LF, Davis N, Elrad H. Trends in prenatal diagnosis of critical cardiac defects in an integrated obstetric and pediatric cardiac imaging center. J Perinatol 2004;24:6748. [31] Hunter S, Heads A, Wyllie J, Robson S. Prenatal diagnosis of congenital heart disease in the northern region of England: benefits of a training programme for obstetric ultrasonographers. Heart 2000;84:2948. [32] Yagel S, Weissman A, Rotstein Z, et al. Congenital heart defects: natural course and in utero development. Circulation 1997;96:5505. [33] Achiron R, Rotstein Z, Lipitz S, et al. Firsttrimester diagnosis of fetal congenital heart disease by transvaginal ultrasonography. Obstet Gynecol 1994;84:6972. [34] Hyett J, Moscoso G, Papapanagiotou G, et al. Abnormalities of the heart and great arteries in chromosomally normal fetuses with increased nuchal translucency thickness at 1113 weeks of gestation. Ultrasound Obstet Gynecol 1996;7: 24550. [35] Hyett JA, Perdu M, Sharland GK, et al. Increased nuchal translucency at 1014 weeks of gestation as a marker for major cardiac defects. Ultrasound Obstet Gynecol 1997;10:2426. [36] Rustico MA, Benettoni A, DOttavio G, et al. Early screening for fetal cardiac anomalies by transvaginal echocardiography in an unselected population: the role of operator experience. Ultrasound Obstet Gynecol 2000;16:6149. [37] Huggon IC, Ghi T, Cook AC, et al. Fetal cardiac abnormalities identified prior to 14 weeks gestation. Ultrasound Obstet Gynecol 2002;20:229. [38] Carvalho JS. Fetal heart scanning in the first trimester. Prenat Diagn 2004;24:10607.

290

Lee & Comstock

[39] Carvalho JS, Moscoso G, Tekay A, et al. Clinical impact of first and second trimester fetal echocardiogaphy on high risk pregnancies. Heart 2004;90:9216. [40] American Institute of Ultrasound in Medicine. Guidelines for the performance of the antepartum obstetrical ultrasound examination. J Ultrasound Med 2003;22:111625. [41] American College of Radiology. ACR practice guideline for the performance of antepartum obstetrical ultrasound. In: Practice guidelines & technical standards. Reston, VA: ACR; 2004. p. 68995. [42] American College of Obstetricians and Gynecologists. ACOG Practice Bulletin. Ultrasonography in pregnancy. Obstet Gynecol 2004;104:144958. [43] Small M, Copel JA. Indications for fetal echocardiography. Pediatr Cardiol 2004;25:20122. [44] Di Salvo DN, Brown DL, Doubilet PM, et al. Clinical significance of isolated fetal pericardial effusion. J Ultrasound Med 1994;13:2913. [45] Yoo SJ, Min JY, Lee YH. Normal pericardial fluid in the fetus: color and spectral Doppler analysis. Ultrasound Obstet Gynecol 2001;18:24852. [46] Comstock CH. Normal fetal heart axis and position. Obstet Gynecol 1987;70:2559. [47] Smith RS, Comstock CH, Kirk JS, Lee W. Ultrasonographic left cardiac axis deviation: a marker for fetal anomalies. Obstet Gynecol 1995;85: 18791. [48] Comstock CH, Smith R, Lee W, Kirk JS. Right fetal cardiac axis: clinical significance and associated findings. Obstet Gynecol 1998;91:4959. [49] Sutton MS, Groves A, MacNeill A, et al. Assessment of changes in blood flow through the lungs and foramen ovale in the normal human fetus with gestational age: a prospective Doppler echocardiographic study. Br Heart J 1994;71: 2327. [50] Rasanen J, Wood DC, Weiner S, et al. Role of the pulmonary circulation in the distribution of human fetal cardiac output during the second half of pregnancy. Circulation 1996;94:106873. [51] Vettraino IM, Huang R, Comstock CH. The normal offset of the tricuspid septal leaflet in the fetus. J Ultrasound Med 2002;21:1099104. [52] DeVore G. The aortic and pulmonary outflow tract screening examination in the human fetus. J Ultrasound Med 1992;11:3458. [53] Yoo SJ, Lee YH, Cho KS. Abnormal three-vessel view on sonography: a clue to the diagnosis of congenital heart disease in the fetus. AJR Am J Roentgenol 1999;172:82530. [54] Vinals F, Heredia F, Giuliano A. The role of the three vessels and trachea view (3VT) in the diagnosis of congenital heart defects. Ultrasound Obstet Gynecol 2003;22:35867. [55] Yagel S, Arbel R, Anteby EY, et al. The three vessels and trachea view (3VT) in fetal cardiac scanning. Ultrasound Obstet Gynecol 2002;20: 3405. [56] Vettraino IM, Lee W, Bronsteen RA, Comstock

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

CH. Sonographic evaluation of the ventricular cardiac outflow tracts [Letter to the Editor]. J Ultrasound Med 2005;24:566. Paladini D, Palmieri S, Lamberti A, et al. Characterization and natural history of ventricular septal defects in the fetus. Ultrasound Obstet Gynecol 2000;16:11822. Turner SW, Hunter S, Wyllie JP. The natural history of ventricular septal defects. Arch Dis Child 1999;81:4136. Johnson P, Maxwell DJ, Tynan MJ, Allan LD. Intracardiac pressures in the human fetus. Heart 2000;84:5963. Allan LD. Atrioventricular septal defect in the fetus. Am J Obstet Gynecol 1999;181(5 Pt 1): 12503. Galindo A, Comas C, Martinez JM, et al. Cardiac defects in chromosomally normal fetuses with increased nuchal translucency at 1014 weeks of gestation. J Matern Fetal Neonatal Med 2003;13: 16370. Gentles TL, Mayer Jr JE, Gauvreau K, et al. Fontan operation in five hundred consecutive patients: factors influencing early and late outcome. J Thorac Cardiovasc Surg 1997;114:37691. Park JK, Taylor DK, Skeels M, Towner DR. Dilated coronary sinus in the fetus: misinterpretation as an atrioventricular canal defect. Ultrasound Obstet Gynecol 1997;10:1269. Bando K, Turrentine MW, Sun K, et al. Surgical management of complete atrioventricular septal defects. A twenty-year experience. J Thorac Cardiovasc Surg 1995;110:154352. Tworetzky W, McElhinney DB, Reddy VM, et al. Improved surgical outcome after fetal diagnosis of hypoplastic left heart syndrome. Circulation 2001;103:126973. McCaffrey FM, Sherman FS. Prenatal diagnosis of severe aortic stenosis. Pediatr Cardiol 1997;18: 27681. Norwood WI, Kirkland JK, Sanders SP. Hypoplastic left heart syndrome: experience with palliative surgery. Am J Cardiol 1980;45:8791. Bove EL, Ohye RG, Devaney EJ. Hypoplastic left heart syndrome: conventional surgical management. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2004;7:310. Mahle WT, Wernovsky G. Neurodevelopmental outcomes in hypoplastic left heart syndrome. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2004;7:3947. Tongsong T, Sittiwangkul R, Wanapirak C, Chanprapaph P. Prenatal diagnosis of isolated tricuspid valve atresia: report of 4 cases and review of the literature. J Ultrasound Med 2004; 23:94550. Sittiwangkul R, Azakie A, Van Arsdell GS, et al. Outcomes of tricuspid atresia in the Fontan era. Ann Thorac Surg 2004;77:88994. Melendres G, Ormsby EL, McGahan JP, et al. Prenatal diagnosis of Ebstein anomaly: a potential pitfall. J Ultrasound 2004;23:5515.

Prenatal Diagnosis of Congenital Heart Disease

291

[73] Pavlova M, Fouron JC, Drblik SP, et al. Factors affecting the prognosis of Ebsteins anomaly during fetal life. Am Heart J 1998;135:10815. [74] Allan LD, Sharland GK, Milburn A, et al. Prospective diagnosis of 1,006 consecutive cases of congenital heart disease in the fetus. J Am Coll Cardiol 1994;23:14528. [75] Tometzki AJ, Suda K, Kohl T, et al. Accuracy of prenatal echocardiographic diagnosis and prognosis of fetuses with conotruncal anomalies. J Am Coll Cardiol 1999;33:1696701. [76] Paladini D, Rustico M, Todros T, et al. Conotruncal anomalies in prenatal life. Ultrasound Obstet Gynecol 1996;8:2416. [77] Botto LD, May K, Fernhoff PM, et al. A population-based study of the 22q11.2 deletion: phenotype, incidence, and contribution to major birth defects in the population. Pediatrics 2003; 112:1017. [78] Comstock CH, Smith RS, Lee W, Kirk JS. Right fetal cardiac axis: clinical significance and associated findings. Obstet Gynecol 1998;85:4959. [79] Smith RS, Comstock CH, Kirk JS, Lee W. Ultrasonographic left cardiac axis deviation: a marker for fetal anomalies. Obstet Gynecol 1995;85: 18791. [80] DeVore GR. Color Doppler examination of the outflow tracts of the fetal heart: a technique for identification of cardiovascular malformations. Ultrasound Obstet Gynecol 1994;4:46371. [81] Collett RW, Edwards JE. Persistent truncus arteriosus: a classification according to anatomic types. Surg Clin N Am 1949;29:124570. [82] Duke C, Sharland GK, Jones AM, Simpson JM. Echocardiographic features and outcome of truncus arteriosus diagnosed during fetal life. Am J Cardiol 2001;88:137984. [83] Butto F, Lucas Jr RV, Edwards JE. Persistent truncus arteriosus: pathologic anatomy in 54 cases. Pediatr Cardiol 1986;7:95101. [84] Volpe P, Paladini D, Marasini M, et al. Common arterial trunk in the fetus: characteristics, associations, and outcome in a multicentre series of 23 cases. Heart 2003;89:143741. [85] Bove EL, Lupinetti FM, Pridjian AK, et al. Results of a policy of primary repair of truncus arteriosus in the neonate. J Thorac Cardiovasc Surg 1993; 105:105765. [86] Allan LD. Sonographic detection of parallel great

[87]

[88]

[89]

[90]

[91]

[92]

[93]

[94]

[95]

[96]

[97]

[98]

arteries in the fetus. AJR Am J Roentgenol 1997; 168:12836. McEwing RL, Chaoui R. Congenitally corrected transposition of the great arteries: clues for prenatal diagnosis. Ultrasound Obstet Gynecol 2004; 23:6872. Bonnet D, Solti A, Butera G, et al. Detection of transposition of the great arteries in fetuses reduces neonatal morbidity and mortality. Circulation 1999;99:9168. Smith RS, Comstock CH, Kirk JS, et al. Doubleoutlet right ventricle: an antenatal diagnostic dilemma. Ultrasound Obstet Gynecol 1999;14: 3159. DeVore GR, Siassi B, Platt LD. Fetal echocardiography. VIII. Aortic root dilatationa marker for tetralogy of Fallot. Am J Obstet Gynecol 1988; 159:12936. Lee W, Smith RS, Comstock CH, et al. Tetralogy of Fallot: prenatal diagnosis and postnatal survival. Obstet Gynecol 1995;86:5838. Hornberger LK, Sanders SP, Sahn DJ, et al. In utero pulmonary artery and aortic growth and potential for progression of pulmonary outflow tract obstruction in tetralogy of Fallot. J Am Coll Cardiol 1995;25:73945. Kirk JS, Comstock CH, Lee W, et al. Fetal cardiac asymmetry: a marker for congenital heart disease. Obstet Gynecol 1999;93:18992. Franklin O, Burch M, Manning N, et al. Prenatal diagnosis of coarctation of the aorta improves survival and reduces morbidity. Heart 2002;87: 679. Allan LD, Chita SK, Anderson RH, et al. Coarctation of the aorta in prenatal life: an echocardiographic, anatomical, and functional study. Br Heart J 1988;59:35660. Hornberger LK, Sahn DJ, Kleinman CS, et al. Antenatal diagnosis of coarctation of the aorta: a multicenter experience. J Am Coll Cardiol 1994;23:41723. Brown DL, Durfee SM, Hornberger LK. Ventricular discrepancy as a sonographic sign of coarctation of the fetal aorta: how reliable is it? J Ultrasound Med 1997;16:959. Allan LD, Sharland GK. The echocardiographic diagnosis of totally anomalous pulmonary venous connection in the fetus. Heart 2001;85: 4337.

293

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 293301

Doppler Ultrasound in Obstetrics


Alfred Abuhamad,
&

MD

Fetal arterial Doppler Umbilical arterial circulation Middle cerebral circulation Fetal venous Doppler

& & &

Fetal cardiac Doppler Fetal Doppler and intrauterine growth restriction References

The Doppler effect, which was first reported by Christian Doppler in 1842 [1,2], describes the apparent variation in frequency of a light or a sound wave as the source of the wave approaches or moves away, relative to an observer. The traditional example that is given to describe this physical phenomenon is the apparent change in sound level of a train as the train approaches and then departs a station. The sound seems higher in pitch as the train approaches the station and seems lower in pitch as the train departs the station. This apparent change in sound pitch, or what is termed the frequency shift, is proportional to the speed of movement of the sound-emitting source. In clinical applications, when ultrasound with a certain frequency (fo) is used to insonate a certain blood vessel, the reflected frequency (fd) or frequency shift is directly proportional to the speed with which the red blood cells are moving (blood flow velocity) within that particular vessel. This frequency shift of the returning signal is displayed in a graphic form as a time-dependent plot. In this display, the vertical axis represents the frequency shift and the horizontal axis represents the temporal change of this frequency shift as it relays to the events of the cardiac cycle [Fig. 1]. This frequency shift is highest during systole, when the blood flow is at its fastest, and lowest during enddiastole, when the blood flow is at its slowest in the peripheral circulation. Given that the velocity of flow in a particular vascular bed is inversely pro-

portional to the downstream impedance to flow, the frequency shift therefore derives information on the downstream impedance to flow of the vascular bed under study. The frequency shift is also dependent on the cosine of the angle that the ultrasound beam makes with the targeted blood vessel [see formula in Fig. 1]. Given that the insonating angle is difficult to measure in clinical practice, indices that rely on ratios of frequency shifts were developed to quantitate Doppler waveforms. By relying on ratios of frequency shifts, these Doppler indices are thus independent of the effects of the insonating angle of the ultrasound beam. Doppler indices that are commonly used in obstetrical practice are shown in Fig. 2.

Fetal arterial Doppler Umbilical arterial circulation


The umbilical arterial circulation is normally a lowimpedance circulation [Fig. 3] , with an increase in the amount of end diastolic flow with advancing gestation [3,4]. Umbilical arterial Doppler waveforms reflect the status of the placental circulation, and the increase in end diastolic flow that is seen with advancing gestation is a direct result of an increase in the number of tertiary stem villi that takes place with placental maturation [5,6]. Diseases that obliterate small muscular arteries in placental tertiary stem villi result in a progressive decrease in

Department of Obstetrics and Gynecology, Eastern Virginia Medical School, Hofheimer Hall, Suite 310, 8254 Fairfax Avenue, Norfolk, VA 23507, USA E-mail address: abuhamaz@evms.edu
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2005.12.001

ultrasound.theclinics.com

294

Abuhamad

Fig. 1. The Doppler effect (fd) is dependent on the velocity of flow (V ) of the blood within a vessel, the initial frequency of the ultrasound beam (fc), and the cosine of the angle (A) that the ultrasound beam makes with the direction of flow. The Doppler effect is displayed on the monitor as a time-dependent plot of the frequency shift (fd) within a cardiac cycle.

end-diastolic flow in the umbilical arterial Doppler waveforms until absent, and then reverse flow during diastole is noted [Fig. 4] [7,8]. Reversed diastolic flow in the umbilical arterial circulation represents an advanced stage of placental compromise, and is associated with more than seventy percent of placental arterial obliteration [912]. The presence of absent or reversed end diastolic flow in the umbilical artery is commonly associated with severe intrauterine growth restriction and oligohydramnios [13,14]. Doppler waveforms of the umbilical arteries can be obtained from any segment along the umbilical cord. Waveforms obtained from the placental end of the cord show more end diastolic flow than

waveforms obtained from the abdominal cord insertion [15,16]. Differences in Doppler indices of arterial waveforms obtained from different anatomic locations of the same umbilical cord are generally minor and have no significance on clinical practice [3,4].

Middle cerebral circulation


The cerebral circulation is normally a highimpedance circulation with continuous forward flow throughout the cardiac cycle [Fig. 5] [17,18]. The middle cerebral artery is the most accessible cerebral vessel to ultrasound imaging in the fetus, and it carries more than 80% of cerebral blood flow [19,20]. In the presence of fetal hypoxemia, central

Fig. 2. Doppler indices that are commonly used in obstetrics. D, diastole; PI, pulsatility index; RI, resistive index; S, systole.

Doppler Ultrasound in Obstetrics

295

Fig. 3. Umbilical artery Doppler waveforms in a normal fetus in the third trimester of pregnancy. Note the lowimpedance circulation with increased end-diastolic velocity.

redistribution of blood flow occurs, resulting in an increased blood flow to the brain, heart, and adrenals, and a reduction in flow to the peripheral and placental circulations. This blood flow redistribution is known as the brain-sparing reflex, and plays a major role in fetal adaptation to oxygen deprivation [Fig. 6] [17,2022]. The right and left middle cerebral arteries represent major branches of the circle of Willis in the fetal brain. The circle of Willis, which is supplied by the internal carotids and vertebral arteries, can be imaged with color flow Doppler ultrasound in a transverse plane of the fetal head obtained at the base of the skull [Fig. 7]. In this transverse plane, the proximal and distal middle cerebral arteries are seen in their longitudinal view, with their course almost parallel to the ultrasound beam [see Fig. 7]. Middle cerebral artery Doppler waveforms, obtained from the proximal portion of the vessel, immediately after its origin from the circle of Willis, have shown the best reproducibility [18,23].

Fetal venous Doppler


Doppler waveforms obtained from the central venous circulation in the fetus reflect the physiologic status of the right ventricle. Specific information with regards to right ventricular preload, myocardial compliance, and right ventricular end-diastolic pressure can be derived from Doppler flow studies of the inferior vena cava and ductus venosus in the fetus [2433]. Inferior vena cava Doppler waveforms can be obtained from a coronal plane of the chest and abdomen. In this view, the inferior vena cava can be imaged as it enters into the right atrium, joined by the ductus venosus and the left hepatic vein [Fig. 8]. The inferior vena cava can be studied at two locations: at the inlet into the right atrium, or in the segment between the entrance of the renal vein and the ductus venosus. A good correlation coefficient exists between these two measurement sites, and the location that provides the smallest angle of insonation with the blood flow should be

Fig. 4. Umbilical artery Doppler waveform in a severely growth restricted fetus showing abnormal Doppler waveforms with reversed diastolic flow.

296

Abuhamad

Fig. 7. Color Doppler at the base of the fetal brain showing the circle of Willis. MCA, middle cerebral artery.

Fig. 5. Middle cerebral artery Doppler waveforms in a normal fetus in the third trimester of pregnancy. Note the high-impedance circulation with decreased enddiastolic velocity.

chosen [25,31]. Inferior vena cava Doppler waveforms are triphasic in shape, with the first phase corresponding to ventricular systole, the second phase to early diastole, and the third phase to late diastole or the atrial kick [Fig. 9]. Ductus venosus Doppler waveforms can be easily obtained from a transverse view of the fetal abdomen at the same anatomic plane of the abdominal circumference. By superimposing color flow Dopp-

ler to the gray-scale image, the ductus venosus can be identified as it branches from the umbilical vein [Fig. 10]. Turbulence is commonly seen within the ductus venosus given its narrow lumen [see Figs. 9, 10]. The presence of turbulence on color flow Doppler helps in identifying the ductus venosus in early gestations. Ductus venosus Doppler waveforms are biphasic in shape, with the first phase corresponding to ventricular systole, the second phase to early diastole, and the nadir of the second phase to late diastole or the atrial kick [Fig. 11].

Fetal cardiac Doppler


Doppler indices in fetal echocardiography are quantitative parameters, and are for the majority, angle dependent. To obtain accurate Doppler indices in fetal echocardiography, the sample volume is placed distal to the respective valves, the insonat-

Fig. 6. Middle cerebral artery Doppler waveforms in a growth-restricted fetus showing a low-impedance circulation with an increase in the end-diastolic velocity.

Fig. 8. Coronal view of the fetal chest and abdomen with color Doppler showing the ductus venosus (DV) and the hepatic vein (superior to the ductus venosus) joining the inferior vena cava (IVC) before it enters the right atrium.

Doppler Ultrasound in Obstetrics

297

Fig. 9. Doppler velocity waveforms of the inferior vena cava in a normal fetus in the third trimester of pregnancy. See text for details.

ing angle should be within 15 to 20 of the direction of blood flow, Doppler waveforms should be obtained during fetal apnea, and multiple measurements should be made. Color Doppler is used to direct placement of the sample volume; placing the sample volume at the brightest colors of the blood flow segment will ensure the best measurements. Fig. 12 shows Doppler indices commonly used in fetal echocardiography. The fetal circulation is different from the adult circulation in many aspects. The fetal circulation is in parallel, rather than in series, and the right ventricular cardiac output is greater than the left ventricular cardiac output [6,7,34,35]. The progressive development of organs during gestation influences blood distribution and vascular impedance [6,34]. With advancing gestation, ventricular compliance is increased, total peripheral resistance is decreased, preload is increased, and combined cardiac output is increased [6,34]. Compliance of the fetal left heart increases more rapidly than compli-

Fig. 10. Transverse plane of the fetal abdomen at the level of the abdominal circumference. Color Doppler ultrasound shows the ductus venosus (arrow) as it branches from the umbilical vein.

ance of the fetal right heart with advancing gestation [6,34]. The pulmonary vascular resistance is high in the fetus and the pulmonary arterial pressure is almost systemic [9,10,32,36]. Flow to the pulmonary vascular bed is maintained at a low rate with a noted increase toward the end of gestation [7,9,32,35]. Cardiac output in the fetus is mainly affected by preload and ventricular compliance [6,34]. The presence of right to left shunts at the level of the foramen ovale and ductus arteriosus has a significant impact on cardiac flow patterns and affects the distribution of blood and oxygen to various organs. Flow across the foramen ovale contributes to the majority of blood entering the left ventricle and more than two thirds of right ventricular output is directed to the ductus arteriosus [7,13,35,37]. This shunting mechanism ensures the delivery of blood with high oxygen content to the coronary and cerebral circulations. Doppler waveforms across the atrioventricular valves are bicuspid in shape. The first peak (E wave), corresponds to early ventricular filling of diastole, and the second peak (A wave) corresponds to atrial systole or the atrial kick. Unlike in postnatal life, the velocity of the A wave is higher than that of the E wave in the fetus [6,15,34,38]. This highlights the importance of the role that atrial systole plays in cardiac filling in the fetus. E/A ratio increases with advancing gestation and reflects ventricular diastolic function [6,15,34,38]. E and A velocity peaks are higher in the right ventricle, and this right ventricular dominance is noted from the first trimester [6,15,17,34,38,39]. Shifting to left ventricular dominance starts in utero toward the end of gestation [6,34]. E/A ratio is an index of ventricular preload and compliance [6,34]. Doppler waveforms across the semilunar valves are uniphasic in shape. Indices most commonly used for the semilunar Doppler waveforms include the peak systolic velocity (PSV) and the time to peak velocity (TPV). PSV and TPV increase with

298

Abuhamad

Fig. 11. Doppler velocity waveforms of the ductus venosus in a normal fetus in the third trimester of pregnancy. See text for details.

advancing gestation across the semilunar valves [2,3,5,6,8,13,19,21,23,33,34,37]. PSV is higher across the aorta than across the pulmonary artery because of a decreased afterload and a smaller diameter across the aorta [2,3,5,6,8,13,19,21,23, 33,34,37]. These Doppler indices reflect ventricular contractility, arterial pressures, and afterloads.

Fetal Doppler and intrauterine growth restriction


Arterial Doppler abnormalities, at the level of the umbilical and middle cerebral arteries (brain-sparing reflex), confirm the presence of hypoxemia in the growth-restricted fetus, and present early warning signs. Once arterial centralization occurs, however, no clear trend is noted in the observational period, and thus arterial redistribution may not be helpful for the timing of the delivery [11,12,14,

4042]. On the other hand, the presence of reversed diastolic flow in the umbilical arteries is a sign of advanced fetal compromise, and strong consideration should be given for delivery, except for extreme prematurity. Cesarean section should be given preference in this setting, because labor may cause further fetal compromise. The current literature suggests that venous Doppler abnormalities in the inferior vena cava and ductus venosus and abnormal fetal heart rate monitoring, even in its computerized version, follow arterial Doppler abnormalities and are thus associated with a more advanced stage of fetal compromise [16,18,20,22,3639,43,44]. Furthermore, in the majority of severely growthrestricted fetuses, sequential deterioration of arterial and venous Doppler precedes biophysical profile score deterioration [22,37]. At least one third of fetuses show early signs of circulatory deregula-

Fig. 12. Doppler indices that are commonly used in fetal echocardiography. (A) Peak systolic velocity (PSV) is the peak velocity achieved during one cardiac cycle. (B) Acceleration time (AT) is the time it takes the velocity to reach its peak in one cardiac cycle. (C ) Time velocity integral (TVI) is the integral of the planimetric area under the curve. TVI expresses the distance that the red blood cells would have to cover with a constant area of the flow section.

Doppler Ultrasound in Obstetrics

299

tion 1 week before biophysical profile deterioration, and that in most cases, Doppler deterioration preceded biophysical profile deterioration by 1 day [22,37]. The occurrence of such abnormal late stage changes of vascular adaptation by the intrauterine growth-retarded (IUGR) fetus appears to be the best predictor of perinatal death, independent of gestational age and weight [39,43]. In a longitudinal study on Doppler and IUGR fetuses, all intrauterine deaths and all neonatal deaths, with the exception of one case, had late Doppler changes at the time of delivery, whereas only a few of the surviving fetuses showed such changes [39,43]. This sequential deterioration of the hypoxemic, growth-restricted fetus is rarely seen at gestations beyond 34 weeks [18,45,46]. Indeed, normal umbilical artery Doppler is common in growthrestricted fetuses in late gestations, and cerebroplacental ratios have poor correlation with outcome of IUGR fetuses at greater than 34 weeks of gestation [35,47]. Caution should therefore be exercised when Doppler is used in the clinical management of IUGR fetuses beyond 34 weeks of gestation. The pathophysiology of fetal growth restriction has not been fully described because recent studies have highlighted the presence of significant variation in fetal adaptation to hypoxemia. The pattern of incremental deterioration of arterial Doppler abnormalities, followed by venous Doppler abnormalities, then followed by fetal heart tracings and biophysical profile abnormalities, is not seen in about 20% of preterm fetuses [16,36]. Furthermore, only 70% of IUGR fetuses show significant deterioration of all vascular beds by the time they were delivered, and about 10% showed no significant circulatory change by delivery time [22,37]. In a prospective, observational study, more than 50% of IUGR fetuses delivered because of abnormal fetal heart rate tracings did not have venous Doppler abnormalities [39]. In view of these findings, the universal introduction of venous Doppler in the clinical management of the growth-restricted fetus should await the results of randomized trials on this subject. IUGR is associated with several changes at the level of the fetal heart involving preload, afterload, ventricular compliance, and myocardial contractility. An increase in afterload is seen at the level of the right ventricle because of increased placental impedance [25,48]. A decrease in afterload is noted at the level of the left ventricle because of decreased cerebral impedance associated with the brain-sparing reflex [25,48]. These changes in afterload result in a redistribution of the cardiac output from right to left ventricle [25,48]. Preload is reduced at both atrioventricular valves because of

hypovolemia and decreased filling associated with IUGR [21,24,26,30,31,45,49,50]. This decrease in preload is reflected by a decrease in the E/A ratio, decreased atrial peak, and decreased time velocity integral at the mitral and tricuspid valves [21,24,26, 45,49,50]. Evidence of reduced myocardial contractility in the presence of severe IUGR has also been reported. Ventricular ejection force, an index of ventricular systolic function that is independent of preload and afterload is decreased at the level of the right and left ventricle in fetal growth restriction [31,36]. IUGR fetuses that have reduced ventricular ejection force have a shorter time to delivery, a higher incidence of nonreassuring fetal heart rate tracing, and a lower pH at birth when compared with controls [31,36]. A significant correlation between the severity of fetal acidosis at cordocentesis and ventricular ejection force values validates the association of this index and the severity of fetal compromise [31,36]. Myocardial cell damage, demonstrated by elevated levels of cardiac troponin-T, is seen in some fetuses that have severe growth restriction [25,28]. This advanced stage of fetal compromise is associated with signs of increased systemic venous pressure, a change in the distribution of cardiac output, a rise in right ventricle afterload, and a high incidence of tricuspid regurgitation [25,28]. These findings suggest that Doppler abnormalities in the proximal venous system of the growth restricted fetus suggest fetal myocardial cell damage and increased systemic venous pressure [25,28]. The fetal heart plays a central role in the adaptive mechanisms for hypoxemia and placental insufficiency. As discussed in this article, longitudinal data on the hemodynamic sequence of the natural history of fetal growth restriction show that the umbilical artery and middle cerebral artery are the first variables to become abnormal [27,34]. These arterial Doppler abnormalities are followed by abnormalities in the right cardiac diastolic indices, followed by the right cardiac systolic indices, and finally by both left diastolic and systolic cardiac indices [27,34]. Preserving the left systolic function as the last variable to become abnormal ensures an adequate left ventricular output, which supplies the cerebral and coronary circulations. Several of the Doppler changes seen in association with fetal IUGR in the peripheral circulation are directly related to the adaptation of the fetal heart. The current management of IUGR involves Doppler at the peripheral arterial circulation (middle cerebral and umbilical arteries), central venous vessels (ductus venosus and inferior vena cava), and cardiotocography. Adding cardiac Doppler may improve management of the IUGR

300

Abuhamad

fetus, but studies are lacking on the prospective clinical evaluation of the IUGR fetus with cardiac Doppler. It is becoming more obvious, however, that those changes in the central venous circulation reflect an advanced stage of fetal compromise, commonly associated with myocardial dysfunction and damage.

[14]

[15]

References
[1] Chervenak FA, Skupski DW, Romero R, et al. How accurate is fetal biometry in the assessment of fetal age? Am J Obstet Gynecol 1998; 178:67887. [2] Doppler C. Uber das farbige Licht der Dopplersterne und einigr anderer Gestirne des Himmels [On the coloured light of double stars and certain other stars of the heavens]. Royal Bohemian Society 1842;2:46582. [3] Fleischer A, Schulman H, Farmakides G, et al. Umbilical artery velocity waveforms and intrauterine growth retardation. Am J Obstet Gynecol 1985;151:5025. [4] Ott WJ. The diagnosis of altered fetal growth. Obstet Gynecol Clin North Am 1988;15:23763. [5] Giles WB, Trudinger BJ, Baird PJ. Fetal umbilical artery flow velocity waveforms and placental resistance: pathological correlation. Br J Obstet Gynaecol 1985;92:318. [6] Manning FA. Intrauterine growth restriction. Diagnosis, prognostication, and management based on ultrasound methods. In: Manning FA, editor. Fetal medicine: principles and practice. Norwalk (CT): Appleton & Lange; 1995. p. 8794. [7] Hadlock FP, Deter RL, Harrist RB, et al. A dateindependent predictor of intrauterine growth retardation: femur length/abdominal circumference ratio. AJR Am J Roentgenol 1983;141:97984. [8] Trudinger BJ, Stevens D, Connelly A, et al. Umbilical artery flow velocity waveforms and placental resistance: the effects of embolization of the umbilical circulation. Am J Obstet Gynecol 1987; 157:14438. [9] Brown HL, Miller Jr JM, Gabert HA, et al. Ultrasonic recognition of the small-for-gestational-age fetus. Obstet Gynecol 1987;69:6315. [10] Creasy RK, Resnick R. Intrauterine growth retardation. In: Creasy RK, Resnick R, editors. Maternal fetal medicine: principles and practice. Philadelphia: W.B. Saunders; 1984. p. 491. [11] Kingdom JC, Burrell SJ, Kaufmann P. Pathology and clinical implications of abnormal umbilical artery Doppler waveforms. Ultrasound Obstet Gynecol 1997;9:27186. [12] Morrow RJ, Adamson SL, Bull SB, et al. Effect of placental embolization on the umbilical arterial velocity waveform in fetal sheep. Am J Obstet Gynecol 1989;161:105560. [13] Bernstein IM, Horbar JD, Badger GJ, et al. Morbidity and mortality among very-low-birth-

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

weight neonates with intrauterine growth restriction. The Vermont Oxford Network. Am J Obstet Gynecol 2000;182:198206. Copel JA, Reed KL, Maulik D, editors. Doppler ultrasound in obstetrics and gynecology. New York: Raven Press; 1995. Kok JH, den Ouden AL, Verloove-Vanhorick SP, et al. Outcome of very preterm small for gestational age infants: the first nine years of life. Br J Obstet Gynaecol 1998;105:1628. Trudinger BJ. Doppler ultrasonography and fetal well being. In: Reece EA, Hobbins JC, Mahoney M, editors. Medicine of the fetus and mother. Philadelphia: J.B. Lippincott; 1992. p. 70122. Fattal-Valevski A, Leitner Y, Kutai M, et al. Neurodevelopmental outcome in children with intrauterine growth retardation: a 3-year followup. J Child Neurol 1999;14:7247. Mari G, Abuhamad AZ, Brumfield J, et al. Doppler ultrasonography of the middle cerebral artery peak systolic velocity in the fetus: reproducibility of measurement [abstract #669]. Am J Obstet Gynecol 2001;185:261. Soothill PW, Ajayi RA, Campbell S, et al. Relationship between fetal acidemia at cordocentesis and subsequent neurodevelopment. Ultrasound Obstet Gynecol 1992;2:803. Veille JC, Hanson R, Tatum K. Longitudinal quantitation of middle cerebral artery blood flow in normal human fetuses. Am J Obstet Gynecol 1993;169:13938. Hanna BD, Nelson MN, White-Traut RC, et al. Heart rate variability in preterm brain-injured and very-low-birth-weight infants. Biol Neonate 2000;77:14755. Mari G, Deter RL. Middle cerebral artery flow velocity waveforms in normal and small-forgestational-age fetuses. Am J Obstet Gynecol 1992; 166:126270. Nijhuis IJ, ten Hof J, Mulder EJ, et al. Fetal heart rate in relation to its variation in normal and growth retarded fetuses. Eur J Obstet Gynecol Reprod Biol 2000;89:2733. Bracero LA, Roshanfekr D, Byrne DW. Analysis of antepartum fetal heart rate tracing by physician and computer. J Matern Fetal Med 2000;9:1815. Devoe L, Golde S, Kilman Y, et al. A comparison of visual analyses of intrapartum fetal heart rate tracings according to the new National Institute of Child Health and Human Development guidelines with computer analyses by an automated fetal heart rate monitoring system. Am J Obstet Gynecol 2000;183:3616. Hecher K, Hackeloer BJ. Cardiotocogram compared to Doppler investigation of the fetal circulation in the premature growth-retarded fetus: longitudinal observations. Ultrasound Obstet Gynecol 1997;9:15261. Huisman TW, Stewart PA, Wladimiroff JW. Flow velocity waveforms in the fetal inferior vena cava during the second half of normal pregnancy. Ultrasound Med Biol 1991;17:67982.

Doppler Ultrasound in Obstetrics

301

[28] Reed KL, Appleton CP, Anderson CF, et al. Doppler studies of vena cava flows in human fetuses. Insights into normal and abnormal cardiac physiology. Circulation 1990;81:498505. [29] Reuss ML, Rudolph AM, Dae MW. Phasic blood flow patterns in the superior and inferior venae cavae and umbilical vein of fetal sheep. Am J Obstet Gynecol 1983;145:708. [30] Ribbert LS, Visser GH, Mulder EJ, et al. Changes with time in fetal heart rate variation, movement incidences and haemodynamics in intrauterine growth retarded fetuses: a longitudinal approach to the assessment of fetal well being. Early Hum Dev 1993;31:195208. [31] Rizzo G, Arduini D, Romanini C. Inferior vena cava flow velocity waveforms in appropriate- and small-for-gestational-age fetuses. Am J Obstet Gynecol 1992;166:127180. [32] Rizzo G, Capponi A, Talone PE, et al. Doppler indices from inferior vena cava and ductus venosus in predicting pH and oxygen tension in umbilical blood at cordocentesis in growthretarded fetuses. Ultrasound Obstet Gynecol 1996;7:40110. [33] Vindla S, James D, Sahota D. Computerised analysis of unstimulated and stimulated behaviour in fetuses with intrauterine growth restriction. Eur J Obstet Gynecol Reprod Biol 1999; 83:3745. [34] Alfirevic Z, Neilson JP. Doppler ultrasonography in high-risk pregnancies: systematic review with meta-analysis. Am J Obstet Gynecol 1995;172: 137987. [35] Bahado-Singh RO, Kovanci E, Jeffres A, et al. The Doppler cerebroplacental ratio and perinatal outcome in intrauterine growth restriction. Am J Obstet Gynecol 1999;180:7506. [36] Baschat AA, Gembruch U, Reiss I, et al. Relationship between arterial and venous Doppler and perinatal outcome in fetal growth restriction. Ultrasound Obstet Gynecol 2000;16:40713. [37] Baschat AA, Gembruch U, Harman CR. The sequence of changes in Doppler and biophysical parameters as severe fetal growth restriction worsens. Ultrasound Obstet Gynecol 2001;18: 5717. [38] Hecher K, Bilardo CM, Stigter RH, et al. Monitoring of fetuses with intrauterine growth restriction: a longitudinal study. Ultrasound Obstet Gynecol 2001;18:56470. [39] Ferrazzi E, Bozzo M, Rigano S, et al. Temporal sequence of abnormal Doppler changes in the peripheral and central circulatory systems of the

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

severely growth-restricted fetus. Ultrasound Obstet Gynecol 2002;19:1406. Baschat AA, Gembruch U, Gortner L, et al. Coronary artery blood flow visualization signifies hemodynamic deterioration in growthrestricted fetuses. Ultrasound Obstet Gynecol 2000;16:42531. Baschat AA, Gembruch U, Weiner CP, et al. Longitudinal changes of arterial and venous Doppler in fetuses with intrauterine growth restriction [abstract]. Am J Obstet Gynecol 2001; 184:103. Senat MV, Schwarzler P, Alcais A, et al. Longitudinal changes in the ductus venosus, cerebral transverse sinus and cardiotocogram in fetal growth restriction. Ultrasound Obstet Gynecol 2000;16:1924. Behrman RE, Lees MH, Peterson EN, et al. Distribution of the circulation in the normal and asphyxiated fetal primate. Am J Obstet Gynecol 1970;108:95669. Pardi G, Cetin I, Marconi AM, et al. Diagnostic value of blood sampling in fetuses with growth retardation. N Engl J Med 1993;328:6926. Harrington K, Thompson MO, Carpenter RG, et al. Doppler fetal circulation in pregnancies complicated by pre-eclampsia or delivery of a small for gestational age baby: 2. Longitudinal analysis. Br J Obstet Gynaecol 1999;106:45366. Hecher K, Campbell S, Doyle P, et al. Assessment of fetal compromise by Doppler ultrasound investigation of the fetal circulation. Arterial, intracardiac, and venous blood flow velocity studies. Circulation 1995;91:12938. Wladimiroff JW, vd Wijngaard JA, Degani S, et al. Cerebral and umbilical arterial blood flow velocity waveforms in normal and growthretarded pregnancies. Obstet Gynecol 1987;69: 7059. Gramellini D, Folli MC, Raboni S, et al. Cerebralumbilical Doppler ratio as a predictor of adverse perinatal outcome. Obstet Gynecol 1992;79: 41620. Arduini D, Rizzo G. Prediction of fetal outcome in small for gestational age fetuses: comparison of Doppler measurements obtained from different fetal vessels. J Perinat Med 1992;20:2938. Fong KW, Ohlsson A, Hannah ME, et al. Prediction of perinatal outcome in fetuses suspected to have intrauterine growth restriction: Doppler US study of fetal cerebral, renal, and umbilical arteries. Radiology 1999;213:6819.

303

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 303319

Sonography of the Lower Uterine Segment


Lucy Chie,
& & & &

MD

, Deborah Levine,

MD

b,*

The lower uterine segment in normal pregnancy Uterine dehiscence and rupture Pitfalls in assessing cervical length and funneling Placental imaging in the lower uterine segment

& & & & & &

Placenta previa Placenta accreta, increta, and percreta Leiomyomas in the lower uterine segment Ectopic pregnancy in the uterine scar Summary References

The lower uterine segment (LUS) is a dynamic component of the uterus that plays an important role in obstetrics. Historically, the LUS has been defined in metric, anatomic, and physiologic terms. Metrically, it is the portion of the uterus toward term that lies within one fingerbreadth of the internal os of the cervix. Anatomically, it lies inferiorly to the natural vesicouterine reflection. Physiologically, it is the part of the uterus which passively stretches in labor and takes hardly any active role in the expulsion of the fetus [1]. The LUS in late pregnancy forms from what in the nonpregnant patient and in early pregnancy are the isthmus and upper portion of the anatomic cervix [Figs. 1, 2] [2]. Obstetricians commonly use the LUS for cesarean delivery because of its relatively amuscular nature in later pregnancy. However, scar tissue in the LUS puts the patient at risk for uterine dehiscence and rupture. When comparing patients who have had prior LUS cesarean section to those with a prior vertical incision, the LUS cesarean section patients have an associated decreased risk of uterine rupture in subsequent pregnancies, as well as lower risks of
a

intra-abdominal adhesion formation and bladder injury. Cesarean delivery currently accounts for over a quarter of the deliveries in the United States, and thus increasing attention has been given to the anterior aspect of the LUS. Imaging of the LUS aids in the management of labor and delivery in patients at risk for these suspected conditions, as well as for placenta accreta and cesarean scar ectopic pregnancies. Important pitfalls in imaging the LUS include contractions that can mimic a funneled cervix and make assessment for placenta previa difficult, as well as an overly distended bladder that can compress the LUS and cause a normal placenta to resemble placenta previa. In this article, the authors discuss and illustrate the various sonographic appearances of the LUS and the implications of these findings.

The lower uterine segment in normal pregnancy


During pregnancy, the uterus undergoes remarkable changes, increasing in capacity 500 to 1000 fold to

Department of Obstetrics and Gynecology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA b Department of Radiology, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA * Corresponding author. E-mail address: dlevine@bidmc.harvard.edu (D. Levine).
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.002

ultrasound.theclinics.com

304

Chie & Levine

Fig. 1. Location of prior cesarean section as shown in nongravid uterus. Transvaginal sagittal image of the uterus during a sonohysterogram reveals a wedge-shaped myometrial defect in the region of the cesarean section scar. The myometrium anterior to the defect is measured with calipers. This defect is in a portion of the uterus that in the pregnant state forms the LUS.

accommodate on average 5 liters at term [3]. Uterine enlargement primarily involves stretching and marked hypertrophy of the smooth muscle cells. During the first trimester, uterine hypertrophy is likely due to the action of estrogen and possibly of progesterone, whereas beyond the first 12 weeks, the changes are in large part related to pressure effects of the growing fetus and increasing amniotic fluid. The uterus usually reaches its term weight by the early second trimester; thus during the second half of pregnancy, uterine enlargement continues in a nonuniform manner, with marked concentration of smooth muscle cells at the fundus. The myome-

trium comprises the major portion of the uterus and consists of smooth muscle bundles united by connective tissue. The number of muscle fibers diminishes caudally, such that in the cervix, only 10% of the tissue mass is muscle [4]. As growth proceeds, the uterine fundus becomes dome-shaped, and the LUS thins. Several investigators have studied the sonographic changes in uterine wall thickness with pregnancy. The uterine wall consists of three layers: the serosa, myometrium, and endometrium. During pregnancy, additional layers that can be visualized with ultrasound include the decidual layer and the placenta.

Fig. 2. Location of scar as shown in gravid uterus in a patient 14 weeks pregnant who had a cesarean section in prior pregnancy after prolonged labor. The scar in this patient (arrows) is seen lower than that in Fig. 1, and appears to be in the anatomic cervix. This low position illustrates that the anatomic cervix late in pregnancy is a portion of the LUS.

Sonography of the Lower Uterine Segment

305

Fig. 3. Normal appearance of myometrium at 34 weeks gestational age. Linear transabdominal image shows normal-appearing myometrium. The myometrial thickness should be measured perpendicular to the length of the myometrium (arrows), rather than where the calipers are placed in this example, and measures 4 mm in this example.

At term, the uterine corpus wall typically measures 15 mm or less in thickness. Sonographically, the myometrium is measured as the echo-homogenous layer between the serosa and the decidua [Fig. 3]. In a study of 25 low-risk singleton pregnancies, Degani and colleagues [5] followed myometrial thickness from early second trimester to term. They found that uterine wall thickness remained relatively constant throughout gestation in all locations except at the anterior LUS. With increasing gestation, they reported a significant decrease in thickness of the anterior LUS, from approximately 8 mm to less than 6 mm. They also found that placental location influenced the extent of uterine hypertrophy. They found an increased anterior LUS thickness of 10.3 1.5 mm with an anterior placenta, versus 8.9 1.4 mm with placentas in other locations [5]. In a study of 52 low-risk singleton pregnancies, Buhimschi and coworkers [6] found that myometrial thickness of the unscarred LUS at term was similar in nonlaboring and laboring patients, 4.68 mm and 4.66 mm, respectively. In a study of patients who had and did not have cesarean deliveries, Gotoh and colleagues [7] found no difference in myometrial thickness measurements (6.76.8 mm) between these two groups at 19 weeks, but significantly different thickness measurements after 29 weeks. At 39 weeks, myometrial thickness in cesarean and control patients measured 2.1 mm and 3 mm, respectively. In summary, these studies demonstrate that: (1) the anterior LUS during gestation undergoes changes in myometrial thickness, unlike the rest of the uterine corpus; (2) placental location can influence this measurement; (3) labor may not affect the myometrial thickness of the unscarred LUS at term; and (4) ce-

sarean scar can significantly decrease the myometrial thickness at term.

Uterine dehiscence and rupture


As the incidence of cesarean deliveries rises, the number of patients who face the decision between a trial of labor and a repeat cesarean section increases. The most serious complication of attempted vaginal birth after cesarean section is uterine rupture. Uterine rupture refers to the complete nonsurgical separation of the uterine wall, resulting in communication between the uterine and peritoneal cavities [3]. Although uterine rupture can occur spontaneously or from other uterine scars (ie, myomectomy, laparoscopy, hysteroscopy), cesarean scar is by far the most common cause of uterine rupture. This can lead to life-threatening consequences, including maternal hemorrhage and fetal-placental extrusion into the peritoneal cavity. Uterine dehiscence refers to the incomplete separation or thinning of the uterine wall, occurs in an estimated 4% of patients who had prior cesareans, and usually does not result in major clinical problems [Fig. 4]. Although the risk of symptomatic uterine rupture is low (0.7%) in patients who had prior LUS cesarean undergoing trial of labor [3], the associated morbidity and mortality demands careful selection of patients and close management in labor. Because the incidence of uterine rupture is much lower than that of uterine dehiscence, clearly not every thin uterus will rupture. What constitutes clinically important uterine wall thinning and increased risk of uterine rupture has been investigated in several studies [713], with a wide range

306

Chie & Levine

Fig. 4. Thin LUS at 32 weeks gestational age. (A) Transabdominal curvilinear image shows that the myometrium is difficult to measure secondary to pressure from the fetal head on the LUS. (B) Image taken with linear transducer with head slightly away shows thin myometrium (calipers) measuring less than 1 mm. Note the fetal scalp (S).

of safe values of LUS thickness, from 1.6 mm or more to 3.5 mm or more [Table 1]. The various methods and definitions used for uterine rupture and dehiscence as well as the nonuniform techniques to measure the LUS likely account for some of the differing cutoff values reported. In the largest study of patients who had prior cesarean, Rozenberg and colleagues [8] scanned 642 patients at 36 to 38 weeks using transabdominal ultrasonography, and found the overall frequency of defective uterine scars was 4.0% (15 ruptures, 10 dehiscences), and that the frequency of defects rose as the thickness of the lower uterine segment decreased [see Table 1].

The utility of imaging the LUS to assess uterine dehiscence and the risk of rupture has been debated. In a survey of Canadian obstetricians [14], 16% of respondents reported using LUS thickness as measured by ultrasound at or near term to determine which women were suitable candidates for trial of labor after cesarean. Although the high negativepredictive value of the LUS thickness may be useful in detecting patients at low risk for uterine rupture, the low positive-predictive value has limited the adoption of LUS measurement as a screening tool in the management of patients who have scarred uteri. When uterine dehiscence is identified, the risk of uterine rupture remains unclear, and the management

Table 1: LUS thickness and incidence of dehiscence Incidence of dehiscence (%) GA (weeks) 3638 near term 3840 <3 (N=19) <2 (N=20) <2 (N=23) 1.6 (N=7) 3 (N=18) 2.5 (N=8) 2 (N=1) 3 (N=19) 2.5 (N=5) 2 (N=1) 100 77.8 90.9 90.9 45.5 90.9 81.8 63.6 100 100 36 7 days before CD term term 100 88 32 100 82 83 88 88.6 68 84 100 72 84 100 45.0 73.2 93.0 88.6 4.0 LUS thickness cutoff (mm) Sensitivity (%) Specificity (%) PPV (%) NPV (%)

Study

TA/TV

# with prior CD

Rozenberg et al, 1996 [9] 6 NR NR NR 4.7 2.82

TA

642

TA

84

4.5 (N=177) 3.5 (N=136) 2.5 (N=51) <2 (N=14)

6.9 11.8 15.7 35.7 87 64 73.9 25.9 55.6 71.4 100 58.8 69.2 100

100 99.3 97.1 100 100 100 100 98.7 94.4 95.5 86.2 94.7 91.3 86.2

TA

50

TA

39

TV

348

TV

186

Fukuda et al, 1988 [12] Tanik et al, 1996 [8] Suzuki et al, 2000 [13] Gotoh et al, 2000 [7] Asakura et al, 2000 [10] Sen et al, 2004 [11]

TA TV

71

Sonography of the Lower Uterine Segment

Numbers in bold are those given by the authors as the cutoff value having the best performance. Abbreviations: CD, cesarean delivery; GA, gestational age; NPV, negative-predictive value; NR, not reported; PPV, positive-predictive value; TA, transabdominal; TV, transvaginal.

307

308

Chie & Levine

Fig. 5. Transabdominal imaging revealing a 3.4 cm defect in the LUS in a patient with history of vaginal birth after cesarean section. The patient first had this defect diagnosed at 21 weeks gestational age, but despite counseling regarding risks of continuing the pregnancy, decided to remain pregnant. At 30 weeks, a bulge is seen in the LUS, extending outside the normal uterine contours. Calipers denote the diameter of the defect. The patient underwent classical cesarean at 31 weeks gestational age due to nonreassuring fetal testing. During the procedure, a 4 cm LUS defect was palpated.

Fig. 6. Uterine dehiscence/rupture in a patient with history of prior uterine rupture and cesarean delivery. (A) Sagittal transabdominal image at 35 weeks gestational age shows extremely thin myometrium (arrow). The study was limited by patient body habitus, and this region could not be adequately assessed transvaginally. (B) Sagittal MR image shows myometrium anteriorly (arrows) but no visible myometrium just above the bladder in the LUS (arrowhead). Because the uterine contour was intact on the sonogram, and the patient was asymptomatic, the decision was made to follow her closely. The patient was admitted to the hospital for observation and given steroids to help the fetal lungs mature. Delivery was by cesarean section 9 days later when she had abdominal pain. At the time of delivery, a 4 cm gap was observed in the lower uterine segment, held together by dense adhesions. Due to partial volume averaging on MR it is very difficult to assess the thin lower uterine segment in pregnancy. Clinical care is based on the imaging findings, the patients symptoms, and the gestational age of the pregnancy.

Sonography of the Lower Uterine Segment

309

of these patients can pose clinical dilemmas. At what lower limit is trial of labor unsafe? At what cutoff value should early intervention (such as strict bed rest or early delivery) be considered? When assessing for risk of uterine dehiscence, the LUS is scanned in transverse and sagittal plane for any obvious regions of thinning or abnormal uterine contour. The measurement of LUS thickness is made at the thinnest region. It is also important to describe defects in the LUS, including symmetry, movement, ballooning, and presence of a defect [Fig. 5] [15,16]. Although both transabdominal and transvaginal approaches have been employed, one small study [17] found less inter- and intraobserver variability with transvaginal sonography. MRI also has been used to investigate the LUS, but is limited in that

when the LUS is thin (less than 1 mm), partial volume averaging over a 3 to 8 mm slice will lead to overcalling defects using this imaging technique [Fig. 6] [1820]. At the authors institution, we measure the LUS when it subjectively appears to be thin. A measurement of less than 2 mm is documented in the report and communicated to the referring clinician. Both transabdominal and transvaginal scans are performed. Because scars can cause adherence between the bladder and LUS, a full bladder will actually cause the LUS to thin. Pressure from the fetal head can also compress the LUS and make the measurement thinner than it otherwise would be. The authors attempt to optimize the study by having the examination performed with a small

Fig. 7. Contraction of the LUS obscures a shortened and funneled cervix at 16 weeks gestational age. (A) Transvaginal image shows thick myometrium anteriorly and posteriorly. The calipers were placed as a guess as to cervical length. The image in (B) was obtained a few minutes after the image in (A). Note that the funneled cervix was totally obscured by the contraction. It is not adequate to guess at the location of the internal os; to perform an adequate cervical examination, one must wait for the LUS contraction to resolve.

310

Chie & Levine

amount of urine in the bladder and no compression of the LUS by the fetal presenting part. A careful search is made for any uterine contour abnormalities, which would suggest dehiscence. Decisions regarding trial of labor are made by the patient and referring clinician.

Pitfalls in assessing cervical length and funneling


The most commonly used sonographic parameters in the assessment of premature labor are cervical length, dilation of the cervix, and the appearance of

the internal cervical os. Although it is beyond the scope of this article to detail the technical aspects of cervical sonography, it is important to realize that contractions in the LUS and overdistension of the bladder in the LUS can mimic a funneled cervix [Fig. 7] [2123], and that these same conditions can mask a shortened cervix [Fig. 8]. The artifact caused by overdistension of the bladder can be recognized by the abnormally long cervix if measured from the purported internal os to the external os. Transvaginal scan of the cervix largely eliminates this artifact. The artifact caused by contraction is recognized by the thickened myometrium above

Fig. 8. Pitfalls of overfull bladder and contraction at lower uterine segment masking a shortened cervix. (A) Compression of the cervix by a full bladder may give the impression of a long cervix. In this image there is also a contraction at the lower uterine segment that adds to the impression of a long cervix. The contraction can be recognized by the thickness of the myometrium in the anterior and posterior lower uterine segment. This can give the false appearance of funneling above a long, closed cervix. The 3.7 cm measurement of the cervix (calipers) in this image is not sufficient for accurate measurement because of the overly distended bladder and uterine contraction. (B) Cervical funneling and 2.5 cm cervix (calipers) was demonstrated by transvaginal scan with an emptied bladder a few minutes after the image in (A) was taken.

Sonography of the Lower Uterine Segment

311

the cervix. This condition can be clarified by extending the period of observation until the contraction disappears, usually within 20 minutes.

Placental imaging in the lower uterine segment


Localization of the placenta is an important indication for obstetrical imaging because it directly impacts the mode of delivery. An unsuspected placenta previa or placenta accreta can significantly increase the morbidity of pregnancy [3].

Placenta previa
Placenta previa, implantation of the placenta over the cervical os, occurs in roughly 1 in 200 to 1 in 400 pregnancies [3]. Incidence increases with maternal age, parity, prior curettage for spontaneous or induced abortions, prior cesarean delivery, smoking, living at higher altitudes, and multiple gestation [2428]. Placenta previa at term necessitates cesarean delivery, and the clinician must be prepared for increased blood loss and potential cesarean hysterectomy at the time of delivery. Although transabdominal ultrasound was initially used for the diagnosis of placenta previa, this technique has been found to have false-positive and negative rates of approximately 7% [29]. A study by Smith and colleagues [30] showed that when a low placenta is observed transabdominally, the improved visualization provided by transvaginal sonography changes the diagnosis in 26% of cases. Transperineal approach has also been described with a negative-predictive value of 100% and positive-predictive value of 90%[31]; however, when placenta previa is suspected, transvaginal ultrasound has now become the standard modality for confirmation, with diagnostic accuracy greater than 99% [32]. Patients can be reassured that a gentle vaginal examination for placenta previa is safe [33,34]. Using a protocol of confirming placenta previa with transvaginal scanning can decrease the need for repeat examinations later in pregnancy, decrease patient anxiety, and eliminate the need for decreased activity in patients who have an incorrect diagnosis of placenta previa. Transvaginal ultrasound not only assists in diagnosis of placenta location, but can measure the exact distance from the placental edge to the internal os. Four categories of abnormality in placental location have been defined: (1) complete placenta previa [Fig. 9], when the placenta covers the internal os completely (may be asymmetric with only a portion of the placenta crossing the os); (2) partial placenta previa, when the placenta partially covers the internal os [Fig. 10]; (3) marginal placenta
Fig. 9. Complete previa at 32 weeks gestational age. Transabdominal sagittal image shows the placenta (P) centered over the internal os.

previa, when the placental edge just reaches the margin of the internal os; and (4) low placenta, when the placental edge does not reach, but is within 2 cm of the internal os [Fig. 11] [3]. Ultrasound usually cannot distinguish between partial and marginal placenta previa, so they are commonly grouped together. It is helpful to diagnose not only placenta previa but also low placenta before delivery, because the potential for uterine atony and hemorrhage exists with any placenta in the LUS. Using transvaginal ultrasonography, Oppenheimer and coworkers [35] demonstrated that when the lower placental margin was less than 2 cm away from the internal cervical os in the third trimester, women often experienced bleeding and required a cesarean delivery. Ultrasonography has a high sensitivity in detection of low placentas, but also has a high falsepositive rate, especially early in pregnancy. In a study of 3696 patients [36], placenta previa was noted in 1.5% of pregnancies at 18 to 23 weeks gestation, but in only 0.14% of pregnancies at term. Thus, at the time of most second trimester anatomical survey scans, usually only a potential diagnosis of placenta previa can be made. The apparent migration of the placenta away from the cervical os results from: (1) the imprecision with which the placenta edges can be defined at earlier gestations; (2) elongation of the LUS as pregnancy progresses [37]; and (3) trophotropism, in which the placenta best grows in the regions of greatest blood supply. A placenta that appears to migrate away from the internal os likely never had placental villi attachments across the os. The later in gestation at which the previa can be visualized, the more likely it will be present at delivery [Table 2]. When placenta previa is diagnosed early in pregnancy, the patient

312

Chie & Levine

Fig. 10. Placenta previa at 30 weeks gestational age. Transvaginal sagittal image shows the placenta (P) with placental edge (arrow) overlying the internal cervical os. The designation of this as a partial previa or marginal previa is less important than the description of the findings, because cesarean delivery will be needed if the appearance does not change; however, in this case, subsequent imaging revealed placental migration, and the patient underwent vaginal delivery at 39 weeks gestational age.

should undergo rescanning at about 28 weeks of gestational age to ensure that placenta previa is still present. This allows for visualization of the tip of the placenta, whereas later in gestation, as the head descends, it may be difficult to visualize the placental margin. Various predictors have been described to determine the likelihood of persistence of previa at delivery. The first predictor is the degree of overlap of

Fig. 11. Low placenta without previa at 19 weeks gestational age. Transvaginal sagittal image shows a posterior placenta (P) 1.40 cm away (calipers) from the internal os.

the placental edge over the internal os [see Table 2] [35,36,3841]. A second predictor is placental edge thickness: those who have thin placental edge, measuring less than 1 cm in thickness or presenting at an angle less than 45, have a significantly higher rate of vaginal delivery; whereas those who have a thick placental edge have increased rates of emergent cesarean deliveries, hemorrhage, accreta, and preterm delivery [42]. A third predictor is the rate of apparent placental migration: those patients requiring cesarean for complications attributable to previa have a mean rate of placental migration of 0.3 mm/week, versus 5.4 mm/week for those who have a vaginal delivery or cesarean for other indication [43]. A common pitfall in the diagnosis of placenta previa is an overdistended bladder compressing the anterior and posterior myometrium, resulting in the placental margin appearing adjacent to the internal os, when the margin is actually removed from the cervix. Thus, if there is concern for placenta previa with a full bladder, examination should be repeated with an empty bladder. Another pitfall in the diagnosis of placenta previa involves imaging the LUS during uterine contraction [Fig. 12]. This should be suspected when the myometrium appears thick (greater than 1.5 cm) in the LUS. When resolving these pitfalls, it is important to remember that after the bladder is emptied, a contraction frequently occurs. Therefore, it is important to wait

Table 2: Overlap of placenta over internal os and persistence of previa at term

Study 0 6.2 2.4 0.14 NR 0.16 80 0.32 100 15

# of screened patients

Gestational age (weeks)

Extension of placenta over internal os (mm) Sensitivity (%)

Incidence of placenta previa at ultrasound (%)

Incidence of placenta previa at term (%)

Specificity (%) 94.2 97.7

PPV (%) 5.2 5.1

NPV (%) 100 100

1252

913

6428

1216

3696

1823 0.7 0.3 NR

Hill et al, 1995 [38] Taipale et al, 1997 [39] Taipale et al, 1998 [36] Dashe et al, 2002 [40] 15 25 0 100 80 NR

NR

(N=363) (N=131) (N=97) (N=171) (N=158) 0 10 0 1.1 NR NR 0.17

99.4 99.8 NR

100 100 NR

Sonography of the Lower Uterine Segment

Lauria et al, 1996 [41]

2910

1519 2023 2427 2831 3235 1524 2440

100 100 100

NR 84 73

19 40 12 34 49 62 73 14 38 65

100 100 100

313

314

Chie & Levine

Fig. 12. Contraction causing appearance of placenta previa at 17 weeks gestational age. (A) Sagittal transabdominal image with a uterine contraction. The clue to the presence of the contraction is the thickness of the anterior and posterior LUS myometrium. Note the placenta (P) appearing to cross the internal os. (B) Sagittal view a few minutes after the contraction resolves. The placental margin is seen to be away from the internal os. The patient subsequently underwent vaginal delivery at term.

to image the potential placenta previa until after the contraction has resolved. Townsend and coworkers [44] found that 72% of second trimester placenta previa were false-positive, and that two thirds of these were due to either uterine contractions or to a full bladder [44]. A final pitfall is the circumferential low placenta that may be seen in the LUS anterior and posteriorly, but never actually implants on the cervix.

Placenta accreta, increta, and percreta


With any case of placenta previa or having history of cesarean delivery, the risk of placenta accreta, increta, and percreta is increased. Placenta accreta and its variants are forms of abnormal placentation in which trophoblastic invasion penetrates beyond the decidual layer into the myometrium (accreta), within the myometrium (increta), or through the uterine serosa (percreta) [45]. Diagnosis of these clinical entities before delivery is crucial in decreasing the associated maternal morbidity and mortality resulting from maternal hemorrhage. This aids in the preoperative preparation of blood products, possible catheterization for uterine artery embolization, and possible cesarean hysterectomy. In cases of placenta previa, the number of prior cesarean sections affects the frequency of placenta accreta and its variants, occurring in up to 5% in the unscarred uterus with previa, 24% after one cesarean section with previa, and 67% after four or more cesarean sections with previa [25,46]. Because of the rise in cesarean deliveries, placenta accreta, occur-

ring 1 in 30,000 deliveries in 1950, now occurs in roughly 1 in 2500 deliveries [46]. The diagnosis of placenta accreta is made by the following criteria: (1) loss of the normal hypoechoic retroplacental myometrial zone, (2) thinning or disruption of the hyperechoic uterine serosabladder interface, and (3) presence of focal exophytic masses [Fig. 13]. Using the above criteria, Finberg and Williams [47] found that ultrasound had a sensitivity of 93% (depicting 14 of 15 cases of placenta accreta) and a specificity of 79% (negative ultrasound results in 15 of 19 patients who had no histopathologic evidence of placenta accreta) for the diagnosis of placenta accreta. In addition, the presence of lacunar vascular spaces within the placental parenchyma, also called the Swiss cheese appearance, has been found to be a risk factor for placenta accreta, even in the absence of other suspicious ultrasound findings [47]. Guy and coworkers [48] found lacunar vascular spaces to be correlated with placenta accreta in 14 of 16 patients who had placenta previa. Doppler evaluation and MR imaging have been purported to further aid in diagnosis; however, Levine and colleagues [49] found gray-scale ultrasonography adequate in cases of anterior placenta previa when the retroplacental clear space was normal (>2 mm), with Doppler or MRI providing little additional information. Those cases who had less than 2 mm retroplacental clear space were aided with further imaging, depending on the location of the placenta. Anterior placentas were best further visualized using power Doppler and transvaginal scanning with a partially full bladder, whereas

Sonography of the Lower Uterine Segment

315

Fig. 13. Placenta accreta at 26 weeks gestational age in a patient with two prior cesarean sections and placenta previa. (A) Transabdominal sagittal image shows a thickened placenta with cystic spaces. There is loss of the normal myometrium anteriorly. (B) Transvaginal transverse image shows absent myometrium (arrowhead) in a small region in the LUS. Note the varix (arrow) within the bladder. This varix is not in continuity with the placental vessels, and therefore should not be mistaken for a sign of placenta percreta. The delivery was by cesarean section at 37 weeks. Despite knowledge of the placenta accreta, the delivery was complicated by hemorrhage, requiring six units of packed red blood cells and two units of fresh frozen plasma.

MRI was helpful in one case who had a posterior uterine scar from a prior myomectomy and the placenta overlying the scar [49].

Leiomyomas in the lower uterine segment


The incidence of leiomyomas during pregnancy ranges from 0.09% to 3.9% [50]. With the increasing age of the obstetric population and the widespread use of sonography, leiomyomas are becoming more frequently detected in pregnancy. The impact of leiomyomas during pregnancy depends on their size, number, and location [Fig. 14]. Leiomyomas in the LUS increase the likelihood of malpresentation, cesarean birth, postpartum hemorrhage, and retained placenta [5052]. In addition, leiomyoma degeneration may occur during pregnancy and pre-

sent with abdominal pain, which must be differentiated from other potential etiologies of pain. Contrary to popular belief, the literature suggests that only 20% of myomas increase in size during pregnancy. Lev-Toaff and coworkers [51] found that in most instances, larger myomas, measuring 6 to 12 cm, became smaller in pregnancy. The ultrasound criteria for establishing the diagnosis of leiomyoma in pregnancy include spherical shape, distortion of myometrial contour, different acoustical structure than myometrium, speckled pattern of internal echoes increasing in density with increased ultrasound sensitivity, and no enhancement of echoes behind the mass [50,51]. In addition, color-flow Doppler can help differentiate a leiomyoma from a uterine contraction, in which a splaying of blood vessels around the leiomyoma mass usually can be seen [53,54]. Leiomyomas

316

Chie & Levine

may also occasionally mimic endometrial polyps, ovarian masses, or stool-filled large bowel [55]. When a leiomyoma is visualized in the LUS, care should be taken to visualize its relation to the cervix, to assess if vaginal birth will be possible [see Fig 14].

Ectopic pregnancy in the uterine scar


A scar ectopic pregnancy is defined as a gestation surrounded in its entirety by myometrium and fibrous tissue scar, separated from the endometrial cavity and not within fallopian tube. The true incidence is unknown because the literature primarily consists of case reports; however, the reports of such ectopic pregnancies have grown in the last few years, likely as a result of the increasing number of cesareans, widespread use of reproductive technology, and improved diagnosis and earlier detection with transvaginal ultrasonography [56,57]. Mechanisms of scar pregnancies include invasion of the myometrium through a microscopic dehiscent tract from

Fig. 14. Fibroid in LUS deviating cervix anteriorly. Sagittal transabdominal image shows a posterior fibroid (calipers) deviating the cervix anteriorly. Note the position of the internal os (arrow). A fibroid in this location will typically preclude vaginal delivery; however, at times the relationship of the fibroid to the cervix will change as the uterus enlarges, therefore follow-up sonography is required.

Fig. 15. Cesarean section scar ectopic pregnancies at 6 weeks gestational age after in vitro fertilization. (A) Transabdominal image through a full bladder shows a retroflexed uterus with two gestational sacs (arrows) in the region of the cesarean section scar. B, bladder. (B) Transvaginal image again shows the two gestational sacs (arrows). Note the sacs are located low in the uterus, separate from the endometrial cavity. The patient was treated with both methotrexate and with sonographically guided potassium chloride injection into the gestational sacs.

Sonography of the Lower Uterine Segment

317

prior uterine surgery (ie, curettage, cesarean, myomectomy, hysteroscopy, and manual removal of the placenta), or possibly in vitro fertilization (IVF) and embryo transfer in the absence of prior uterine surgery [58]. Differential diagnosis of a cesarean section ectopic pregnancy includes intrauterine pregnancy with placenta accreta, spontaneous abortion, and cervical pregnancy [58]. In this entity, a sagittal view of the uterus shows an empty uterine cavity, an empty cervical canal, and development of the sac in the anterior part of the uterine wall, with absence of normal myometrium between the bladder and the gestational sac [Fig. 15]. The gestational sac should be well-perfused on Doppler imaging [58]. Given the rarity of this clinical entity, no universal treatment guidelines exist. Although a single case report of expectant management to 35 weeks has been described, the patient experienced significant hemorrhage and required cesarean hysterectomy [59]. Thus, the proposed management strategies primarily have involved surgical or minimally invasive techniques. Surgical approaches typically involve laparotomy with hysterotomy or hysterectomy [60]. Minimally invasive techniques such as laparoscopic resection have been described [61]; however, if future fertility is a consideration, there is concern that the repair of the uterine wall in this fashion may be compromised. In addition, although curettage has been performed, it is generally not recommended, because it is associated with high risk of uterine perforation and hemorrhage. Nonsurgical approaches include systemic methotrexate, direct injection of methotrexate, aspiration of the sac with transvaginal ultrasound guidance, direct injection of hyperosmolar glucose, direct injection with potassium chloride, and uterine artery embolization (or some combination of the aforementioned treatments) [6169]. Subsequent pregnancy outcomes in patients who had a previous cesarean scar pregnancy and who did not undergo hysterectomy have been reported [70]. Although four of the seven pregnancies reported in the series were uneventful, one was complicated by uterine rupture at 38 weeks and another two required cesarean hysterectomies for placenta accreta at 32 weeks [70]. Thus such patients need to be counseled on the potential morbidity associated with any future pregnancy.

cesarean delivery is one of the most common surgical procedures, imaging of the LUS will play an increasing role in assisting in the management of patients in subsequent pregnancies.

References
[1] Donald I. Practical obstetric problems. London: Lloyd-Luke; 1964. p. 4212. [2] Creasy RK, Resnik R. Anatomic alterations in the reproductive tract. In: Creasy RK, Reznik R, Ia J, editors. Maternal-fetal medicine. Philadelphia: WB Saunders; 1999. p. 901. [3] Cunningham FG, Gant NF, Leveno KJ, et al. Williams obstetrics. 21st edition. New York: McGraw-Hill Companies; 2001. p. 168. [4] Schwalm H, Dubrauszky V. The structure of the musculature of the human uterusmuscles and connective tissue. Am J Obstet Gynecol 1966;94: 391404. [5] Degani S, Leibovitz Z, Shapiro I, et al. Myometrial thickness in pregnancy: longitudinal sonographic study. J Ultrasound Med 1998;17:6615. [6] Buhimschi CS, Buhimschi IA, Malinow AM, et al. Myometrial thickness during human labor and immediately post partum. Am J Obstet Gynecol 2003;188:5539. [7] Gotoh H, Masuzaki H, Yoshida A, et al. Predicting incomplete uterine rupture with vaginal sonography during the late second trimester in women with prior cesarean. Obstet Gynecol 2000; 95:596600. [8] Tanik A, Ustun C, Cil E, et al. Sonographic evaluation of the wall thickness of the lower uterine segment in patients with previous cesarean section. J Clin Ultrasound 1996;24:3557. [9] Rozenberg P, Goffinet F, Phillippe HJ, et al. Ultrasonographic measurement of lower uterine segment to assess risk of defects of scarred uterus. Lancet 1996;347:2814. [10] Asakura H, Nakai A, Ishikawa G, et al. Prediction of uterine dehiscence by measuring lower uterine segment thickness prior to the onset of labor: evaluation by transvaginal ultrasonography. J Nippon Med Sch 2000;67:3526. [11] Sen S, Malik S, Salhan S. Ultrasonographic evaluation of lower uterine segment thickness in patients of previous cesarean section. Int J Gynaecol Obstet 2004;87:2159. [12] Fukuda M, Fukuda K, Mochizuki M. Examination of previous caesarean section scars by ultrasound. Arch Gynecol Obstet 1988;243:2214. [13] Suzuki S, Sawa R, Yoneyama Y, et al. Preoperative diagnosis of dehiscence of the lower uterine segment in patients with a single previous Caesarean section. Aust N Z J Obstet Gynaecol 2000;40:4024. [14] Brill Y, Kingdom J, Thomas J, et al. The management of VBAC at term: a survey of Canadian obstetricians. J Obstet Gynaecol Can 2003;25: 30010. [15] Michaels WH, Thompson HO, Boutt A, et al.

Summary
The LUS is a dynamic portion of the uterus. Careful imaging is needed to avoid the pitfalls in diagnosis associated with overfilling of the bladder and LUS contractions. Further research is needed to elucidate how imaging of the LUS can be best used to screen for uterine dehiscence and rupture. Because

318

Chie & Levine

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

Ultrasound diagnosis of defects in the scarred lower uterine segment during pregnancy. Obstet Gynecol 1988;71:11220. Hamar BD, Levine D, Katz NL, et al. Expectant management of uterine dehiscence in the second trimester of pregnancy. Obstet Gynecol 2003; 102:113942. Jastrow N, Antonelli E, Boulvain M, et al. Interobserver and intraobserver variability in ultrasound measurment of lower uterine segment after a previous cesarean section. Ultrasound Obstet Gynecol 2003;22(Suppl):40. Hamrick-Turner JE, Cranston PE, Lantrip BS. Gravid uterine dehiscence: MR findings. Abdom Imaging 1995;20:4868. Hebisch G, Kirkinen P, Haldemann R, et al. Vergleichende untersuchung am unteren uterinsegment nach sectio caesarea mittels ultraschall und magnetresonanztomographie [Comparative study of the lower uterine segment after Cesarean section using ultrasound and magnetic resonance tomography]. Ultraschall Med 1994;15:1126. Hasbargen U, Summerer-Moustaki M, Hillemanns P, et al. Uterine dehiscence in a nullipara, diagnosed by MRI, following use of unipolar electrocautery during laparoscopic myomectomy: case report. Hum Reprod 2002;17:21802. Confino E, Mayden KL, Giglia RV, et al. Pitfalls in sonographic imaging of the incompetent uterine cervix. Acta Obstet Gynecol Scand 1986; 65:5937. Karis JP, Hertzberg BS, Bowie JD. Sonographic diagnosis of premature cervical dilatation. Potential pitfall due to lower uterine segment contractions. J Ultrasound Med 1991;10:837. Yost NP, Bloom SL, Twickler DM, et al. Pitfalls in ultrasonic cervical length measurement for predicting preterm birth. Obstet Gynecol 1999; 93:5106. Gilbert WM, Nesbitt TS, Danielsen B. Childbearing beyond age 40: pregnancy outcome in 24,032 cases. Obstet Gynecol 1999;93:914. Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior cesarean section. Obstet Gynecol 1985;66:8992. Rose GL, Chapman MG. Aetiological factors in placenta praeviaa case controlled study. Br J Obstet Gynaecol 1986;93:5868. Meyer MB, Tonascia JA. Maternal smoking, pregnancy complications, and perinatal mortality. Am J Obstet Gynecol 1977;128:494502. Ananth CV, Demissie K, Smulian JC, et al. Placenta previa in singleton and twin births in the United States, 1989 through 1998: a comparison of risk factor profiles and associated conditions. Am J Obstet Gynecol 2003;188:27581. Cotton DB, Read JA, Paul RH, et al. The conservative aggressive management of placenta previa. Am J Obstet Gynecol 1980;137:68795. Smith RS, Lauria MR, Comstock CH, et al. Transvaginal ultrasonography for all placentas that appear to be low-lying or over the internal

[31]

[32]

[33]

[34]

[35]

[36]

[37] [38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

cervical os. Ultrasound Obstet Gynecol 1997;9: 224. Hertzberg BS, Bowie JD, Carroll BA, et al. Diagnosis of placenta previa during the third trimester: role of transperineal sonography. AJR Am J Roentgenol 1992;159:837. Timor-Tritsch IE, Monteagudo A. Diagnosis of placenta previa by transvaginal sonography. Ann Med 1993;25:27983. Leerentveld RA, Gilberts EC, Arnold MJ, et al. Accuracy and safety of transvaginal sonographic placental localization. Obstet Gynecol 1990;76: 75962. Timor-Tritsch IE, Yunis RA. Confirming the safety of transvaginal sonography in patients suspected of placenta previa. Obstet Gynecol 1993; 81:7424. Oppenheimer LW, Farine D, Ritchie JW, et al. What is a low-lying placenta? Am J Obstet Gynecol 1991;165:10368. Taipale P, Hiilesmaa V, Ylostalo P. Transvaginal ultrasonography at 1823 weeks in predicting placenta previa at delivery. Ultrasound Obstet Gynecol 1998;12:4225. Lavery JP. Placenta previa. Clin Obstet Gynecol 1990;33:41421. Hill LM, DiNofrio DM, Chenevey P. Transvaginal sonographic evaluation of first-trimester placenta previa. Ultrasound Obstet Gynecol 1995;5: 3013. Taipale P, Hiilesmaa V, Ylostalo P. Diagnosis of placenta previa by transvaginal sonographic screening at 1216 weeks in a nonselected population. Obstet Gynecol 1997;89:3647. Dashe JS, McIntire DD, Ramus RM, et al. Persistence of placenta previa according to gestational age at ultrasound detection. Obstet Gynecol 2002;99:6927. Lauria MR, Smith RS, Treadwell MC, et al. The use of second-trimester transvaginal sonography to predict placenta previa. Ultrasound Obstet Gynecol 1996;8:33740. Ghourab S. Third-trimester transvaginal ultrasonography in placenta previa: does the shape of the lower placental edge predict clinical outcome? Ultrasound Obstet Gynecol 2001;18: 1038. Oppenheimer L, Holmes P, Simpson N, et al. Diagnosis of low-lying placenta: can migration in the third trimester predict outcome? Ultrasound Obstet Gynecol 2001;18:1002. Townsend RR, Laing FC, Nyberg DA, et al. Technical factors responsible for placental migration: sonographic assessment. Radiology 1986;160:1058. Sumawong V, Nondasuta A, Thanapath S, et al. Placenta accreta. A review of the literature and a summary of 10 cases. Obstet Gynecol 1966;27: 5116. Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for placenta previa-placenta accreta. Am J Obstet Gynecol 1997;177:2104.

Sonography of the Lower Uterine Segment

319

[47] Finberg HJ, Williams JW. Placenta accreta: prospective sonographic diagnosis in patients with placenta previa and prior cesarean section. J Ultrasound Med 1992;11:33343. [48] Guy GP, Peisner DB, Timor-Tritsch IE. Ultrasonographic evaluation of uteroplacental blood flow patterns of abnormally located and adherent placentas. Am J Obstet Gynecol 1990;163: 7237. [49] Levine D, Hulka CA, Ludmir J, et al. Placenta accreta: evaluation with color Doppler US, power Doppler US, and MR imaging. Radiology 1997; 205:7736. [50] Phelan JP. Myomas and pregnancy. Obstet Gynecol Clin North Am 1995;22:8015. [51] Lev-Toaff AS, Coleman BG, Arger PH, et al. Leiomyomas in pregnancy: sonographic study. Radiology 1987;164:37580. [52] Katz VL, Dotters DJ, Droegemeuller W. Complications of uterine leiomyomas in pregnancy. Obstet Gynecol 1989;73:5936. [53] Trampe BS, Pryde PG, Stewart KS, et al. Color Doppler ultrasonography for distinguishing myomas from uterine contractions in pregnancy. J Reprod Med 2001;46:7914. [54] Kessler A, Mitchell DG, Kuhlman K, et al. Myoma vs. contraction in pregnancy: differentiation with color Doppler imaging. J Clin Ultrasound 1993;21:2414. [55] Richenberg J, Cooperberg P. Ultrasound of the uterus. In: Callen PW, editor. Ultrasonography in obstetrics and gynecology. Philadelphia: WB Sanders Co.; 2000. p. 82935. [56] Maymon R, Halperin R, Mendlovic S, et al. Ectopic pregnancies in caesarean section scars: the 8 year experience of one medical centre. Hum Reprod 2004;19:27884. [57] Jurkovic D, Hillaby K, Woelfer B, et al. Firsttrimester diagnosis and management of pregnancies implanted into the lower uterine segment cesarean section scar. Ultrasound Obstet Gynecol 2003;21:2207. [58] Fylstra DL. Ectopic pregnancy within a cesarean scar: a review. Obstet Gynecol Surv 2002;57: 53743. [59] Herman A, Weinraub Z, Avrech O, et al. Follow up and outcome of isthmic pregnancy located in

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

a previous caesarean section scar. Br J Obstet Gynaecol 1995;102:83941. Larsen JV, Solomon MH. Pregnancy in a uterine scar sacculusan unusual cause of postabortal haemorrhage. A case report. S Afr Med J 1978; 53:1423. Lee CL, Wang CJ, Chao A, et al. Laparoscopic management of an ectopic pregnancy in a previous caesarean section scar. Hum Reprod 1999; 14:12346. Godin PA, Bassil S, Donnez J. An ectopic pregnancy developing in a previous caesarian section scar [erratum appears in Fertil Steril 1997; 68(1):187]. Fertil Steril 1997;67:398400. Roberts CL, Weston MJ. Bilateral massive ovarian edema: a case report. Ultrasound Obstet Gynecol 1998;11:657. Doubilet PM, Benson CB, Frates MC, et al. Sonographically guided minimally invasive treatment of unusual ectopic pregnancies. J Ultrasound Med 2004;23:35970. Chou MM, Hwang JI, Tseng JJ, et al. Cesarean scar pregnancy: quantitative assessment of uterine neovascularization with 3-dimensional color power Doppler imaging and successful treatment with uterine artery embolization. Am J Obstet Gynecol 2004;190:8668. Ghezzi F, Lagana D, Franchi M, et al. Conservative treatment by chemotherapy and uterine arteries embolization of a cesarean scar pregnancy. Eur J Obstet Gynecol Reprod Biol 2002; 103:8891. Seow KM, Huang LW, Lin YH, et al. Cesarean scar pregnancy: issues in management. Ultrasound Obstet Gynecol 2004;23:24753. Nawroth F, Foth D, Wilhelm L, et al. Conservative treatment of ectopic pregnancy in a cesarean section scar with methotrexate: a case report. Eur J Obstet Gynecol Reprod Biol 2001;99:1357. Ravhon A, Ben-Chetrit A, Rabinowitz R, et al. Successful methotrexate treatment of a viable pregnancy within a thin uterine scar. Br J Obstet Gynaecol 1997;104:6289. Seow KM, Hwang JL, Tsai YL, et al. Subsequent pregnancy outcome after conservative treatment of a previous cesarean scar pregnancy. Acta Obstet Gynecol Scand 2004;83:116772.

321

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 321334

3-Dimensional Obstetric Ultrasound: Tips of the Trade


Maryam Tarsa, MDa, Dolores H. Pretorius, Deborah DAgostini, RDMSb
& & & & & & &

MD

b,*

Acquisition of volumes Mechanical positioning Display modalities Scanning Tips Archiving 3-D US and 4-D US data Transvaginal 3-D sonography Usefulness of 3-D and 4-D US in obstetrics

& & & & & &

Good normal face Cleft lip and palate Extremities and skeleton Pitfalls/limitations/artifacts Summary References

There is no question that the practice of perinatal medicine has changed due to advances in radiology, particularly ultrasound. Prenatal diagnosis of fetal abnormalities has become one of the most important aspects of maternal fetal medicine. Patients and referring physicians heavily rely on correct diagnosis of potential congenital abnormalities. With a diagnosis in hand, parents have the option to meet with the pediatric subspecialty team and are prepared for possible events after the birth of their child. Three-dimensional ultrasound (3-D US) has revolutionized the field of imaging with its ability to use multiple planes in which the area of interest may be displayed. Given its ability to examine complex anatomic structures by using surface analysis and volumetric measures, it has become one of the major modalities used in assessing fetal abnormalities. The role of 3-D US in obstetrics is evolving rapidly. There are many advantages in using this modality including more accurate diagnosis about fetal abnormalities, improved comprehension of fetal anatomy by families, and
a

improved maternal fetal bonding [13]. The importance of reassurance provided by 3-D US in the absence of any fetal abnormalities especially in high-risk families should also be considered [4]. This article briefly reviews some applications of 3D-US in obstetrics. A summary of image acquisition and display including some how to techniques are provided.

Acquisition of volumes
The key factors in obtaining a 3-D image are USs positioning flexibility and data acquisition speed [5,6]. The quality of images obtained depends on acquisition speed. Fast speeds are required for moving structures such as a fetal limb. The challenge in producing a good quality 3-D image stems from the difficulty locating the position of the 2-D US image within the volume of investigation. The current 3-D transducers primarily use a position-sensing device. The data-regarding position may be obtained using electromagnetic position sensor, rotation device, or stepping motors in the scan head

Department of Reproductive Medicine, University of California, San Diego Medical Center, 200 West Arbor Drive, San Diego, CA 92103, USA b Department of Radiology, University of California, San Diego Medical Center, 200 West Arbor Drive, San Diego, CA 92103, USA * Corresponding author. Thornton Hospital, 9300 Campus Point Drive, 7756, La Jolla, CA 92037. E-mail address: dpretorius@ucsd.edu (D.H. Pretorius).
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.004

ultrasound.theclinics.com

322

Tarsa et al

[7]. Most commercially available equipment being used today acquires 3-D US data using a mechanically driven transducer where the scan head moves back and forth. This article discusses techniques using this type of data acquisition. Volumes can be acquired using hand-held transducers that are moved across the abdomen in a smooth sweep, but assumptions must be made regarding the distance between images and thus measurements are generally not accurate.

Mechanical positioning
In general, multiple digitally recorded 2-D US images are recorded and then reconstructed immediately into a 3-D US volume that can be displayed in different orientations. The obvious problem with reconstructing images is geometric inaccuracies. To avoid this issue, the relative position and angulation of each image must be known. Therefore, every time the transducer moves, the images are obtained at a predefined angle and distance interval. Thanks to advances in computer technology, these planar images are stored in the US equipment computer system. 3-D US is a display of the acquired static volumeno motion can be seen,

such as bowel peristalsis or fetal movement. Fourdimensional ultrasound (4-D US) adds time to the information acquired and allows for continuously updated volume display allowing for motion of the fetus to be observed. This is also referred to as real time 3-D US or live 3-D US. There are three different mechanisms used to produce 3-D US data: linear, tilt, and rotational motion. In linear scanning the images are arranged parallel to each other. Tilt scanning, most commonly used in obstetrics, has images arranged in a fanlike shape. The disadvantage of this method is loss of resolution with increasing depth. The resolution will be best in the focal zone and attention must be made to identify the area of interest during 2-D setup scanning to optimize the focal zone positioning arrows. Rotation scanning is used in some endovaginal tranducers for first trimester obstetrics and gynecology, arranging images in propeller like geometry.

Display modalities
Multiplanar and volume rendering are the two primary display modalities. In multiplanar or orthogonal mode, the stored volume is displayed in three

Fig. 1. Multiplanar image of 12-week embryo. Marker dot is located on the fetal stomach. Top left image is coronal, bottom left is sagittal, and top right is transverse through embryo.

3-Dimensional Obstetric Ultrasound

323

planes perpendicular to each other. Where the three planes intersect, a marker dot is placed. Using this dot, the same structure can be located and displayed on all three different planes [Fig. 1]. The marker dot is invaluable in following structures that traverse the volume, such as hydrosalpinx and cerebral ventricles. It is also widely used in gynecology to display the endometrium in a coronal projection. When evaluating very fine structures such as the primary palate, it is important to acquire volumes in the optimal plane of resolution (ie, axial plane for the palate) to examine it in the multiplanar display. The volume can be evaluated by scrolling through each plane in a parallel fashion, similar to scrolling through axial CT images. We often say this is like going through the volume like a deck of cards. The volume can be rotated in the three planes, X, Y, and Z, to optimize visualization of anatomic structures. It is ideal to display anatomy in a standard orientation (eg, with the fetal face upright; frontal or coronal), in a profile view (sagittal), and transversely (axial). We have found it easiest to rotate the image that is recognizable in the Z plane first, as it rotates the volume like a record player and the action is more readily recognizable. Determining how to rotate a volume to see the desirable information is clearly the crux of the 3-D US beginner. When a specific region needs to be seen, it is acquired in a recognizable plane, the marker dot placed on it, and then rotated to a standard orientation. The volume is then evaluated by scrolling through the parallel planes to assess anatomy and size, if necessary. It is important to realize that the right and left side of the structure are not identified on the volume. The operator must label right and left based on 2-D US informa-

Fig. 3. Maximum intensity-rendered image of the thorax and spine at 19 weeks. The posterior ribs and spine are visualized. (Courtesy of Philips Medical Systems, Bothell, WA; with permission.)

Fig. 2. Surface-rendered image of a normal face at 31 weeks. (Courtesy of Philips Medical Systems, Bothell, WA; with permission.)

tion. By rotating the volume 180, the right and left will be reversed. The rendering display is used to demonstrate information within the entire volume. There are multiple display modes emphasizing different structures. In surface rendering, a sharp interface between two structures must be present, such as the fetal skin surface and the amniotic fluid or blood in the heart and heart walls. Surface-rendered images of the fetal face are impressive and very realistic for both families and medical care providers [Fig. 2]. The 3-D US stored volume can be rotated in the three cardinal directions. In learning how to rotate the image it is important to understand which knob to turn to move the image in the desired direction. In general, rotation of the rendered image using the X axis moves the face up and down (similar to saying yes), using the Y axis moves the face side to side (similar to saying no), and using the Z axis moves the face around (similar to a record player). It is important to understand how to do these rotations on the equipment you have in laboratory because different equipment will rotate differently. Sometimes the rendered image is linked to the multiplanar images and sometimes it is not. Sometimes the X, Y, and Z are only connected to the rendered image and other times to the multiplanar images. As you adjust to your own equipment, you will be able to predict which way the knobs will move the image although at first, it will be necessary to rotate all three knobs to find the desired rotation. Different light mode and filtering levels can change the quality of the image. Each US vendor uses different terms to identify changes in rendering

324

Tarsa et al

Fig. 4. Acquiring 4-D volumes using the rendering line to set up the acquisition. (A) Acquisition plane of fetal face. Notice that no structures are present between the face and the rendering line (dotted line). (B) Rendered image of face resulting from sweep in (A). Notice that the 3-D sweep is made with the face in a profile or sagittal plane while the rendered image (B) is frontal, or 90 from the acquisition plane.

mode: surface, transparency, x-ray, opacity, threshold, maximum, minimum, smoothness, and so forth. Each of these modes emphases certain structures, such as surface, bones, and vessels. The modes can often be used in combination to optimize visualization. For example, the skeleton is

often displayed with the maximum intensity mode [Fig. 3] or the x-ray mode. Using this technique, soft tissue structures potentially can be eliminated. The threshold knob is perhaps the most important parameter to optimize in displaying surface features. When the threshold is turned

Fig. 5. Gestational sac in a septate uterus. On 3-D US, the sac was seen in the right horn of a septate uterus (arrow ). Top left image is the uterus transverse, and top right uterus is sagittal, and bottom left image is coronal showing two horns and normal myometrium in the fundus.

3-Dimensional Obstetric Ultrasound

325

up, the low-level echoes are eliminated from the image. If, however, the threshold is turned up too high, the image disappears before your eyes.

Scanning Tips
First, set up your image parameters as you would generally for a 2-D US except use a little more contrast in the 2-D images. A poor 2-D image will result in a poor 3-D image, whereas good 2-D images will likely result in good 3-D images. After optimizing your image quality, scan over the area of interest as per a standard 2-D US. Second, determine which plane is most important anatomically for the region of interest. For example, the best views of the palate are obtained in axial plane. It is in this acquisition plane that you have the best resolution in the volume. Scan over the region in a linear fashion, as the 3-D transducer will do automatically and position the transducer directly over the center of the region of interest since the volume will be swept out equally on either side of the

center of the transducer. Third, the optimal image is obtained when there is no motion. Therefore, only the motion from the transducer position changes is desired. This is best obtained by asking the patient to hold her breath during the acquisition, even when performing endovaginal scans. Fourth, hold the transducer steady and push the button to acquire a volume. Fifth, review the volume quickly by scrolling through one of the planes to make sure the images collected are from the desired region of interest and that there was minimal motion. If using 4-D US, similar scanning tips are used. The transducer automatically acquires continuous volumes by moving the acoustic array within the probe. When setting up the volume (ie, scanning in 2-D), it is important to find an area with fluid adjacent to the fetus. Turn on 4-D and place the rendering line adjacent to the structure in the amniotic fluid so that there are no intervening structures [Fig. 4]. When the desired rendered image is seen then it can be displayed on the monitor as a

Fig. 6. Standard orientation of the fetal face. Upper left image in profile. Upper right image is symmetrical orbits. Lower left image is coronal view. Lower right image is rendered image of fetal face. The marker dot is on the fetal nose.

326

Tarsa et al

single image but as the fetus moves viewing the 2-D localizing image will again be necessary. Generally we use only a 2-image display to localize for 4-D imaging as we find that it is difficult for our brains to evaluate more than two moving images simultaneously. However, if we are having difficulty in figuring out what structure is obstructing our view, we will occasionally turn on a four-image display to assist in understanding the orientation issues. Generally, the operator holds the transducer still on the maternal abdomen and allows the fetus to move rather than moving the transducer around the abdomen. Occasionally, very minor, slow movements of the transducer can be helpful. One advantage of 4-D US scanning is that the volume can be steered while scanning the patient. This is not possible with 3-D US because the volume is stationary and is viewed after the acquisition. But in 4-D US, the volume can be viewed from any angle while the scanning is being performed. This means that when one side of the face is up against the placenta, the volume can be rotated (using the X, Y, and Z knobs) to view the fetus from a more optimal angle. This is different from rotating the rendered image 90,180, or

270 to see the fetus in a more anatomic position, a technique that is available on equipment today. The volume steering technique is a true advantage of 4-D US scanning, and we believe a major assistance to evaluating movement of fetus arms and legs. In the past, this has been called beam steering but truthfully, the beam angle is staying the same and the volume is being rotated to view it optimally.

Archiving 3-D US and 4-D US data


Volume data are different than 2-D image data. It is important for the sonographer/physician to realize that saving data for review at a future time requires an understanding of the types of archiving possibilities. When a single image is archived then we save it as any 2-D image. However, this new technology allows us many new options. If the entire volume is needed for future manipulation, then the volume must be saved. If a cine clip of the 3-D volume is needed, then a video clip must be saved. If a cine loop from the 2-D image data is desired then a different video clip mechanism is in play to save it. Multiple volumes from a 4-D acqui-

Fig. 7. Display of the upper lip in the multiplanar view at 29 weeks. Face is tipped posteriorly slightly to show the fetal lip (arrow) optimally in the coronal plane. Marker dot is on the upper lip in all 3D multiplanar images.

3-Dimensional Obstetric Ultrasound

327

sition can be saved or a single volume from the 4-D acquisition can be saved. Data can be archived on the equipment itself and often downloaded onto a CD-rom or other storage device for review on a workstation with 3-D capabilities. Additionally, jpg images and video clips can be downloaded onto a CD-rom, which can be given to the patient for their review with family and friends. Each manufacturer uses their own terminology to identify these different archiving capabilities, and the operator must learn what is available on the machine they are using. The size of the images, volumes, and video clips will vary significantly depending on the transducer used, the size of the region of interest box, and the equipment vendor.

uterus can help differentiate a cornual ectopic pregnancy from a septate uterus with a gestational sac in one horn of the uterus [Fig. 5]. Detailed surface and multiplanar images of the embryo [8,9] and the fetal face can be obtained as early as 9 weeks gestation using transvaginal 3-D US [10]. Fetal abnormalities are now being detected with transvaginal 3-D US [10,11] including the brain [12], spine [13], face, and skeleton. More recently, evaluation of cervical volumes in pregnant patients has been undertaken [14]. This may prove useful in the evaluation of cervical incompetence.

Usefulness of 3-D and 4-D US in obstetrics Transvaginal 3-D sonography


3-D transvaginal scanning has been used not only for assessment of uterine anatomy and detection of congenital anomalies but also for location of ectopic pregnancies, particularly cornual (interstitial) and cervical ectopic. The coronal plane of the Multiple studies have compared the performance of 3-D and 2-D US in detecting fetal anomalies prenatally. A review of the literature has shown a 51% to 64% diagnostic advantage of 3-D US [15,16]. 3-D US has been shown to be helpful in assessing anatomy and pathology, evaluating the extent of disease, detecting a variety of abnormali-

Fig. 8. Narrowed region of interest to display fetal face. The boundaries of the region of interest box are narrowed to eliminate echoes in the volume that do not contribute to the surface of the face.

328

Tarsa et al

ties (a few not seen on 2-D US), and counseling patients. A few of these will be mentioned later.

Cleft lip and palate


Cleft lip and palate is one of the most common congenital abnormalities and is highly associated with other congenital abnormalities [19]. 3-D US is a useful tool in correct identification and determination of the extent of the facial clefting. Accuracy of diagnosis of facial clefting is higher using 3-D US when compared with 2-D US [20,21]. A clear view of a fetal face with anomalies has a tremendous role in making decisions regarding continuation of pregnancy, bonding, and conceptualization of the postnatal surgical care of the newborn. Processing of volumes of fetal face using 3-D US has been extensively studied yet we continue to learn new ways to obtain optimal pictures. It is important to standardize the orientation so that the face is viewed symmetrically. This is done by acquiring, in a standard plane if possible, either sagittal, axial, or coronal. After the volume has been acquired the face should be rotated so that the orbits are symmetrical. This is performed easily if the marker dot is positioned on the center of the nose and the volume is rotated; this often requires a

Good normal face


3-D US is a valuable tool in examining the fetal face. Multiple abnormalities including cleft lip/ palate, micrognathia, dysplastic ear, and midface hypoplaisa have been examined using this technique [17]. Enhanced images of fetal face may be obtained if there is some amniotic fluid adjacent to the fetal face. This is achieved by changing maternal position or tapping on the maternal abdomen near and distant from the transducer. A gentle pushing of the fetus or fluid by the palm of the hand can be helpful. The best surface-rendered face images are often obtained sagittally or just obliquely. Using 2-D US technique, the facial profile should be obtained first. The region of interest boundaries are then placed around the entire face and the 3-D US volume swept. Both planar and rendered images are useful in assessing the fetal facial features [18]. Rendered images are best obtained between 22 and 30 weeks.

Fig. 9. Multiplanar view of normal primary palate (anterior alveolar ridge). The face is in profile in upper left. The marker dot is on the palate. The green reference line overlying the upper lip on the rendered image localizes the palate on the upper right-hand axial image.

3-Dimensional Obstetric Ultrasound

329

Fig. 10. Multiplanar and rendered view of cleft in primary palate and lip (anterior alveolar ridge). The left upper is an image of the profile; the right upper is cleft of primary palate (short arrow) in an axial view; bottom left image is coronal view of cleft lip (long arrow), and bottom right is a rendered image of cleft lip. The green line represents the level of cleft palate on the coronal view.

Fig. 11. Rendered view of normal primary palate (anterior alveolar ridge) and cleft palate. (A) Normal primary palate. (B) Cleft of primary palate (arrow). These images were obtained by increasing the threshold on the surfacerendered image.

330

Tarsa et al

Fig. 12. Rendered view of scoliosis in 22-week fetus. Scoliosis is at T5 level.

3-D provides additional clues from the perpendicular views and from then rendered view to confirm that the maxilla or anterior alveolar ridge are being evaluated. A reference line can be seen crossing the rendered view at the level of the upper lip, identifying the appropriate axial image of the palate [Figs. 9 and 10]. This can also be confirmed using the marker dot on the multiplanar images. After the face is standardized, the axial plane is scrolled through mm by mm to assess each tooth bud in the anterior alveolar ridge (primary palate or hard palate). The palate can also be examined on the rendered view. This can be done by looking at the face directly and increasing the threshold until only the bones of the face are visualized [Fig. 11]. If the threshold is increased too high, then pseudo clefts can be seen. Interestingly, when skeletal parameters are used, we do not obtain optimal images to display the cleft. An additional pitfall occurs when shadowing from other structures lead to a break in the maxilla. It is, therefore, very important to confirm any suspected cleft palate seen on the rendered views with the multiplanar view, particularly the plane of acquisition where shadows are most obvious. Although preliminary data suggest that the soft palate or secondary palate can be evaluated with 3-D US and 4-D US [2224], we have not been able to consistently evaluate this

slight rotation in X, Y, and Z planes. The profile should be lined up so that the face is pointing upward in the upper-left image [Fig. 6]. The best multiplanar views of the lip are obtained in coronal plane with face tipped slightly posteriorly [Fig. 7]. For optimal-rendered images of the face, the region of interest boundaries can be narrowed, eliminating much of the echoes in the volume that do not contribute to the surface information [Fig. 8]. This technique can be used with both 3-D and 4-D imaging and are quite helpful as the gestation progresses, often allowing quite good images in the late third trimester when only a sliver of fluid is present in front of the face. The palate can be evaluated in both multiplanar and rendered views. The multiplanar method is more familiar to most sonographers because the axial view is routinely obtained during 2-D scanning. The face should be acquired in the axial plane so that there is minimal shadowing from adjacent structures (the upper half of the palate or extremities) and that there is optimal resolution in the axial plane. 3-D offers additional information in some cases because the face position can be standardized. On 2-D imaging, it can be difficult to differentiate the palate from the mandible and although we have rules for making the distinction, in clinical practice we have found that sometimes we incorrectly identify the maxilla. For this reason,

Fig. 13. Rendered image of club foot. The leg is upright (short arrows) and the foot is turned inward (short arrows).

3-Dimensional Obstetric Ultrasound

331

Fig. 14. Normal hand (A) and a picture of bilateral polydactyly (B). Arrows point to extra digits.

Fig. 15. Artifact from an intrauterine device (IUD) simulating an IUD within the sac. The white echo arising from the IUD is a comet tail from reverberations off the IUD and then a shadow as it goes through the uterus behind the gestational sac. The marker dot is on the comet tail in all three planes. Upper left image is the acquisition plane, and the comet tail is well recognized emanating from the echogenic IUD (arrows), which is outside of the gestational sac. In both the upper right image and lower left image, the comet tail is seen as a curvilinear structure which could be misinterpreted as IUD. In reality, it is only an artifact.

332

Tarsa et al

Fig. 16. Rendering artifact creasting the false image of hole in the head. This occurs due to the boundaries of the rendering box passing through the cranium.

region, and we believe that further experience is needed to evaluate it confidently.

Extremities and skeleton


Using surface rendering, maximum intensity and x-ray mode, fetal bony structures and extremities can be visualized clearly. The surface-rendered images are often helpful in assessing motion of the ankles, opening and closing of the hands, and movement of the arms and legs. They are also helpful in counting digits of the hands. The maximum intensity mode or increasing the threshold on the surfacerendering mode (described above to evaluate the palate) are useful in evaluating the bones (ribs, long bones, clavicles, spine), which can be seen

with reliability and reproducibility [25,26]. Using these techniques is it possible to evaluate deformities of fetal spine [Fig. 12], thorax, or long bones. We have found 3-D to be particularly helpful in evaluating for club foot, both because the ankle can be displayed and evaluated in a standard orientation (without movement on 3-D US) [Fig. 13] and movement of the ankle can be assessed using 4-D US. Some investigators have suggested that 3-D US is the tool of choice in evaluating the hands and feet due to capability of rotating the volumes [Fig. 14] [26]. However, fast movement of hands and feet, positioning of the extremities adjacent to the uterine wall, or some other suboptimal position can potentially be an obstacle in their evaluation. Although 4-D US may prove to be more useful, the primary stumbling blocks remain the same.

Fig. 17. Motion artifact presented as conjoined twins at 24 weeks (A). Actual image of face is shown in (B).

3-Dimensional Obstetric Ultrasound

333

Pitfalls/limitations/artifacts
Despite advantages of 3-D US mentioned previously, there are still some problems in the 3-D technique that should be taken into account: 1. Similar to learning any new imaging modality, the learning curve is variable in different individuals. The examiner must adjust to examining the stored volume to obtain the desired view. Also, common means of orientation such as left, right, cranial, and caudal are not necessarily related to the position on the screen. The sidedness should be directly correlated with the image of the fetus because the volume can be rotated in all directions. 2. During acquisition of the volume, the probe as well as the fetus should remain stationary to avoid movement artifacts. This may take as long as 4 seconds depending on the size of volume acquired. 3. Quality of surface-rendered images is as good as the quality of original data. The examiner needs to have knowledge on how to obtain the best image by rotating and eliminating structures that disturb the surface views (for example, umbilical cord or extremities obstructing the facial views). For this reason, surface reconstruction in oligohydramnios cases is nearly impossible. 4. Artifacts that simulate pathology are common in 3-D US and 4-D US as they are in 2-D US imaging. The shadows that we routinely see in 2-D US emanating from calcified structures (bones) or other dense objects are easily recognized. The shadows seen in 3-D US are easily recognized in the plane of acquisition but are not obvious in the reconstructed planes. These can lead to misinterpretation of the study [Fig. 15]. Region of interest boundaries can lead to apparent defects such as a hole in the head or a missing limb [Fig. 16]. Motion can lead to an apparent image of conjoined twins [Fig. 17].

is a promising tool for the evaluation of the fetus. Well-designed studies are needed to show its clinical effectiveness is daily practice.

References
[1] Merz E, Bahlman F, Weber G. Volume scanning in the evaluation of fetal malformations: a new dimension in prenatal diagnosis. Ultrasound Obstet Gynecol 1995;5:2227. [2] Maier B, Steiner H, Wilnerroither H, et al. The psychological impact of three-dimensional fetal imaging on the fetomaternal relationship. In: Baba K, Jurkovic D, editors. Three dimensional ultrasound in obstetrics and gynecology. New York: Parthenon; 1997. p. 6774. [3] Ji E, Pretorius DH, Newton R, et al. Effects of ultrasound on maternal-fetal bonding: a comparison of 2-dimensional vs. 3-dimensional imaging. J Ultrasound Obstet Gynecol 2005;25:4737. [4] Pretorius DH, House H, Nelson TR, et al. Evaluation of normal and abnormal lips in fetuses: comparison between three- and two-dimensional sonography. AJR Am J Roentgenol 1995;165(5): 12337. [5] Baba K, Satoh K, Sakamato S, et al. Development of an ultrasonic system for three-dimensional reconstruction of the fetus. J Perinatol Med 1989; 17:1924. [6] Nelson T, Pretorius D. Interactive acquisition and visualization of sonographic volume data. Int J Imaging Syst Technol 1997;8:2637. [7] Nelson T, Downey DB, Pretorius DH, et al. Three dimensional ultrasound. Philadephia: Lippincott, Williams & Wilkins; 1999. [8] Benoit B, Kuysesic S, Kurjak A, et al. 3-dimensional ultrasound in imaging of multifetal pregnancy. Ultrasound Rev Obstet Gynecol 2001;1(4):3016. [9] Hull A, James G, Salerno C, et al. Three dimensional ultrasonography and assessment of the first trimester fetus. JUM 2001;20:28793. [10] Merz E, Weber G, Bahlman F, et al. Application of transvaginal and abdominal three-dimensional ultrasound for the detection or exclusion of malformations of the fetal face. Ultrasound Obstet Gynecol 1997;9(4):23743. [11] Jurkovic D, Geipel A, Grubboek K, et al. Threedimensional ultrasound for the assessment of uterine anatomy and detection of congenital anomalies: a comparison with hysterosalpingography and sonography. Ultrasound Obstet Gynecol 1995;5(4):233. [12] Benoit B, Hafner T, Bekavac I, et al. 3-dimensional sonoembryology. Ultrasound Rev Obstet Gynecol 2001;1(2):12837. [13] Blaas H, Eik-Nes S, Isaksen C. The detection of spina bifida before 10 gestational weeks using two and three dimensional ultrasound. Ultrasound Obstet Gynecol 2000;16:259. [14] Roavas L, Sladkevicius P, Strobel E, et al. Intraobserver and interobserver reproductibility of

Summary
Clinical data on the application of 3-D US and its role as a screening modality in obstetrics and as an adjunct to 2-D US in the diagnosis of fetal anomalies are limited. Although investigators have been using 3-D US for more than 12 years in their clinical practices, our knowledge base is still at the fundamentals level. As this new technology is dispersed widely and is used by a vast number of clinical practitioners, we are sure that advancements will be forthcoming. In summary, 3-D US

334

Tarsa et al

[15]

[16]

[17]

[18]

[19]

[20]

three-dimensional gray-scale and power Doppler ultrasound examinations of the cervix in pregnant women. Ultrasound Obstet Gynecol 2005; 26(2):1327. Dyson R, Pretorius DH, Budorick NE, et al. Three dimensional ultrasound in the evaluation of fetal anomalies. Ultrasound Obstet Gynecol 2000;16: 3218. Timor-Tritsch I, Platt L. Three-dimensional ultrasound experience in obstetrics. Curr Opin Obstet Gynecol 2000;14:56975. Pretorius D, Richards RD, Budorick NE, et al. Three-dimensional ultrasound in the evaluation of fetal anomalies. Radiology 1997;205(Suppl):245. Devonald K, Elwood DA, Griffiths KA, et al. Volume imaging: three-dimensional appreciation of the fetal head and face. J Ultrasound Med 1995; 14:91925. Jones M. Etiology of facial cleft: prospective evaluation of 428 patients. Cleft Palate J 1988; 25:1620. Johnson D, Pretorius DH, Budorick NE, et al. Fetal lip and primary palate: three dimensional versus two dimensional US. Radiology 2000;217: 2369.

[21] Chmait R, Pretorius DH, Jones M, et al. Prenatal evaluation of facial clefts with two dimensional and adjunctive three dimensional ultrasonography: a prospective trial. Am J Obstet Gynecol 2002;187(4):9469. [22] Aubry M, Aubry J. Prenatal diagnosis of cleft palate: contribution of color doppler ultrasound. Ultrasound Obstet Gynecol 1992;2(3):2214. [23] Campbell S, Lees C, Moscoso G, et al. Ultrasound antenatal diagnosis of cleft palate by a new technique: the 3D reverse face view. Ultrasound Obstet Gynecol 2005;25:128. [24] Lee W, Kirk JS, Shaheen W, et al. Fetal cleft lip and palate detection by three-dimensional ultrasonography. Ultrasound Obstet Gynecol 2000; 16:31420. [25] Johnson DD, Pretorius DH, Riccabona M, et al. Three-dimensional ultrasound of the fetal spine. Obstet Gynecol 1997;89(3):4348. [26] Ploeckinger-Ulm B, Ulm MR, Lee ME, et al. Antenatal depiction of fetal digits with three dimensional ultrasonography. Am J Obstet Gynecol 1996;175:5714.

335

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 335356

Practical Approach to the Adnexal Mass


Maitray D. Patel,
& & & & &

MD

Normal anatomy Overview of ultrasound analytic approach to the adnexal mass Unilocular smooth-walled anechoic cyst Physiologic causes of findings that may raise concern When physiologic process is unlikely, has a characteristic pattern been established? Non-neoplastic cyst Endometrioma Hydrosalpinx

& &

Peritoneal inclusion cyst Cystic teratoma Benign and malignant cystic neoplasms (cystadenomas and cystadenocarcinomas) Other potentially characteristic adnexal masses Further imaging or surgical exploration Summary References

Evaluation of an adnexal mass, either presenting on physical examination, suspected based on clinical history, or identified on routine pelvic sonography, is a common task for the sonologist. While the clinical context is very important, for the vast majority of sonographically identified adnexal masses, the subsequent management of the patient will be highly dependent on the sonologists interpretation of the imaging findings. The sonologist who merely measures the size of a mass and who subsequently offers a differential diagnosis that includes nearly every adnexal abnormality, including malignancy, has failed in his or her opportunity to contribute meaningfully to the care of the patient. Using a practical approach [Fig. 1] and with knowledge of the sonographic patterns of adnexal pathology, the sonologist is better equipped to make reasoned conclusions and useful recommendations for patient management.

Normal anatomy
An understanding of the expected sonographic appearance of the ovary is important so that one

does not confuse normal structures with pathology. The ovary is highly dynamic, with constant formation and regression of physiologic cysts even before menarche. During the menstrual cycle, hormonally mediated ovulation begins with the recruitment of about five to eight preantral follicles, which are visible as small cysts during the early proliferative phase measuring about 2 to 4 mm in diameter [1]. A dominant follicle emerges by days 8 to 10 of the cycle, generally measuring about 10 mm in diameter, exceeding the diameter of the other follicles. Occasionally two dominant follicles will develop, but not in the same ovary [2]. Both dominant and nondominant follicles increase in size until ovulation; the dominant follicle is expected to become 2 to 2.5 cm in average diameter. In contrast, nondominant follicles typically do not exceed 1.1 cm in size [Fig. 2] [3]. With ovulation, the dominant follicle ruptures, typically losing much if not all of its internal fluid and collapsing into a sonographically solid appearing structure (if visible at all). This becomes the corpus luteum; the margins of the corpus luteum are hypervascular, leading to a ring of fire appear-

Department of Radiology, Mayo Clinic, 13400 E. Shea Blvd., Scottsdale, AZ 85259, USA E-mail address: patel.maitray@mayo.edu
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.003

ultrasound.theclinics.com

336

Patel

Fig. 1. Suggested approach to the sonographically identified adnexal mass.

ance on color Doppler sonography, and they may be thick and somewhat irregular, reflecting the loss of wall tension following rupture [Fig. 3] [4]. The corpus luteum is typically smaller than the mature follicle from which it arose, measuring about 1.5 cm. Hemorrhage into the corpus luteum can lead to re-expansion, resulting in a hemorrhagic ovarian cyst [Fig. 4]; the size of the cyst varies, but it is not unusual for it to be up to 4 cm in diameter. Recognizing this normal physiologic process is paramount to avoiding the temptation to view every ovarian cyst as a mass requiring treatment or follow-up imaging. Some experts have advocated

that sonologists avoid using the term cyst to describe anything in the ovary which is likely to be secondary to normal physiologic events [5]. Certainly, use of the terms follicle, dominant follicle, and corpus luteum in sonographic reports serves to describe these structures without inadvertently misleading others to believe that they are findings that are potentially pathologic. With even just a little experience, and with some attention to the phase of the menstrual cycle at which time the premenopausal patient is being imaged, sonologists should have no difficulty in ignoring these expected normal ovarian findings when they appear

Fig. 2. Dominant follicle. Sonogram of an ovary in a premenopausal woman shows a unilocular smooth-walled anechoic cyst measuring 2.5 cm in maximum diameter, with multiple other follicles in the ovarian parenchyma. This is the expected appearance of a dominant follicle in a premenopausal woman.

Practical Approach to the Adnexal Mass

337

Fig. 3. Corpus Luteum. (A) Sonogram of an ovary (demarcated by electonic calipers) in a premenopausal woman shows the corpus luteum (arrow) as a predominantly solid-appearing structure within the ovarian parenchyma, with a thick wall and internal echoes. (B) The rim of the corpus luteum shows hypervascularity as compared with the rest of the ovary, resulting in an appearance on color Doppler sonography that has been called a ring-of-fire.

typical, as they will in the vast majority of cases. Nevertheless, some of these physiologic structures will develop into larger masses or exhibit atypical features requiring an analytic approach by the sonologist for further evaluation.

Overview of ultrasound analytic approach to the adnexal mass


The sonographer who uses a practical approach to imaging recognizes that there are specific categories or groupings of pathologic processes that can result in an adnexal mass. The approach seeks to determine if a sonographically identified mass ex-

hibits an imaging pattern that is reasonably characteristic of a particular category. Some categories are broad, encompassing several different pathologic entities, whereas others are more specific, referring to a single pathologic entity; as a result, the categories overlap for some entities. The categories with potentially characteristic imaging features are as follows, listed in a morphologic spectrum from unilocular and entirely cystic to variable/multilocular/partially cystic to entirely solid: (1) nonneoplastic cyst; (2) hemorrhagic ovarian cyst; (3) endometrioma; (4) hydrosalpinx; (5) peritoneal inclusion cyst; (6) benign cystic neoplasm; (7) cystic teratoma; (8) cystic malignancy; (9) ectopic

338

Patel

Fig. 4. Hemorrhagic ovarian cyst. Sonogram of an ovary in a premenopausal woman shows a cyst (demarcated by electronic calipers) measuring over 6 cm in diameter, containing multiple internal fibrin strands. These characteristic fibrin strands should not be misinterpreted as septations.

pregnancy; (10) abscess/inflammatory mass; (11) torsed ovary; (12) exophytic or broad ligament myoma; (13) ovarian fibroma; and (14) solid mass. At the most basic level, the sonologists approach to the sonographically identified adnexal mass can be understood as a two-phase process. First, the sonologist determines if the mass exhibits a characteristic pattern indicating the likely cause or category. If so, the mass is managed as appropriate for that type of pathology in the clinical context of the patient. If not, the sonologist considers if a subsequent diagnostic test (including repeat follow-up ultrasound) can be used to safely allow potential characterization of the mass into a particular pathologic category and thereby avoid surgery or improve surgical planning. If so, the subsequent test or follow-up should be recommended. If not, diagnostic surgical evaluation will be indicated. In essence, the imager confronting an adnexal mass asks two questions: do I know what this is with reasonable certainty? If not, is there a test or strategy that has a reasonable chance of enabling me to know what this is and that would make a difference in the care of the patient? Note the emphasis on pattern recognition in this approach; it all boils down to whether one can be reasonably certain of the pathology. Research indicates that subjective evaluation of ovarian masses using pattern recognition achieves high sensitivity and specificity for discriminating malignant pathology from benign pathology [6,7]. Furthermore, investigations have quantified a high likelihood ratio for specific sets of observations that enable sonologists to confidently discriminate particular causes

of adnexal pathology [811]. When specific sonographic observations and considerations are placed in the framework of the basic practical approach (do I know what this is? if not, is there a test or strategy that will enable me to know what this is and make a difference to the care of the patient?), the sonologist is able to render effective recommendations for the management of the identified adnexal mass. Fig. 1 details a specific algorithm that expands on this basic practical approach by detailing a step-bystep analysis that facilitates understanding what a mass is likely to be and what should be done next. The critical questions to be answered are: (1) is the mass a unilocular anechoic smooth-walled cyst?; if so, how likely is it to be a non-neoplastic cyst that will not require treatment?; (2) if the mass is not a cyst that is unilocular, anechoic, and smoothwalled, is it possible that normal physiologic changes might account for the aberrant observations?; if so, how likely is the mass to be a nonneoplastic cyst that will not require treatment?; (3) if normal physiologic changes do not likely explain the aberrant observations, can one identify other observations that enable classification into a particular category?; (4) if sonographic observations are insufficient to allow characterization as a particular pathologic category, are there other diagnostic tests that could be used that might reasonably assist in placing the mass in a single category?; if so, would that be clinically meaningful? The remainder of this article focuses on detailing the sonographic observations and considerations relevant to the questions posed by the algorithm,

Practical Approach to the Adnexal Mass

339

with primary attention to those masses with cystic features. The adnexal mass related to ectopic pregnancy or inflammation/abscess almost always arises in a specific clinical setting that has unique analytic considerations separate from the focus of this discussion, and these entities will not be further discussed. Likewise, there is no discussion regarding the sonographic patterns of ovarian torsion and various categories of adnexal solid masses. The recommended imaging algorithm for an adnexal mass can and should be applied to these entities, but discussion of the analysis of these masses is beyond the scope of this article.

Unilocular smooth-walled anechoic cyst


The first step in analysis of the sonographically identified adnexal mass is to decide if it meets the criteria for a unilocular anechoic smooth-walled cyst. These cysts are common in premenopausal women, since essentially every nonhemorrhagic dominant follicle could be thus described. With the increasing use of ultrasound, it has been firmly established over the last few decades that these cysts are also common in postmenopausal women. The incidence of such cysts in postmenopausal women has been reported to be up to 20% [1214]. These cysts are round or oval in shape, have no internal structure or echogenicity, and have a clearly demonstrable wall without defined surface projections or nodularity [Fig. 5]. A cyst with slight crenulation of the wall resulting in mild irregularity

would still be considered smooth-walled if no clearly defined surface projections are evident. There are practically two pathologic possibilities when a unilocular anechoic smooth-walled adnexal cyst is identified: either the mass is a non-neoplastic cyst or it is a benign cystic neoplasm. The term non-neoplastic cyst is a very useful imaging designation but not highly specific. A number of specific pathologic entities can fall into this category, including physiologic cysts (cysts that develop in the course of ovulation, namely the follicular cyst and corpus luteum cyst), theca lutein cysts, serous inclusion cysts, endometriomas, peritoneal inclusion cysts, and paraovarian/paratubal cysts. Some of these pathologic entities may demonstrate other sonographic features that enable more specific categorization; for example, sonographic identification of a separate ovary will allow more specific categorization of a unilocular smooth-walled anechoic cyst as a paraovarian cyst. Furthermore, not every individual pathologic entity in this list will always appear as a unilocular anechoic smooth-walled cyst; for example, the vast majority of endometriomas will contain low-level echoes [8]. It is important to have confidence in the fact that the risk of malignancy is extremely low when a mass is assuredly characterized as a unilocular smooth-walled cyst. The data supporting this conclusion are extensive. Based on the natural history of more than 3000 unilocular ovarian cysts identified in postmenopausal women measuring 10 cm or less in diameter, Modesitt and colleagues [13] calculate a risk of malignancy of less than 0.1%

Fig. 5. Unilocular smooth-walled anechoic cyst. Sonogram of the right ovary in this postmenopausal patient shows a cyst meeting criteria for designation as a unilocular smooth-walled anechoic cyst. In this case, the cyst measures just over 3.5 cm in maximum diameter. Minimal irregularity at the 5 oclock position of the cyst is not sufficient to raise concern.

340

Patel

with 95% confidence interval. The rare unilocular smooth-walled cyst that does eventually prove to be malignant (usually borderline) has papillary projections or septations identified on follow-up [13]. These rare malignancies may be misclassified as unilocular smooth-walled cysts on initial evaluation by failure to identify the wall nodularity [15]. Size is also an important consideration; the rare borderline tumor or malignancy that appears to be a unilocular anechoic smooth-walled cyst is nearly always over 5 cm in diameter [15]. Distinguishing between a non-neoplastic cyst and a benign neoplasm is clinically relevant. Nonneoplastic cysts are often self-limiting and resolve without intervention. The main risk of an asymptomatic non-neoplastic ovarian cyst that does not resolve spontaneously is that it can theoretically induce torsion of the ovary, though the risk of this event must be very low, given the commonality of non-neoplastic cysts and the infrequency of torsion. Thus, if one can be reasonably confident that an adnexal mass is due to a non-neoplastic cyst, there are several management options. One could choose to ignore the mass as long as the patient remained asymptomatic, assuming that it will disappear over time. Alternatively, one could elect to percutaneously aspirate and treat the cyst [Fig. 6] [16]; this course of action is less frequently chosen, usually only for those non-neoplastic cysts that are of certain size and that have not resolved over time. Indications for this maneuver might include localized patient discomfort related to the volume of the mass and its effect on adjacent organs, such as the urinary bladder. Of course, surgical manage-

ment remains an option if the patient is symptomatic or desires removal. On the other hand, a unilocular smooth-walled cyst that has a reasonable chance of being a benign neoplasm is almost always surgically removed. First, there is a concern about the rare malignancy appearing to be a unilocular smooth-walled cyst. Even if one could be further assured of the benignity of a mass by evaluation of CA-125 levels or other considerations [17], benign neglect of a mass which has features suggesting a benign neoplasm rather than a non-neoplastic cyst is not a long-term option as it can be expected to lead only to a situation that will have to be addressed later, either emergently if the mass induces ovarian torsion or electively as the mass enlarges and causes symptoms. Percutaneous drainage of a cystic mass likely to be a benign neoplasm rather than a nonneoplastic cyst could also be considered, but this approach remains controversial [16]. There are four factors to consider when trying to distinguish between a non-neoplastic cyst and a benign neoplasm as the cause of a unilocular anechoic smooth-walled mass: location, growth rate, menopausal status, and size. Location is the most obvious consideration, as identification of a clearly separate ovary nearly eliminates the possibility of a benign neoplasm [Fig. 7]. There are exceptions to every rule, and in this regard, there are reports of paraovarian cystadenomas [18,19]; nevertheless, for practical purposes one can reasonably designate a mass as very likely being a nonneoplastic cyst when a clearly separate ipsilateral ovary is identified.

Fig. 6. Percutaneous aspiration of a unilocular smooth-walled anechoic cyst. Sonogram of a persistent left adnexal cyst in a perimenopausal woman documents the percutaneous insertion of a needle. The cyst was completely aspirated. In this case, the cyst had been demonstrated to be 50% larger on a prior sonogram 18 months ago. It was stable since the previous sonogram 6 months ago.

Practical Approach to the Adnexal Mass

341

Fig. 7. Paraovarian cyst. Sonogram of the left adnexa in an asymptomatic postmenopausal woman shows a unilocular smooth-walled anechoic cyst (demarcated with electronic calipers). An adjacent normal appearing ovary is also identified (arrow). Other images (not shown) confirmed the impression that the cyst was abutting but not arising from the ovary. Follow-up sonograms were not requested as long as the patient remained asymptomatic.

Consideration of the growth rate of the cyst can also be fruitful in distinguishing between a nonneoplastic cyst and a benign neoplasm. Cysts that can be documented to have appeared suddenly, always in premenopausal women, are clearly nonneoplastic physiologic cysts. Thus, a patient who had a sonogram within the preceding few months demonstrating normal ovaries without mass, who then returns for evaluation of rapid onset of unilateral adnexal pain wherein a 4 cm ovarian cyst is identified, clearly has a non-neoplastic cyst as the cause of the mass. Neoplasms do not exhibit this type of hypergrowth. Similarly, cysts that are shown to have long-term stability are undoubtedly non-neoplastic (however, a pathologist may designate such a cyst as a cystadenoma if it eventually gets surgically removed, because the pathologic distinction between a cystadenoma and some types of non-neoplastic ovarian cysts can be imprecise). Finally, any cyst that demonstrates reduction in size is clearly not neoplastic. Size and hormonal status are important considerations when a unilocular smooth-walled cyst is encountered, not only to suggest whether the cyst is non-neoplastic or a benign neoplasm but also determine the intensity of additional testing to recommend. Keeping in mind that normal physiologic events are expected to result in the development of a dominant follicle and a corpus luteum in every normal ovulatory cycle in a premenopausal woman, unilocular smooth-walled cysts below a certain size in premenopausal women can be assumed to be non-neoplastic in the absence of

contrary clinical features. There is no magic number below which benign neoplasm becomes theoretically impossible; likewise, there is no magic number above which non-neoplastic cyst is theoretically excluded. A practical approach that I use is to consider a unilocular smooth-walled anechoic cyst in a premenopausal or perimenopausal woman measuring 4 cm or less in average diameter as very likely non-neoplastic in cause (I use the 4-cm threshold as a reasonable approach; some would use 3 cm, even if the woman is premenopausal). Conversely, a similar cyst in the same patient measuring 8 cm or larger is very likely a benign neoplasm. For postmenopausal women, I reduce the threshold by 3 cm; thus, a unilocular smoothwalled cyst measuring less than 1 cm in diameter is very likely non-neoplastic, and a similar cyst measuring over 5 cm is very likely a benign neoplasm. Those cysts that fall between these ranges are considered to be either a non-neoplastic cyst or benign neoplasm (distinction not yet possible). The probabilities favor non-neoplastic cyst at the lower end of the size range and benign neoplasm at the higher end of the size range, but additional testing (sonographic follow-up) would be appropriate. Many sonographically identified unilocular smoothwalled cysts can be ignored; many do not require sonographic follow-up or surgical intervention. In fact, imagers who routinely recommend sonographic follow-up for a cyst that already exhibits features or behavior that indicates that it can be reasonably expected to be non-neoplastic force

342

Patel

Practical Approach to the Adnexal Mass

343

sonographic overuse. When a mass is very likely to be a non-neoplastic cyst, the imaging report might be worded as follows: In the absence of persistent symptoms or other clinical considerations, sonographic follow-up should not be necessary. For example, this would be appropriate for the unilocular smooth-walled cyst measuring less than 4 cm in maximum size in a premenopausal woman, less than 1 cm in size in the post-menopausal woman, any cyst showing long-term stability, and any paraovarian unilocular smooth-walled cyst. (Some would opt to follow these cysts, likely non-neoplastic, at least once to demonstrate resolution or stability. If you choose this option, it is recommended that a reasonable interval is chosen. This might be 6 months for an asymptomatic woman.) This approach recognizes that sonographic follow-up may still be indicated if clinical circumstances necessitate. As noted previously, some unilocular smoothwalled cysts cannot be categorized as being very likely to be non-neoplastic or very likely to be a benign neoplasm based on the sonographic evaluation thus far. The sonologist does not know with reasonable certainty what the mass is. The practical approach is then to ask if there a test or strategy that has a reasonable chance of enabling the sonologist to distinguish between these two possibilities and make a difference in the care of the patient. In such cases, sonographic follow-up will allow assessment of the stability or growth of the mass that may facilitate making the distinction. Obviously, if the mass resolves or is demonstrably smaller on follow-up, it is a non-neoplastic cyst which wont need treatment; conversely, if the mass grows, the suspicion that it is a benign neoplasm increases (though non-neoplastic cysts can also increase in size), likely to require intervention [Fig. 8]. If a unilocular smooth-walled cyst does not change in size on follow-up, it is more likely to be a non-neoplastic cyst, assuming that there has been enough of an interval between the two studies to enable detection of growth. Thus, it is important that follow-up studies be performed with enough of an interval to assist the sonologist in making this determination. It is common for sonologists encountering a unilocular smooth-walled cyst to recommend follow-up in 6 to 8 weeks or after one or two menstrual cycles. No doubt, the vast

majority of non-neoplastic cysts in premenopausal women can be expected to change (get smaller or resolve) in that time-frame; in fact, most will resolve during the course of a single menstrual cycle (3 to 5 weeks). Nevertheless, one should avoid the temptation to request sonographic follow-up at too short an interval for the follow-up to be meaningful. Universal application of the 6-to-8 weeks rule for follow-up is inappropriate, especially in postmenopausal or perimenopausal women but also for many premenopausal women. Though nonneoplastic cysts in perimenopausal or postmenopausal woman are not uncommon and often self-limiting, there should be no expectation for these cysts to resolve or meaningfully change in 6 to 8 weeks. Furthermore, the asymptomatic premenopausal woman with a unilocular smoothwalled cyst is not well-served by the follow-up study performed in 6 to 8 weeks that again shows the cyst unchanged in size, as the time interval is too small to make any conclusions; it could still be a non-neoplastic cyst or a benign neoplasm. Granted, most cysts in premenopausal women will indeed resolve in this short interval, but documenting such resolution so quickly is usually not necessary, and the occasional non-neoplastic cyst will take more time to resolve. In my practice, when I am trying to distinguish between a non-neoplastic cyst and a benign neoplasm as the cause of a typical unilocular smoothwalled anechoic ovarian cyst, I will request sonographic follow-up in 6 months. If there is some atypical feature that I suspect is artifactual or caused by a physiologic process (see ensuing discussion), or if there are even vague clinical symptoms possibly attributed to the mass, I will cut the follow-up interval in half (3 months). In the uncommon situation in which I am even more concerned about the atypical features of the mass but do not feel other testing or intervention is yet justified, or if there are clinical symptoms clearly attributed to the mass which persist, I will further cut the follow-up interval in half (1.5 months, or 6 weeks). When following a unilocular smooth-walled cyst to try to distinguish between a non-neoplastic cyst and a benign neoplasm, actual calculation of the volume of the cyst and the apparent doubling time can be helpful [see Fig. 8]. Small differences

Fig. 8. Benign cystadenoma. (A) Sagittal and (B) coronal sonograms of the right ovary in a postmenopausal woman demonstrate a unilocular smooth-walled cyst measuring 3.3 2.2 1.8 cm. With these measurements, the volume of the cyst is calculated at 6.8 cubic cm. Subsequent follow-up sagittal (C ) and coronal (D) sonograms performed 6 months later demonstrate the cyst to measure 3.7 2.2 2.1 cm. This yields a volume calculation of 8.9 cubic cm. The doubling time is calculated to be 19 months. A benign neoplasm was suspected; the mass was surgically removed and proved to be a serous cystadenoma.

344

Patel

in the measured transverse, craniocaudal, and anteroposterior diameter of a cyst can occur between studies due to technical variability, so reliance on only one plane of measurement when serially imaging a mass can be misleading. Furthermore, small masses can demonstrate significant increase in volume with seemingly minimal diameter changes. For example, a cystic mass that is 2.0 cm in average diameter has doubled in size when the diameter increases to 2.5 cm. The formula for calculation of the volume of an ellipsoid mass (length width height 0.52) is applied. To calculate the estimated doubling time, the time interval between the two measurements is divided by the percent increase in size of the mass. Thus, a cyst that grows from 3.5-cm average diameter to 4.0-cm average diameter in 6 months has an estimated doubling time of 1 year. This would be consistent with a benign neoplasm, and subsequent management would be appropriately directed with this assumption. A unilocular cyst that demonstrates an increase in size is not always a benign neoplasm rather than a non-neoplastic cyst. Confidence in the assessment of interval growth depends on the magnitude of the change in measurement and the interval of observation. Re-evaluation of the size of a mass in a short interval allows technical variability to potentially mislead the observer. For example, a 2.0-cm average diameter mass that is subsequently reevaluated in 6 weeks and that appears to measure 2.1 cm in average diameter has potentially demonstrated a 15% increase in volume, resulting in a calculated doubling time of 40 weeks (6 weeks/0.15 = 40 weeks); obviously, this doubling time calculation is imprecise and unreliable because the magnitude of the measurement change (1 mm) is within the range of technical variability and the interval of observation is small. Even when the observer is more confident regarding the existence of true interval growth, an enlarging unilocular smooth-walled cyst could still be a non-neoplastic cyst rather than a benign neoplasm. Though there is no scientific literature to indicate the range of growth of benign ovarian cystic neoplasms, if the calculated doubling time of a unilocular smoothwalled cyst is very lengthy (exceeding 3 years), the mass could well be non-neoplastic. Nevertheless, unless the calculated doubling time is very lengthy, it is reasonable to conclude that the enlarging unilocular smooth-walled cyst is a benign neoplasm and not a non-neoplastic cyst and manage the mass with this expectation. Once the cyst has demonstrated growth, the institution of other testing (including serologic CA-125 assessment or MRI) will not be clinically meaningful as the results will not allow one to conclude that the cyst is an atypical non-neoplastic cyst. Furthermore, it makes no clini-

cal difference whether the enlarging cyst is a benign neoplasm or a non-neoplastic cyst; in either case, intervention is usually warranted.

Physiologic causes of findings that may raise concern


What if a mass violates one or more of the criteria for designation as a unilocular smooth-walled anechoic cyst? For example, suppose the mass contains a possible or definite septation, has wall irregularity or nodularity, or has diffuse or focal areas of echogenicity? For such lesions, the next step in the practical approach is to determine if the aberrant observation(s) could potentially be caused by a physiologic process. There are three physiologic processes to consider: (1) occurrence of two or more individual cysts next to each other, thus mimicking a multilocular mass with intervening septations; (2) cyst involution resulting in wall irregularity; and (3) intracystic hemorrhage resulting in internal echoes, internal strands and retractile clot. Of these three processes, hemorrhage is the most common in premenopausal women. Hemorrhagic ovarian cysts frequently demonstrate features that can be confused with ovarian neoplasms [9]. Retracting thrombus adhering to the cyst wall may be mistaken for a mural nodule [Fig. 9]. Hemorrhagic cysts contain fine networks of linear echoes, which may be mistaken for septations [see Fig. 4]. True septations, when seen, are highly characteristic of ovarian neoplasms. Thus, this appearance may confuse the examiner into believing that the sonographically observed mass is a neoplasm. These fine linear echoes are not septations but rather a manifestation of clot architecture, likely originating from strands or bands of fibrin. It is probably because of these potentially confusing appearances that some older studies on hemorrhagic cysts most often conclude that they have a nonspecific appearance,considering the diagnosis difficult and describing these masses as the great imitator [20]. It is unfortunate that hemorrhagic cysts are frequently said to be indistinguishable from ovarian neoplasms. In fact, hemorrhagic cysts almost always represent non-neoplastic cysts that will resolve spontaneously [21,22]. All of the 24 cases followed by Okai and colleagues [22] resolved within 8 weeks. Thirty percent of the masses resolved within 2 weeks. Among 38 surgically excised hemorrhagic cysts, none were neoplasms [21]. Importantly, no ovarian malignancies have been erroneously reported to represent hemorrhagic ovarian cysts in series evaluating the diagnostic performance of sonography for this pathologic entity. Therefore, if the examiner can correctly conclude

Practical Approach to the Adnexal Mass

345

Fig. 9. Hemorrhagic ovarian cyst. (A) Sonogram of the right ovary in a premenopausal woman shows clumped material along the wall of an otherwise smooth-walled unilocular cyst at the 5 oclock position. This retracting clot (arrow) has a slightly lower echogenicity than the adjacent wall. (B) Color Doppler sonogram of the cyst shows no detectable blood flow within the clot.

that an ovarian mass is a hemorrhagic cyst, then one may be confident that the lesion will almost certainly resolve spontaneously. Should the lesion not resolve, there is virtually no chance that it represents a misdiagnosed ovarian carcinoma. In contrast to the nonspecific appearance often previously described in the literature, our study indicates that hemorrhagic ovarian cysts have a very specific appearance in the vast majority (90%) of cases [9]. The key gray-scale ultrasound features to observe are the presence of fibrin strands or retracting clot, with absence of suspected septations and wall irregularity secondary helpful findings [see Fig. 4; Fig. 9]. When a unilocular smoothwalled cyst containing fibrin strands is identified, the observed mass is 200 times more likely to be a hemorrhagic ovarian cyst than any other possibility. Even if the sonologist can only be confident that fibrin strands are present but is uncertain as

to whether there are coexisting true septations or wall irregularity, the presence of fibrin strands makes the mass 40 times more likely to be a hemorrhagic ovarian cyst [Fig. 10]. Finally, if one can identify a mural-based structure with features characteristic of a retracting clot, the mass is at least 67 times more likely to be a hemorrhagic ovarian cyst than any other entity. Understanding that the likelihood ratio is this high with these observations should help the less experienced sonologist feel more confident in the diagnosis and allow for observation instead of surgical intervention in appropriate cases. It is critical, therefore, for the sonologist to master the typical appearance of fibrin strands and retracting clot. Jain and colleagues [23] described the appearance of fibrin strands in hemorrhagic cysts as fish netting. Although a mass may contain many septations, the number is usually fewer than 20. In

346

Patel

Fig. 10. Hemorrhagic ovarian cyst. Sonogram of the right ovary in a premenopausal woman, with color Doppler, shows a mass with internal echoes and linear echogenic bands which simulate septations. Fibrin strands are identified in one quadrant of the mass (from 6 to 9 oclock). Color Doppler evaluation confirms the absence of demonstrable blood flow within the mass. The mass resolved on follow-up sonogram performed 6 weeks later.

contrast, fibrin strands are often innumerable. Septations are usually thicker and more reflective than fibrin strands. Importantly, septations visually track for a reasonable and appropriate distance as they are viewed using real-time sonography. Indeed, even on still images, this feature is usually obvious. Unlike septations, fibrin strands are discontinuous and seem to flit from plane to plane either on realtime examination or on captured still images. Similarly, retracting thrombus has a detectably and reliably different appearance than mural nodules. Mural nodules have convex margins, whereas retracting thrombus has concave margins. Neoplastic tissue does not grow with a concave margin. Furthermore, echoes returning from thrombus tend to differ in amplitude compared with the cyst wall and are usually less echogenic. Mural nodules tend to have the same echogenicity as the cyst wall or, in the case of the Rochitansky nodule of cystic teratomas, markedly greater echogenicity than the cyst wall. Color Doppler analysis is important to apply when one is considering the possibility of hemorrhage as a cause for apparent intracyst septations or cyst wall irregularity or nodularity. Retracting clot causing apparent septations or wall nodules should not exhibit blood flow [see Figs. 9 and 10]. Similarly, identification of vascularity within a mass with diffuse low-level echoes identifies the mass as solid, not possibly cystic with internal hemorrhage. If one is entertaining the notion of hemorrhage as a cause for an aberrant finding, the demonstration of blood flow in that structure eliminates that possibility. Unfortunately, the same cannot be said for the other physiologic processes that might possibly

account for observations of apparent septations or wall irregularity. When two cysts lie next to each other, the resulting interface of cyst wall and any intervening ovarian parenchyma can be interpreted as a septation [Fig. 11]. Blood flow may be demonstrated within this septation as a function of the normal ovarian parenchyma trapped between the two cysts. Likewise, when physiologic cysts involute, the wall irregularity of the cyst, corresponding to normal ovarian parenchyma along the surface of the flaccid cyst, can contain vessels. Once the sonologist has determined that a physiologic process (adjacent cyst, cyst involution, or intracyst hemorrhage) could possibly account for the observations that make the mass not be classified as a unilocular anechoic smooth-walled cyst, he or she makes a general assessment of how likely it is that the mass will prove to be non-neoplastic in the same manner as described previously for the unilocular smooth-walled anechoic cyst. Again, factors to consider will include growth rate, location, patient hormonal status, and cyst size. Because of the compelling likelihood that one is dealing with a non-neoplastic hemorrhagic cyst with some observations (fibrin strands and retracting clot), it is reasonable to ignore those hemorrhagic cysts that contain these characteristic features when they are below the 4 cm size threshold articulated previously (assuming that the patient becomes asymptomatic); in such cases, sonographic reports should emphasize that sonographic follow-up may not be necessary. In other cases, sonographic follow-up allows assessment of cyst resolution, stability, or growth, to help distinguish between a nonneoplastic cyst and other possible causes of the mass. In this situation, the time interval chosen

Practical Approach to the Adnexal Mass

347

Fig. 11. Two adjacent unilocular cysts. (A) Sonogram of the right ovary in a premenopausal woman shows an apparent septation between two cystic compartments. Although this could be a bilocular cyst with an internal septation, the sonologist needs to consider that it could reasonably also be two adjacent unilocular cysts (as it was in this case). (B) Color Doppler sonogram demonstrates blood flow in the possible septation between the two cystic compartments. Normal ovarian parenchyma trapped between two unilocular cysts (as in this case) can demonstrate blood flow.

for follow-up is often justifiably shorter than the 6-month period chosen for the asymptomatic unilocular smooth-walled anechoic cyst. As articulated previously, I usually choose to cut the follow-up interval in half (3 months instead of 6 months) when there is a finding potentially caused by a physiologic process or even by another half (i.e., 1.5 months, or 6 weeks) when I think a particularly ominous sonographic finding might possibly be caused by hemorrhage or cyst involution. One final comment regarding hemorrhage and cyst involution as a cause of wall irregularity; these physiologic processes occur almost exclusively in premenopausal women. It would be extremely unusual for intracyst hemorrhage to result in retrac-

tile clot in a woman who has undergone menopause, and these women would not be expected to have an involuting corpus luteum cyst. Of course, the coexistence of two unilocular cysts next to each other, simulating a single multilocular cyst, can occur in women of any age.

When physiologic process is unlikely, has a characteristic pattern been established?


The mass that is not a unilocular anechoic smoothwalled cyst, due to features that are not possibly physiologic in origin, is then analyzed to determine if features characteristic of a particular category have been observed. The categories for a cystic

348

Patel

mass to consider would include non-neoplastic cyst, endometrioma, hydrosalpinx, peritoneal inclusion cyst, benign cystic neoplasm, cystic teratoma, and cystic malignancy. For each of these categories, specific observations can enable the sonologist to reasonably conclude that the mass in question is caused by that entity.

Non-neoplastic cyst
Most non-neoplastic cysts will have already been identified as such by the preceding analysis of unilocularity, wall regularity, and likely physiologic mimics of suspicious pathology. Nevertheless, some cysts will not meet these preceding criteria

but can still be confidently assumed to be nonneoplastic by demonstrating lack of interval growth [Fig. 12]. Thus, the cyst which appears to have wall irregularity not due to hemorrhage (because the irregularity has not resolved) could still be classified as likely due to a non-neoplastic cyst by demonstrating no growth with serial sonography over a sufficiently long observation interval (at least 2 years). Having established this stability, the mass can then be managed in a manner appropriate for a non-neoplastic cyst. This would include continued periodic sonographic re-evaluation, percutaneous aspiration with or without sclerotherapy or methotrexate injection, or surgical removal.

Fig. 12. Non-neoplastic cyst with wall irregularity. (A) Sonogram of the left ovary in a postmenopausal woman shows a small 1-cm cyst with wall irregularity (arrow). Color Doppler images (not shown) did not demonstrate blood flow in the area of focal wall irregularity. (B) The cyst was periodically followed with ultrasound. Follow-up sonogram 2 years later demonstrates no change in the size or appearance of the cyst or focal area of irregularity (arrow). The mass is likely a non-neoplastic cyst and is justifiably further managed conservatively.

Practical Approach to the Adnexal Mass

349

Fig. 13. Endometrioma. Sonogram of the left ovary shows a mass with diffuse low-level internal echoes and multiple hyperechoic wall foci (best seen at 8 to 9 oclock).

Endometrioma
The presence of diffuse low-level internal echoes is an important feature that helps to consider the possibility of endometrioma as the cause of an adnexal mass; our prior study demonstrated that 95% of endometriomas exhibit diffuse lowlevel internal echoes, with technical factors likely accounting for those few endometriomas in which low-level echoes were not demonstrated [Fig. 13] [8]. While the absence of this finding does not exclude endometrioma, it significantly decreases the

likelihood of that diagnosis (negative likelihood ratio=0.1). Nevertheless, the presence of diffuse low-level echoes in a mass is clearly not enough for the sonologist to conclude that the mass is an endometrioma, as some hemorrhagic ovarian cysts, benign neoplasms (including cystic teratoma), and malignancies can also exhibit this feature [Fig. 14]. As discussed previously, hemorrhagic cysts are almost exclusively non-neoplastic and most resolve spontaneously; they are surgically removed only when patients have compelling acute symptoms. The practical approach to the adnexal mass articulated thus far allows the sonographer to consider the possibility that a physiologic process (intracyst hemorrhage in a premenopausal woman) is the cause for the diffuse internal echoes in an otherwise unilocular smooth-walled mass. This justifiably leads to follow-up sonography as a diagnostic strategy to use to distinguish between acute hemorrhagic cyst and endometrioma [see Fig. 14]. If the diffuse low-level echoes persist unchanged on the follow-up examination, acute hemorrhagic cyst can be reasonably excluded from consideration. Another particularly helpful observation to enable distinction between the acute hemorrhagic cyst and the endometrioma is the presence of one or more hyperechoic foci in the wall of a unilocular smooth-walled cyst with diffuse low-level echoes [see Fig. 13]. In our study, hyperechoic wall foci were seen in only 1 of 69 non-neoplastic cysts (a mass that did not resolve with periodic sonographic follow-up and considered a simple cyst at pathology) but were found in 14 of 40 (35%) endometriomas [8]. Based on these data, a smoothwalled mass with low-level internal echoes and

Fig. 14. Two hemorrhagic ovarian cysts. Sonogram of the right ovary in a premenopausal woman shows two masses with diffuse low level-internal echoes. Both cysts show mild wall irregularity (arrows). Both hemorrhagic cysts resolved on follow-up sonography performed 3 months later.

350

Patel

coexisting hyperechoic wall foci is 32 times more likely to be an endometrioma than another adnexal mass; for masses that exhibit this pattern, sonographic follow-up is a low-yield course of action. The pathologic basis of these hyperechoic wall foci has not been established. Given the similarity in appearance to hyperechoic wall foci seen in the gallbladder wall in patients with hyperplastic cholecystoses, due to the presence of cholesterol within polyps or Rokitansky-Aschoff sinuses, it has been postulated that these foci may contain cholesterol, perhaps from the breakdown of cell membranes subsequently phagocytized by giant cells. This may explain why they are seen in endometriomas, which contain chronic collections of cells that have had time to break down, and are not seen in spontaneously resolving lesions such as acute hemorrhagic cysts. The presence of hyperechoic wall foci in a cystic mass allows the observer to discount the possibility that the mass will resolve with expectant management. It does not necessarily mean that the mass is an endometrioma, however; if the theory that these hyperechoic wall foci represent cholesterol deposition in the wall of a chronic mass is true, one could also expect to see these hyperechoic foci in neoplasms. Indeed, neoplasms can exhibit this feature [8].

Thus, to have confidence that a mass with diffuse low-level echoes is an endometrioma, there is considerable benefit from further assessment of wall nodularity, a feature that is associated with neoplasia [24]. In our prior study, excluding masses with wall nodularity from the diagnosis of endometrioma helped to distinguish the endometriomas from neoplasms, especially malignancies [8]. However, as 20% of endometriomas can demonstrate wall nodularity, one can expect to be unable to reasonably distinguish between neoplasm and endometrioma for some masses with diffuse lowlevel echoes [Fig. 15]. Of the 11 masses in our study that demonstrated low-level internal echoes with wall nodularity but no features of cystic teratoma, 6 were endometriomas and 5 were neoplasms. This subgroup of patients may benefit from additional imaging with color Doppler sonography or MRI to try to distinguish between neoplastic and nonneoplastic causes of wall nodularity.

Hydrosalpinx
Sonographic findings associated with hydrosalpinx have been described by multiple investigators. Tessler and colleagues [25] found a tubular structure with folded configuration (incomplete septation) as the most consistent feature in their 12 cases of hydrosalpinx [Fig. 16]. Short linear projections were also seen in approximately half of their patients. Timor-Tritsch and colleagues [26] expanded the analysis of the sonographic features of hydrosalpinx by analyzing the mass shape, wall structure, wall thickness, and extent of ovarian involvement. This analysis suggested that many hydrosalpinges were ovoid or pear-shaped fluid collections containing incomplete septa, short linear projections (cogwheel sign), or small hyperechoic mural nodules (beads-on-a-string). Recent data have refined these observations [11]. The most accurate sonographic diagnosis of hydrosalpinx as the cause of a cystic adnexal mass is achieved by first determining whether or not the mass appears tubular and then focusing on whether a waist sign or small round projections can be identified [see Fig. 16; Fig. 17]. The small round projections correspond to the beads-on-a-string described by Timor-Tritsch and colleagues [26], pathologically caused by fibrosis of endosalpingeal folds. The waist sign refers to diametrically opposed indentations along the wall of the mass. As we recently reported [11], in our experience the combination of tubular shape and waist sign had no false positives for diagnosis of hydrosalpinx, leading to a calculated likelihood ratio exceeding 18.9; the exact likelihood ratio falls somewhere between this value and infinity (more

Fig. 15. Endometrioma with wall nodule. Sonogram of the left ovary shows a cystic compartment with diffuse low-level internal echoes and wall nodule (thick arrow). It is uncertain based on this image whether the entire mass is bilocular with an internal septation (thin arrows) or whether there are two adjacent cysts with intervening ovarian parenchyma. A small amount of normal ovarian parenchyma containing three follicles is present at the 12 oclock position. At pathologic evaluation after laparoscopic surgical removal, no wall vegetations were identified; the apparent wall nodule was likely related to clot fibrosis.

Practical Approach to the Adnexal Mass

351

Fig. 16. Hydrosalpinx. Sonogram of the left adnexa shows a cystic mass with folded tubular configuration. The folded configuration leads to an apparent incomplete septation (arrow). There is a waist sign (asterisks), which refers to the presence of diametrically opposed indentations along the wall of the mass. A short linear projection is present at one of the sites of indentation.

precise determination of the true likelihood ratio of this combination of findings for the diagnosis of hydrosalpinx would have required a larger study population leading to at least one false-positive case). The combination of tubular shape with small round mural projections also performed very well, with likelihood ratio of 22.1; there was one false positive case using this combination, in which a paratubal cyst exhibited both features. Two of the previously described sonographic findings associated with hydrosalpinx had no additional value in predicting that a mass was a hydrosalpinx when combined with the observation of tubular shape. Although the presence of an incomplete septation and short linear projections were findings predictive of hydrosalpinx, each was less predictive than tubular shape or the waist sign; moreover, in contrast to the waist sign and small

round projections, neither incomplete septation nor small linear projection improved diagnostic performance when combined with tubular shape. In other words, the likelihood ratio of tubular shape when combined with incomplete septation or short linear projection for the diagnosis of hydrosalpinx decreased as compared with the likelihood ratio of tubular shape alone. This is because apparent incomplete septations are also identified in some cystic neoplasms. When trying to distinguish a hydrosalpinx from a cystic ovarian neoplasm, the sonographic identification of a normal-appearing ovary ipsilateral to a cystic adnexal mass has an extremely positive effect on the ability of the sonologist to accurately predict that the mass is a hydrosalpinx and not a cystadenoma or cystadenofibroma. Although paraovarian cystadenomas can occur, they are exceedingly rare.

Fig. 17. Hydrosalpinx. Sonogram of the right adnexa shows a tubular cystic mass with small round projections along the wall (arrows), which have been described as beads-on-a-string.

352

Patel

Fig. 18. Peritoneal inclusion cyst. Sonogram of the left adnexa shows a deformed ovary (arrow) with large exophytic cystic collection on one side, as well as smaller collection of fluid with thin septations on the other side.

Thus, careful attention to identification of a normal ovary ipsilateral to the mass suspected of being a hydrosalpinx can clinch the diagnosis.

Peritoneal inclusion cyst


Peritoneal inclusion cysts, also known as peritoneal pseudocysts, occur as a result of trapped fluid between the ovary and adhesions in the peritoneal cavity. They begin in premenopausal patients who have had prior surgery, trauma, peritoneal infection, or endometriosis. The sonographic appearance of these peritoneal inclusion cysts can be pathognomonic [27]; in such cases, the margins of the cystic collection follow the contours of the pelvis, with some areas of acute angulation, and the often deformed ovary can be seen suspended among the adhesions centrally or peripherally [Fig. 18].

none of the other sonographic features associated with cystic teratomas [10]. There are two other important sonographic findings associated with dermoids [10]. The presence of diffuse or regional high-amplitude echoes within the mass, one of these other findings, is due to sebum; nearly 60% of cystic teratomas will exhibit this finding [see Fig. 19]. The other important feature is the presence of hyperechoic lines and dots within the mass, the so-called dermoid mesh [28]. This sonographic appearance is caused by hair and is also found in about 60% of cystic teratomas [see Fig. 19].

Cystic teratoma
A number of sonographic findings have been associated with cystic teratoma [10]. The feature that most commonly defines an ovarian mass as a cystic teratoma is the observation of focal or diffuse highamplitude echoes that attenuate the acoustic beam (shadowing echodensity) [Fig. 19]. There are three types of tissues that can produce this finding: calcified structures like bone and teeth, clumps of hair in a cystic cavity, and fat in a Rokitansky protuberance. It does not matter which of the three situations results in the sonographic identification of focal or diffuse high-amplitude echoes that attenuate the acoustic beam. Each of these tissues predicts that an ovarian mass is a cystic teratoma. Almost 90% of cystic teratomas demonstrate a shadowing echodensity; in only 16% of cases, there will be

Fig. 19. Cystic teratoma. Sagittal endovaginal sonogram of an adnexal mass demonstrates shadowing echodensity (thick arrow), regional area of bright internal echoes (arrowhead), and hyperechoic lines and dots (long arrow). This combination of sonographic features is virtually pathognomonic for a cystic teratoma.

Practical Approach to the Adnexal Mass

353

Each of these three sonographic findings (shadowing echodensity, regional or diffuse bright echoes, hyperechoic lines and dots) can occasionally be identified or mimicked as isolated findings in other types of adnexal masses [10]. Fibrotic wall nodules seen in some endometriomas can produce acoustic shadowing. Hemorrhage can cause regional or diffuse mildly echogenic areas in a cystic mass. Fibrin strands in a hemorrhagic cyst can mimic the dermoid mesh. Amongst the three findings, the presence of a shadowing echodensity is the most accurate single feature enabling diagnosis of the mass as a cystic teratoma, with a likelihood ratio for dermoid exceeding 40 when only this feature is identified. However, these findings become pathognomonic for identifying a cystic teratoma when they are seen in combination [10]. In our study [10] comparing cystic teratomas to nondermoid adnexal masses, 55 (76%) of the 72 dermoids exhibited two or more of these sonographic features, whereas none of the 178 nondermoids exhibited this feature set, resulting in a likelihood ratio exceeding 135. Thus, when sonologists identify any one of these sonographic findings, they can become highly confident regarding the diagnosis of cystic teratoma by carefully scrutinizing the mass for the presence of either of the other associated features [Fig. 20].

Benign and malignant cystic neoplasms (cystadenomas and cystadenocarcinomas)


The preceding discussion has analyzed some features of an adnexal mass that might enable the sonologist to characterize the mass as a benign neoplasm, namely the unilocular smooth-walled cyst

that is enlarging on follow-up examination. For these masses, the demonstration of interval growth is the key feature that enables the sonologist to confidently assert that the mass is not a typical non-neoplastic cyst, which would be expected to exhibit stability or resolution over time [see Fig. 8]. As noted previously, theoretically there may be an atypical non-neoplastic cyst that shows growth over time, at a rate that is indistinguishable from a benign neoplasm; nevertheless, trying to distinguish between a growing non-neoplastic cyst and a growing benign ovarian cystic neoplasm is usually not clinically relevant. Both masses merit surgical removal, and it is reasonable and practical to assume that the mass is a benign cystic neoplasm. In this situation, further testing with serologic evaluation or MRI is unlikely to provide compelling additional data to distinguish between the rare enlarging non-neoplastic cyst and the benign neoplasm. The presence of septations or mural or septal nodules in a cystic mass is compelling evidence that the mass is an ovarian neoplasm (benign or malignant) [Figs. 21 and 22] [29]. Benignancy is favored over malignancy when septations are smooth and relatively thin, and when the mural or septal nodularity is minor. Thick septations, irregular solid areas, poorly defined margins, and coexisting ascites or matted bowel loops are features highly specific but not very sensitive for malignancy [30]. As described previously, after considering and excluding the possibility that the aberrant findings could be reasonably caused by a physiologic process, the sonologist confronted with a mass exhibiting septations or mural nodularity can confidently indicate that the mass is a

Fig. 20. Cystic teratoma. Sagittal transabdominal sonogram of an adnexal mass shows a large regional area of high-amplitude echoes. Additional scrutiny reveals a shadowing echodensity (arrows mark borders of the acoustic shadow). The presence of two features of cystic teratomas virtually assures that the mass is a dermoid.

354

Patel

tional tests might be useful if the results would lead to a different surgical approach (or different surgeon) when the patient goes to surgery. Doppler analysis of the resistive index of flow in the wall of the mass has not been shown to have any reproducible accuracy in distinguishing benign from malignant cystic ovarian neoplasms [31] and cannot be recommended as a high-utility maneuver.

Other potentially characteristic adnexal masses


Other categories of adnexal pathology that have not yet been discussed are beyond the scope of this article. In some cases, there are characteristic features (both clinical and imaging) that enable confident classification of an adnexal masses into one of these categories, which include ectopic pregnancy, abscess/inflammatory mass, torsed ovary, exophytic leiomyoma, ovarian fibroma, and other solid mass.

Fig. 21. Cystadenocarcinoma. Sonogram of the right adnexa reveals a cystic mass with large wall vegetations. Color Doppler evaluation showed obvious blood flow within the nodules (not shown). Surgical treatment and staging was performed by a gynecologic oncologist without additional imaging work-up.

neoplasm. He or she may not be sure if the mass is benign or malignant in such cases; benignancy or malignancy may be favored based on the specific features of the case, but the sonologist may be confident of this assessment in only the most straightforward cases. In keeping with the basic approach, the question to consider after identifying a neoplasm is whether any additional testing should be performed, recommended, or offered. The answer depends on whether it makes any difference. Serologic evaluation of CA-125 levels may help to implicate malignancy over benignancy. MRI may help to better demonstrate malignant-appearing features. These addi-

Further imaging or surgical exploration


Using the algorithm as a practical approach to the adnexal mass, the sonologist should be able to confidently classify many, if not most, adnexal masses into one of the categories of suspected pathology: (1) non-neoplastic cyst; (2) hemorrhagic ovarian cyst; (3) endometrioma; (4) hydrosalpinx; (5) peritoneal inclusion cyst; (6) benign cystic neoplasm; (7) cystic teratoma; (8) cystic malignancy; (9) ectopic pregnancy; (10) abscess/inflammatory mass; (11) torsed ovary; (12) exophytic or broad ligament myoma; (13) ovarian fibroma; (14) solid mass. When the sonographic findings do not allow confidence in establishing one category, further

Fig. 22. Cystadenocarcinoma. Transabdominal sonogram of the right adnexa shows a large cystic and solid mass (demarcated with electronic calipers), with multiple septations, some harboring large septal nodules. The sonologist can confidently diagnose malignancy based on these features.

Practical Approach to the Adnexal Mass

355

testing with MRI can lead to additional observations that enable categorization. The endometrioma with apparently avascular wall nodularity, the atypical hydrosalpinx without sonographically identified adjacent ovary, and the peritoneal inclusion cyst in which the suspended ovary is not visible are some examples of masses that might benefit from further evaluation with MRI, since one might choose to not surgically evaluate the patient if one could be reasonably assured that the mass was an endometrioma, hydrosalpinx, or peritioneal pseudocyst. Similarly, establishing the diagnosis of exophytic leiomyoma or ovarian fibroma using MRI in those cases without characteristic sonographic observations would be useful to avoid surgical evaluation. Ultimately, if imaging is unable to characterize a mass into one of these categories using the suggested algorithm, diagnostic surgical exploration will be necessary.

[7]

[8]

[9]

[10]

[11]

Summary
Gynecologic sonography has matured into a highly effective and accurate tool enabling confident diagnosis of a variety of adnexal masses. Using a practical evidence-based approach, sonologists are well equipped to differentiate expected findings in the normal ovary from pathologic entities and can often generate specific conclusions regarding the cause of an adnexal mass. Mastery of the diagnostic strategies to use when an adnexal mass is identified and the sonographic patterns of various types of adnexal pathology contributes greatly to the proper and cost-effective care of a woman with an adnexal mass.

[12]

[13]

[14]

[15]

References
[1] Bakos O, Lundkvist O, Wide L, Bergh T. Ultrasonographical and hormonal description of the normal ovulatory menstrual cycle. Acta Obstet Gynecol Scand 1994;73:7906. [2] Ron-El R, Nachum H, Golan A, et al. Binovular human ovarian follicles associated with in vitro fertilization: incidence and outcome. Fertil Steril 1990;54:86972. [3] Pache TD, Wladimiroff JW, de Jong FH, et al. Growth patterns of nondominant ovarian follicles during the normal menstrual cycle. Fertil Steril 1990;54:63842. [4] Swire MN, Castro-Aragon I, Levine D. Various sonographic appearances of the hemorrhagic corpus luteum cyst. Ultrasound Q 2004;20:4558. [5] Timor-Tritsch IE, Goldstein SR. The complexity of a complex mass and the simplicity of a simple cyst. J Ultrasound Med 2005;24:2558. [6] Timmerman D, Schwarzler P, Collins WP, et al. Subjective assessment of adnexal masses with the [16]

[17]

[18]

[19]

[20]

use of ultrasonography: an analysis of interobserver variability and experience. Ultrasound Obstet Gynecol 1999;13:116. Valentin L, Hagen B, Tingulstad S, Eik-Nes S. Comparison of pattern recognition and logistic regression models for discrimination between benign and malignant pelvic masses: a prospective cross validation. Ultrasound Obstet Gynecol 2001;18:35765. Patel MD, Feldstein VA, Chen DC, et al. Endometriomas: diagnostic performance of US. Radiology 1999;210:73945. Patel MD, Feldstein VA, Filly RA. The likelihood ratio of sonographic findings for the diagnosis of hemorrhagic ovarian cysts. J Ultrasound Med 2005;24:60715. Patel MD, Feldstein VA, Lipson SD, et al. Cystic teratomas of the ovary: diagnostic value of sonography. AJR Am J Roentgenol 1998;171:10615. Patel MD, Acord DL, Young SW. Likelihood ratio of sonographic findings in discriminating hydrosalpinx from other adnexal masses. AJR 2006;186. Nardo LG, Kroon ND, Reginald PW. Persistent unilocular ovarian cysts in a general population of postmenopausal women: is there a place for expectant management? Obstet Gynecol 2003; 102:58993. Modesitt SC, Pavlik EJ, Ueland FR, et al. Risk of malignancy in unilocular ovarian cystic tumors less than 10 centimeters in diameter. Obstet Gynecol 2003;102:5949. Dorum A, Blom GP, Ekerhovd E, Granberg S. Prevalence and histologic diagnosis of adnexal cysts in postmenopausal women: an autopsy study. Am J Obstet Gynecol 2005;192:4854. Ekerhovd E, Wienerroith H, Staudach A, Granberg S. Preoperative assessment of unilocular adnexal cysts by transvaginal ultrasonography: a comparison between ultrasonographic morphologic imaging and histopathologic diagnosis. Am J Obstet Gynecol 2001;184:4854. Mesogitis S, Daskalakis G, Pilalis A, et al. Management of ovarian cysts with aspiration and methotrexate injection. Radiology 2005;235: 66873. Timmerman D, Bourne TH, Tailor A, et al. A comparison of methods for preoperative discrimination between malignant and benign adnexal masses: the development of a new logistic regression model. Am J Obstet Gynecol 1999; 181:5765. Korbin CD, Brown DL, Welch WR. Paraovarian cystadenomas and cystadenofibromas: sonographic characteristics in 14 cases. Radiology 1998;208:45962. Ghossain MA, Braidy CG, Kanso HN, et al. Extraovarian cystadenomas: ultrasound and MR findings in 7 cases. J Comput Assist Tomogr 2005;29:749. Bass IS, Haller JO, Friedman AP, et al. The sonographic appearance of the hemorrhagic

356

Patel

[21]

[22]

[23]

[24]

[25]

ovarian cyst in adolescents. J Ultrasound Med 1984;3:50913. Baltarowich OH, Kurtz AB, Pasto ME, et al. The spectrum of sonographic findings in hemorrhagic ovarian cysts. AJR Am J Roentgenol 1987; 148:9015. Okai T, Kobayashi K, Ryo E, et al. Transvaginal sonographic appearance of hemorrhagic functional ovarian cysts and their spontaneous regression. Int J Gynaecol Obstet 1994;44:4752. Jain KA, Friedman DL, Pettinger TW, et al. Adnexal masses: comparison of specificity of endovaginal US and pelvic MR imaging. Radiology 1993;186:697704. Sassone AM, Timor-Tritsch IE, Artner A, et al. Transvaginal sonographic characterization of ovarian disease: evaluation of a new scoring system to predict ovarian malignancy. Obstet Gynecol 1991;78:706. Tessler FN, Perrella RR, Fleischer AC, Grant EG. Endovaginal sonographic diagnosis of dilated fallopian tubes. AJR 1989;153:5235.

[26] Timor-Tritsch IE, Lerner JP, Monteagudo A, Murphy KE, Heller DS. Transvaginal sonographic markers of tubal inflammatory disease. Ultrasound Obstet Gynecol 1998;12:5666. [27] Jain KA. Imaging of peritoneal inclusion cysts. AJR Am J Roentgenol 2000;174:155963. [28] Malde HM, Kedar RP, Chadha D, Nayak S. Dermoid mesh: a sonographic sign of ovarian teratoma. AJR Am J Roentgenol 1992;159: 134950. [29] Granberg S, Wikland M, Jansson I. Macroscopic characterization of ovarian tumors and the relation to the histological diagnosis: criteria to be used for ultrasound evaluation. Gynecol Oncol 1989;35:13944. [30] Herrmann Jr UJ, Locher GW, Goldhirsch A. Sonographic patterns of ovarian tumors: prediction of malignancy. Obstet Gynecol 1987;69: 77781. [31] Levine D, Feldstein VA, Babcook CJ, Filly RA. Sonography of ovarian masses: poor sensitivity of resistive index for identifying malignant lesions. AJR Am J Roentgenol 1994;162:13559.

357

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 357383

Use of MR Imaging for Further Evaluation of Sonographically Detected Adnexal Pathology


Liina Poder,
& &

MD*,

Aliya Qayyum,

MD,

Ruth B. Goldstein,

MD

Overview of the role of MR imaging in the evaluation of common adnexal masses Diagnoses with specific MR imaging characteristics Leiomyomas

& &

Endometriosis Dermoids Fibromas Summary References

Adnexal masses are a very common clinical problem in pre- and postmenopausal women. Approximately 5% to 10% of women in the United States will undergo surgery for a suspected adnexal mass, and less than 25% of these women will have a proven malignancy [1]. The age, menopausal status, cancer antigen (CA) 125 levels, and clinical history are important factors to consider when planning further work-up. The recognition of a benign adnexal mass precludes the need for further unnecessary diagnostic procedures, which reduces cost and patient anxiety, and does not necessitate referral to a gynecologic oncologist (see the article by Patel elsewhere in this issue). Increasingly, removal of benign lesions is conducted by way of surgical laparoscopy, with minimal surgical morbidity, which unarguably depends on proper imaging diagnosis [2]. Furthermore, expedited referral of patients who have malignant lesions to a gynecologic oncologist has been shown to correlate with improved survival [1]. An explanation for the improved outcome is the considerable difference in surgical approach to a benign adnexal process compared with a malignant adnexal lesion, which requires a full-staging laparotomy including strip-

ping of the diaphragm with attempted optimal surgical cytoreduction of all tumor sites [1,3]. Women who are clinically at low risk for malignancy and have sonographically indeterminate adnexal lesions most often benefit from further imaging with MR imaging. Sonography (typically endovaginal) plays an integral role in evaluation of gynecologic disease, and is indisputably the primary imaging approach and the most practical modality for assessment of ovarian tumors. Sonography is readily available and has a reported sensitivity of 85% to 97% for predicting malignancy in adnexal masses [46]. There is considerable variation in the reported specificity, however, ranging from 39% to 95%, and the reported accuracy is only moderate at approximately 64% [1,4,7]. In most cases, sonographic establishment of benign features of an adnexal mass precludes the need for additional imaging (see the article by Patel elsewhere in this issue) [1]. Imaging features that suggest malignancy with sonography, MR imaging, and CT are well established and include increased septal and mural thickness (>3 mm), mural nodularity, cystic lesions greater than 4 cm, bilateral adenexal masses, and

Department of Radiology, University of California San Francisco, 505 Parnassus Avenue, L-374, San Francisco, CA 94143-0628, USA * Corresponding author. E-mail address: liina.poder@radiology.ucsf.edu (L. Poder).
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.02.002

ultrasound.theclinics.com

358

Poder et al

solid masses with or without necrosis. Endovaginal sonography is excellent for visualization of fine septations, mural irregularities, and often vascularity within the mural nodules, thereby discriminating simple, benign-appearing cystic masses from malignant masses [810]. Although solid masses in general are suggestive of a malignant etiology, not all solid masses are malignant, and in situations in which there is uncertainty regarding the nature of the adnexal mass or its origin, MR imaging may be used as a problem-solving modality. In women of reproductive age, MR imaging is typically requested for further characterization of masses such as endometriomas, dermoids, or degenerating extrauterine leiomyomas because these may not demonstrate overtly benign sonographic features. A further example is the presence of a retracting clot or internal debris that may mimic a mural nodule within a cystic mass. Doppler interrogation may not always enable confident determination of the vascularity of small nodules. In such cases, MR imaging can be used to readily characterize the tissue as blood products or confirm true vascularity by demonstrating enhancement of the nodule with intravenous gadolinium. MR imaging generally offers little added benefit in an acute setting, and acute pelvic pathology is not discussed in this article. One important exception, however, is appendicitis in a pregnant patient, whereby MR imaging can be used and is very revealing. CT is typically requested for tumor staging or to confirm the presence of calcium or fat within a mass (such as a dermoid) following sonography. Although CT is not generally recommended for adnexal mass characterization, it is possible recognize features of benign versus malignant disease [2,11,12]. In this article, the authors describe the sonographic and MR imaging features of adnexal masses detected with sonography in which MR imaging may be of incremental benefit and discuss a practical radiologic approach to cross-sectional imaging and patient management.

is not recommended as a first-line modality for imaging the adnexa because it is expensive and sonography is sufficient in most instances. From a practical perspective, there a few important technical considerations with respect to MR imaging technique: (1) MR imaging should be performed with a phased-array surface coil to optimize signal to noise; (2) an antiperistaltic agent such as glucagon may be used to reduce motion artifact; (3) application of anterior and posterior fat-saturation bands reduces respiratory artifact from subcutaneous fat that can seriously degrade image quality; and (4) a small field of view (2024 cm) should be used. Images should be obtained in at least two different planes, typically axial and sagittal. The T2-weighted sequences are crucial for delineating the anatomy of the female pelvis, and T2- and T1-weighted sequences are necessary for accurate tissue characterization [7,10,1517]. Leiomyomas and fibrous tumors demonstrate low signal intensity on T2-weighted images, unlike solid malignancies, which are of increased signal intensity. Leiomyomas may be distinguished from fibrous tumors because they demonstrate greater enhancement on gadolinium-enhanced T1-weighted images. Cystic lesions are readily identified by the characteristic signal intensity of their fluid content. Hemorrhage associated with endometriomas is recognized by its low signal intensity on T2-weighted images and high signal intensity on T1-weighted images. Fat is also of high signal intensity on T1-weighted images but may be distinguished from blood products by its decrease in signal intensity with the application of fat suppression to the T1-weighted sequence.

Diagnoses with specific MR imaging characteristics Leiomyomas


Leiomyomas (fibroids) are the most common neoplasms of the uterus and reported to occur in 20% to 30% of women over age 30 years [18,19]. These tumors are usually multiple, cause enlargement of the uterus, and may be symptomatic, most commonly causing pelvic pain and dysfunctional uterine bleeding. Although leiomyomas are benign tumors, a rapid increase in size, particularly in postmenopausal patients, raises the possibility of sarcomatous change [20]. Leiomyomas are readily recognized at sonography but are variable in appearance. Localized leiomyomas appear hypoechoic but may be heterogeneous. Many leiomyomas demonstrate acoustic attenuation or a venetian blind type of shadowing, occasionally without a discrete mass, making it difficult

Overview of the role of MR imaging in the evaluation of common adnexal masses


MR imaging is well established as a modality that enables accurate diagnosis of leiomyomas, endometriomas, fibromas, and cysts. MR imaging provides valuable information on the composition of soft tissue masses by using differences in tissue relaxation times and has been shown to have a sensitivity, a specificity, and an accuracy of 85% to 96.6%, 83.7% to 96%, and 88% to 93%, respectively, when differentiating benign from malignant adnexal masses [1,13,14]. In general, MR imaging

MR Imaging of Adnexal Pathology

359

to assess the lesion size. This characteristic shadowing is helpful in distinguishing pedunculated or exophytic leiomyomas from other solid adnexal masses; however, at times this differentiation remains a challenge. Leiomyomas may contain coarse calcifications and commonly undergo degeneration.

Degeneration may be cystic or hemorrhagic, which may lead to a misdiagnosis of malignancy, particularly if the leiomyoma is exophytic or pedunculated [Fig. 1] [21,22]. MR imaging is helpful in further evaluating leiomyomas for three reasons. First, when a mass is

Fig. 1. A 32-two-year-old woman had pelvic pain and left adnexal/midline fullness on physical examination. Endovaginal sonogram sagittal view (A) and coronal view with Doppler flow (B) images revealed a 5.3 3.5 4.2-cm heterogenous mass (*) probably arising from the region of uterine fundus. Attachment to the uterus, however, was not confidently delineated, and typical acoustic attenuation or shadowing was not present and no internal vascularity was noted with color Doppler (B). Follow-up sagittal T2 MR image (C ) demonstrates high signal intensity within the mass (*), suggestive of degenerative fibroid. Claw sign (arrowheads) supports its origin from the uterus. Sagittal postcontrast MR image (D) shows peripheral intense enhancement (arrowheads) similar to myometrium (arrows) and no central enhancement (*), further confirming the diagnosis of a fundal exophytic degenerating fibroid. Both ovaries were confirmed separate from this mass (not shown).

360

Poder et al

Fig. 1 (continued ).

suspected to be a leiomyoma but is not confidently diagnosed, MR imaging may be helpful. Simple and degenerating leiomyomas have very typical appearances. The MR imaging characteristics of a simple, nondegenerating leiomyoma are a rounded contour, low signal intensity on T1- and T2-weighted sequences, well-circumscribed margins, and a characteristic high signal intensity pseudocapsule on T2-weighted sequences that may not always be apparent but is diagnostic [19,2325]. The enhancement pattern with gadolinium can be variable. Most leiomyomas enhance slightly less or similar to the surrounding myometrium, although avid enhancement is frequently seen and may be used to help distinguish pedunculated, broadligament, or ovarian leiomyomas from fibromas or malignant ovarian neoplasms, which generally only demonstrate slight enhancement [19,20,25]. Degenerating fibroids demonstrate areas of high signal intensity on T2-weighted sequences [Fig. 2] [26]. The types of degeneration are usually indistinguishable at MR imaging. Hyaline degeneration may demonstrate a cobblestone appearance, with areas of increased signal on T2-weighted sequences, and hemorrhagic or red degeneration may exhibit high signal intensity on T1-weighted sequences and

is associated with a lack of gadolinium enhancement secondary to the interruption of blood flow [19]. Hemorrhagic degeneration represents massive hemorrhage and infarction, often as a consequence of rapid leiomyoma enlargement that may occur under hormonal stimulation during pregnancy, and has been associated with use of oral contraceptives. Hemorrhagic degeneration can also be seen as a consequence of fibroid embolization [Fig. 3]. Cystic degeneration occurs in 4% of leiomyomas and is associated with large lesion size [19]. Cystic degeneration is recognized as a well-defined region of fluid signal intensity within the leiomyoma (very high signal intensity on T2-weighted and low signal intensity on T1-weighted sequences) [see Fig. 2]. Fatty degeneration is also readily recognized as a region of high signal intensity on T1-weighted sequences, which becomes a region of low signal intensity with the application of fat suppression [22]. A second situation in which MR imaging is helpful in further evaluating leiomyomas is when the connection of the mass to the uterus (confirming it as a leiomyoma) cannot be confidently determined with sonography. At least 90% of leiomyomas occur in the uterus and approximately

Fig. 2. A 33-year-old woman had pelvic pain and an enlarged uterus on clinical examination. Longitudinal (A) and transverse (B) transabdominal sonogram images revealed a large pelvic mass with two cystic areas (*) suggestive of degenerating fibroid, although not conventional in appearance. The uterus (u) appears deep to the mass. Follow-up sagittal (C ) and axial (D) T2-weighted MR images confirm a large anterior fundal fibroid (arrowheads) measuring 11.1 7.6 7.7 cm, demonstrating two large areas of high T2 signal (*) consistent with cystic degeneration.

MR Imaging of Adnexal Pathology

361

362

Poder et al

Fig. 3. A 47-year-old woman had degeneration of fibroid and evaluation before fibroid embolization. MR imaging demonstrates typical appearance of infarcted fibroid/red degeneration. Sagittal (A) and axial (B) T2-weighted MR images demonstrate very low T2 signal (*) with hemosiderin rim (arrowheads). Axial T1 MR image (C ) shows T1 signal brighter than expected for a typical fibroid (*) and no enhancement on sagittal postgadalinium image (D) (*).

5% occur in the cervix, whereas leiomyomas of the broad ligament or adenexa are rare [27]. Exophytic, pedunculated, or broad ligament leiomyomas may be misdiagnosed as a solid ovarian mass at sonography because the origin of the lesion may be difficult to determine [Fig. 4]. Third, MR imaging is helpful in confirming the diagnosis of very large leiomyomas. Such lesions demonstrate intense shadowing at sonography, which may make establishment of diagnosis or lesion size virtually impossible. MR imaging also helps to evaluate already-known, very large leio-

myomas for interval change in size or morphology [Figs. 5 and 6].

Endometriosis
Endometriosis is defined as functional endometrial tissue located outside the endometrial canal. The ectopic endometrium is hormonally responsive, affects women of reproductive age, and may cause dysmenorrhea, dyspareunia, and infertility. Endometriosis involves the ovaries, fallopian tubes, broad ligament, and posterior cul-de-sac, with decreasing order of frequency [4,28]. Rarely, deposits

MR Imaging of Adnexal Pathology

363

Fig. 4. A 59-year-old postmenopausal woman had recurrent pelvic pain and bloating. Endovaginal sonogram coronal (A and B) and sagittal (C ) images revealed a 3.7 3.1 3.1-cm hypoechoic left adnexal mass (electronic calipers, *) in close contact with uterus (u); left ovary was not identified separately. Normal-appearing right ovary was visualized (not shown). Follow-up T2-weighted sagittal MR image (D) demonstrate low heterogeneous signal (*), with claw sign from the uterus (arrowheads) suggesting an exophytic mass originating from the uterus. Axial pregadolinium image (E ) demonstrates low T1 signal (*) and postgadolinium image (F ) demonstrates low T1 signal with avid enhancement (*), confirming the diagnosis of a left exophytic uterine fibroid.

364

Poder et al

Fig. 4 (continued ).

may be found in remote parts of the abdomen, pelvis, or even the chest. Endometriosis is a disease spectrum that extends from the localized form (otherwise known as an endometrioma) to diffuse endrometriosis characterized by multiple small (23 mm) implants that may not be visualized with imaging. Endometriomas have a variable but characteristic sonographic appearance. Hemorrhage is often recurrent, resulting in blood products of various stages. Sonographically, endometriomas are usually seen as well-defined, multi-, or unilocular complex cystic masses filled with uniform low-level echoes. Occasionally, fluidfluid levels can be observed. The wall of the lesion is avascular but may be nodular. Small hyperechoic foci are often noted in the periphery of the lesion, which represent cholesterol deposits accumulated in the cyst wall.

Similarities between an endometrioma and a complex hemorrhagic cyst can cause confusion, although the clinical history and interval follow-up typically clarify these issues because hemorrhagic cysts, unlike endometriomas, resolve or at least involute over subsequent menstrual cycles. A more troubling diagnostic dilemma can arise when an endometrioma demonstrates a thickened wall or septations on the sonogram, which may lead to a misdiagnosis of malignancy. MR imaging may be used to confirm the diagnosis of endometriosis due to the characteristic appearance of the blood products. The sensitivity of MR imaging for the detection of lesions greater than 1 cm in diameter has been reported to be 82%, although the use of fat-suppressed T1-weighted sequences is thought to improve detection [29]. The presence of large amounts of blood products within

MR Imaging of Adnexal Pathology

365

endometriomas results in regions of high T1 signal intensity, which are maintained with the application of fat suppression. This MR imaging feature is key to the diagnosis of endometriosis. A further helpful MR imaging appearance is the internal layering of the blood products on T2-weighted sequences, which results in a fluidfluid layering of low and high signal intensity known as T2 shad-

ing [4,30,31]. In addition, a peripheral low signal hemosiderin rim due to chronic hemorrhage may be seen. The fibrous wall or septations are of low signal intensity on T1- and T2-weighted sequences and may show some enhancement with gadolinium, but the central contents of the lesion reportedly do not enhance [18,32]. Endometriosis may also involve the fallopian tubes, resulting in obstruc-

Fig. 5. A 29-year-old woman had dysfunctional uterine bleeding and pelvic fullness on physical examination. Sagittal endovaginal (A), transverse transabdominal (B), and transverse endovaginal with Doppler (C ) sonographic images revealed an enlarged uterus with a large posterior transmural mass (*), with central cystic changes (arrowhead in B and C ) measuring at least 12.6 10.1 11 cm, although the entire mass is not completely captured by endovaginal or transabdominal sonography. Mass effect (arrow in A) and distortion of the stripe by this mass (arrowheads in A) were noted. Due to the large size and atypical appearance, an MR imaging was obtained. (D) Sagittal T2 image demonstrates a large posterior fundal mass with mixed but predominantly high T2 signal intensity (*); claw sign (arrowheads) confirms uterine origin. Mass effect on endometrial canal (arrow) was confirmed. (E ) Axial T1 image shows characteristic low T1 signal (*). (F ) Avid enhancement on sagittal postgadolinium image (arrowheads), similar to myometrium, is noted. Areas that do not enhance are consistent with areas of necrosis/degeneration (*). Due to heavy bleeding, patient had transabdominal myomectomy; pathology confirmed a degenerating fibroid.

366

Poder et al

Fig. 5 (continued ).

MR Imaging of Adnexal Pathology

367

tion and dilatation. The tubular anatomy may be difficult to discern at sonography but may be better delineated at MR imaging with the use of multiplanar imaging and a larger field of view compared with sonography [Fig. 7]. Although endometriosis is generally a benign condition, it is associated with a small increase in gynecologic malignancy. A study from Sweden in 1997 reported that the risk of ovarian cancer was 4.2% in the presence of endometriosis [3335].

Clear-cell and endometrioid carcinomas are the most common malignancies to occur in the rare instances of cancer arising from ovarian endometriosis, whereas extragonadal tumors are usually endometriod tumors and sarcomas [34]. The identification of enhancing nodules or a solid mass within an endometriotic lesion favors the development of malignancy [Fig. 8] [33]. Another entity to be aware of is desidualization of ovarian endometriosis. Desidualization is a change

Fig. 6. A postmenopausal female had a history of stable degenerating fibroid. Longitudinal (A) and transverse with Doppler (B) transabdominal sonographic images, with Doppler performed in 2004, demonstrated interval increase in the size of the suspected cystic degenerating fibroid (*); no suspicious mural nodularity on ultrasound was detected. Increase in size triggered further evaluation with MR imaging. Follow-up sagittal (C ) and axial (D) T2 images demonstrate high T2 signal consistent with degeneration (*). Axial T1 preimage (E ) and postimage (F ) demonstrate a small enhancing mural nodule (arrowhead) that on T2 images (arrowhead) demonstrates low signal and was believed to represent residual leiomyomatous tissue. Sagittal (G) and transverse (H ) transabdominal sonogram 1 year later demonstrated further increase in size in the cystic component (arrowheads) and, most important, interval increase in size of the mural nodule (electronic calipers, *). Repeat MR imaging was performed following the sonogram done in (G) and (H ). Sagittal (I ) and axial T2 (J ) images confirmed interval growth and avid enhancement (K ) of the mural nodule (*). Surgical removal followed, and pathology revealed low-grade sarcoma without myometrial invasion.

368

Poder et al

Fig. 6 (continued ).

MR Imaging of Adnexal Pathology

369

Fig. 6 (continued ).

370

Poder et al

Fig. 7. A 42-year-old woman who had muscular dystrophy presented with pelvic fullness and was found to have right adnexal mass by palpation. Longitudinal (A) and transverse (B) endovaginal sonographic images with Doppler (C ) examination revealed a 9.0 10-cm avascular cystic structure (*) containing low-level echoes, bright reflections, and an incomplete fold or septum (arrow). Given the appearance, a neoplastic process was included in the differential diagnosis. Follow-up axial T1 with fat saturation (D) and axial T2-weighted (E) MR images confirmed characteristic hyperintense T1 and hypointense T2 signal, characteristic of blood products within this right adnaxal mass (*). Increase in size and symptoms were noted on follow-up and the lesion was subsequently removed. Pathology confirmed endometriotic cyst.

MR Imaging of Adnexal Pathology

371

Fig. 7 (continued ).

372

Poder et al

Fig. 8. A 37-year-old woman had pelvic discomfort. Transverse (A) and Doppler (B) transabdominal sonogram images showed a left adnexal mass with low-level echoes, without internal flow (*). To improve fertility, left adnexal mass was removed by way of laparoscopy, and endometrioma was confirmed. Multiple in vitro fertilization attempts followed; a recurrent large cyst in the left ovary was noted during intraoperative egg retrieval. MR imaging was obtained after laparoscopy. Axial T1 (C ) and T1 fat-saturated (D) images demonstrate blood products in this left adnexal mass, consistent with a recurrent endometrioma (*). Two mural nodules were noted (arrowheads). In addition, axial T1 fat-saturated (E) image demonstrated right ovarian small endometrial implants (arrowheads). (F ) Enhancement of one of the mural nodules on this axial image, postgadolinium (arrow). As discussed previously, an enhancing nodule in endometrioma would be highly concerning for malignancy. CA 125 level was 209. Pathology demonstrated serous borderline tumor that was surgically removed. Nine months later, the patient returned with recurrence and metastatic disease.

MR Imaging of Adnexal Pathology

373

Fig. 8 (continued ).

374

Poder et al

Fig. 9. A 29-nine-year-old woman had early gestation and incidentally noted complex left adnexal mass on first trimester sonogram. Sagittal (A) and transverse (B) with Doppler (C ) transabdominal sonogram images demonstrated a solid and cystic mass with internal low-level echoes measuring at least13.9 cm (*). CA 125 level was elevated (equivocal in the setting of pregnancy). The solid material was believed to be secondary to blood of different stages, but malignancy could not be excluded. (D) Transvaginal sonogram demonstrated a smaller lesion (*) with more typical low-level echoes suggestive of endometrioma is noted in the right ovary. Follow-up coronal T2 (E) MR image reveals a gravid uterus (arrow) and a tubular hyperintense but heterogeneous structure (*) originating from the left adnexa, abutting the gravid uterus and extending into the cul-de-sac. Follow-up axial T1 (F ) and T2 (G) MR images redemonstrate the hyperintense left adnexal tubular structure (*) containing blood products of various stages. The tubular morphology suggests involvement of the fallopian tube, although ovarian follicles in the periphery were noted (not shown), indicating also involvement of the left ovary. In addition, a small, more typical (high T1/low T2 signal) endometrioma was demonstrated on the right ovary (arrowhead in F and G). Although the MR imaging appearance was benign, surgery was performed in the third trimester to avoid labor dystocia and possible rupture of the mass. Pathology of bilateral adnexa confirmed endometriosis with associated decidual reaction. Biopsy of omentum revealed endometriosis with decidual reaction.

MR Imaging of Adnexal Pathology

375

Fig. 9 (continued ).

376

Poder et al

Fig. 9 (continued ).

that takes place in normal endometrium, under hormonal influence, to optimally accommodate the gestation. Formation of ectopic deciduas (deciduosis) is a well-confirmed phenomenon that has been attributed to similar hormonal effects not only on ectopic endometrium (such as endometriosis) but also on normal subceolomic mesothelium [36]. Deciduosis of ectopic endometrial tissue is a rare but important pitfall that may mimic malignant change in endometriosis in pregnant women [37] because it is associated with nodularity of the wall of the endometriotic lesion. These papillary excrescencies demonstrate intense vascularity on sonographic images, and enhancement at MR imaging is not typically evaluated because intravenous contrast is generally not recommended in pregnancy. Deciduosis is usually asymptomatic and typically presents as an incidental finding during first or second trimester sonograms, although cases with intraperitoneal hemorrhage and deciduosis of the appendix presenting as an acute appendicitis have been described [37]. This phenomenon is rare but certainly worth knowing about because imaging in pregnancy, as is treatment, is often limited and controversial. In most reported cases so far, the treatment has been surgical removal, but awareness of this benign identity may guide the surgical planning [Fig. 9]. Tumor markers conventionally used to evaluate worrisome pelvic masses are not helpful because they are physiologically elevated in the gravid state [33,37].

formation of a dermoid is very rare, occurring in only 2% of cases and usually in the older patient population. Sonographically, dermoids have a variable appearance, although certain features are considered specific, such as an echogenic mural nodule or dermoid plug, which usually contains hair, teeth, or fat. The dermoid plug casts a typical acoustic shadow due to the hair fibers, resulting in multiple linear hyperechoic foci that appear to float within the cyst, causing a dot-dash or dermoid mesh pattern. The acoustic shadowing from the echogenic focus may obscure the posterior wall of the cystic mass, referred to as the tip-of-theiceberg sign. Diagnostic dilemmas may arise at sonography if the dermoid does not demonstrate typical features or is very large or very small. A large dermoid may distort the normal pelvic anatomy or obscure other pelvic structures. A small dermoid may not demonstrate the specific sonographic features. Small dermoids can mimic endometriomas on sonography and vice versa. When fertility is of concern, it is important to make the distinction. Even small endometriomas are commonly removed to improve fertility. MR imaging is an easy and preferred technique for differentiation of small dermoids and endometriomas. Furthermore, it may be difficult to distinguish the hyperechoic appearance of the fat within dermoids from bowel, which can result in failure to identify the lesion [38]. Generally, a noncontrast pelvic CT is sufficient to clarify the diagnosis because the presence of fat and calcification within the mass is readily identified. The typical features of a dermoid at MR imaging include a high signal intensity on T1-weighted sequences and an intermediate signal intensity on T2-weighed sequences due to the lipid-laden cyst fluid. Fat suppression allows differentiation from blood products on T1-weighted images. Another feature of dermoids is the presence of chemical shift artifact at the fatfluid interface on opposedphase T1-weighted gradient echo images [4]. The only other ovarian mass to demonstrate the presence of fat is an immature teratoma. Immature teratomas are usually very large at presentation and characterized by a prominent solid component containing coarse calcifications and only a small focus of fat [Fig. 10] [4].

Fibromas Dermoids
Mature cystic teratomas or dermoids account for 15% to 25% of ovarian neoplasms, and up to 15% are bilateral. Dermoids are usually found on gynecologic examination as a palpable adnexal mass. They are often asymptomatic, but complications include torsion and rupture. Malignant transFibromas are the most common sex cord tumors, accounting for 4% of all ovarian neoplasms. Fibromas are benign lesions that occur in pre- and postmenopausal women and often detected as a solid palpable mass on gynecologic examination. Fibromas may be misdiagnosed on the sonogram as exophytic fibroids or ovarian cancer because they

MR Imaging of Adnexal Pathology

377

Fig. 10. (A) Sagittal transabdominal sonogram demonstrates a dermoid that has typical features that are easy to recognize on sonography: dermoid mesh (dot-dash) (arrows) and dermoid plug (*). (B) In this patient, who had a more atypical appearance of a dermoid, a sagittal transabdominal sonogram demonstrates a highly echogenic and sound beam attenuating mass (*) in the right ovary, and follicles were noted in the periphery (arrowheads). (C ) Sagittal endovaginal image of right adnexa demonstrates another unusual-appearing dermoid with fat lobules (*). Dermoid mesh is less obvious but present (arrows). (D, E) Two separate but similar echogenic ovarian lesions (*,D) with differential diagnosis of enodometrioma versus dermoid cyst. This setting is an appropriate one in which to use MR imaging for confirmation of diagnosis. Lesions proved to be endometrioma (D) and dermoid (E) on MR imaging (not shown). (F ) Typical appearance of dermoid using CT. Axial image confirms presence of fat (*). (G, H, I ) An example of typical appearance and signal characteristics of a dermoid cyst on MR imaging. Axial T1 (G), axial T2 (H), axial T1 fat-supressed (I ), images confirm high T1/high T2 and diagnostic low signal on T1 fatsuppressed image in this right adnexal mass (*), in keeping with a characteristic appearance of a dermoid cyst.

378

Poder et al

Fig. 10 (continued ).

MR Imaging of Adnexal Pathology

379

Fig. 10 (continued ).

380

Poder et al

Fig. 11. A 78-year-old woman had a palpable right adnexal lump. Sagittal (A) and transverse (B) endovaginal sonogram images revealed a solid mass (D) with strong posterior acoustic shadowing (*). Ovaries were not seen separately. Follow-up axial T1 (C), T2 (D), and postgadolonium (E) MR images demonstrate a strongly hypointense nonenhancing mass, in keeping with a fibroma (*).

MR Imaging of Adnexal Pathology

381

Fig. 11 (continued ).

appear solid, typically hypoechoic, and cause sound beam attenuation. The sonographic appearance may vary because fibromas have also been described as hyperechoic masses with increased through transmission. Another feature of fibromas is the presence of dense calcifications, which can be helpful in suggesting the diagnosis [1,2,39]. Fibromas have a very characteristic MR imaging appearance: low signal intensities on T2- and T1-weighted sequences and slight enhancement on gadolinium-enhanced T1-weighted sequences. The difference in enhancement is helpful in distinguishing fibromas from simple subserosal and pedunculated leiomyomas.

In such cases, the main differential diagnosis is a cystadenofibroma, which contains solid and cystic components [4,18,31]. Ascites is reported to occur in conjunction with fibromas in 40% of cases, particularly if the fibroma is large. Fibromas are also associated with pleural effusions in Meigs syndrome [Fig. 11] [4,18,31]. At sonography, the criteria most useful in predicting ovarian malignancy are lesion size greater than 4 to 6 cm, thickness of the septa greater than 3 mm, papillary projections, nodularity, necrosis, hemorrhage, and masses with solid components; however, the findings of benign and malignant may overlap. Correlation with the patients age and clinical history, CA 125 levels can be helpful (although elevated CA 125 can be caused by fibroids, pregnancy, menstruation, endometriosis, liver disease, and dermoids). Further, in early stages of ovarian malignancy, CA 125 level is elevated in less than 50% of patients [14,40]. At MR imaging, several types of tissue and fluid can be distinguished on the basis of signal characteristics, which helps to narrow the differential diagnosis. Because no signal intensity characteristics specific for malignant epithelial tumors have been described, however, tumors must be distinguished based on morphologic criteria.

Summary
Sonography remains the primary imaging modality for the evaluation of adnexal masses. Sonography and MR imaging demonstrate excellent sensitivity

Fig. 12. Management paradigm for adnexal mass.

382

Poder et al

(100% and 85%99%, respectively) in distinguishing benign from malignant adnexal masses, although MR imaging is reported to have a greater specificity (84%96%) compared with sonography (39%95%). In most cases, sonography can reliably distinguish benign from malignant masses and be used to guide patient management [Fig. 12]. Diagnostic dilemmas can arise when evaluating lesions that are not clearly benign, especially in premenopausal women who are at low risk for ovarian malignancy by age and CA 125 levels [1]. It is usually in this situation that MR imaging is used to compliment sonography. MR imaging is useful for three principle reasons: (1) its tissue characterization may be specific for certain etiologies as described earlier; (2) its multiplanar capabilities facilitate anatomic localization; and (3) its larger field of view compared with sonography is helpful in visualizing the extent of the disease process. In addition, MR imaging and CT can assist in the staging and preoperative surgical planning of ovarian cancerin particular, the identification of patients who may not be candidates for optimal primary surgical cytoreduction [1,4].

[10]

[11]

[12]

[13]

[14]

[15]

[16]

References
[1] Sohaib SA, Mills TD, Sahdev A, et al. The role of magnetic resonance imaging and ultrasound in patients with adnexal masses. Clin Radiol 2005; 60(3):3408. [2] Jung SE, Lee JM, Rha SE, et al. CT and MR imaging of ovarian tumors with emphasis on differential diagnosis. Radiographics 2002;22(6): 130525. [3] Hricak H, Chen M, Coakley FV, et al. Complex adnexal masses: detection and characterization with MR imagingmultivariate analysis. Radiology 2000;214(1):3946. [4] Jeong YY, Outwater EK, Kang HK. Imaging evaluation of ovarian masses. Radiographics 2000;20(5): 144570. [5] Ferrazzi E, Zanetta G, Dordoni D, et al. Transvaginal ultrasonographic characterization of ovarian masses: comparison of five scoring systems in a multicenter study. Ultrasound Obstet Gynecol 1997;10(3):1927. [6] Reles A, Wein U, Lichtenegger W. Transvaginal color Doppler sonography and conventional sonography in the preoperative assessment of adnexal masses. J Clin Ultrasound 1997;25(5):21725. [7] Balachandran A, Iyer RB. Imaging of ovarian cancer. Appl Radiol 2005;34(9):1929. [8] Timor-Tritsch IE, Goldstein SR. The complexity of a complex mass and the simplicity of a simple cyst. J Ultrasound Med 2005;24(3): 2558. [9] Sohaib SA, Sahdev A, Van Trappen P, et al. Characterization of adnexal mass lesions on MR

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

imaging. AJR Am J Roentgenol 2003;180(5): 1297304. Okamoto Y, Tanaka YO, Nishida M, et al. MR imaging of the uterine cervix: imagingpathologic correlation. Radiographics 2003;23(2): 42545 [quiz: 5345]. Kim KA, Park CM, Lee JH, et al. Benign ovarian tumors with solid and cystic components that mimic malignancy. AJR Am J Roentgenol 2004; 182(5):125965. deSouza NM, ONeill R, McIndoe GA, et al. Borderline tumors of the ovary: CT and MRI features and tumor markers in differentiation from stage I disease. AJR Am J Roentgenol 2005; 184(3):9991003. Saini A, Dina R, McIndoe GA, et al. Characterization of adnexal masses with MRI. AJR Am J Roentgenol 2005;184(3):10049. Rieber A, Nussle K, Stohr I, et al. Preoperative diagnosis of ovarian tumors with MR imaging: comparison with transvaginal sonography, positron emission tomography, and histologic findings. AJR Am J Roentgenol 2001;177(1):1239. Siegelman ES, Outwater EK. Tissue characterization in the female pelvis by means of MR imaging. Radiology 1999;212(1):518. Szklaruk J, Tamm EP, Choi H, et al. MR imaging of common and uncommon large pelvic masses. Radiographics 2003;23(2):40324. Tempany CM, Zou KH, Silverman SG, et al. Staging of advanced ovarian cancer: comparison of imaging modalitiesreport from the Radiological Diagnostic Oncology Group. Radiology 2000;215(3):7617. Leyendecker JR. Practical guide to abdominal and pelvic MRI. Philadelphia: Lippincott Williams & Wilkins; 2004. Ueda H, Togashi K, Konishi I, et al. Unusual appearances of uterine leiomyomas: MR imaging findings and their histopathologic backgrounds. Radiographics 1999;19(Spec No):S13145. Salem S, Wilson SR. Gynecologic ultrasound: abnormalities of the myometrium. In: Rumack CM, Wilson SR, Charboneau JW, editors. Diagnostic ultrasound, vol 1. 3rd edition. St. Louis (MO): Elsevier Mosby; 2005. p. 53840. Reddy NM, Jain KA, Gerscovich EO. A degenerating cystic uterine fibroid mimicking an endometrioma on sonography. J Ultrasound Med 2003; 22(9):9736. Murase E, Siegelman ES, Outwater EK, et al. Uterine leiomyomas: histopathologic features, MR imaging findings, differential diagnosis, and treatment. Radiographics 1999;19(5):117997. Ooto H, Nanbu Y, Nonogaki H, et al. Treatment with LH-RH analog, buserelin, for uterine leiomyoma: assessment with MR imaging. Endometriosis Kenkyukai Kaishi 1989;10:2459. Oguchi O, Mori A, Kobayashi Y, et al. Prediction of histopathologic features and proliferative activity of uterine leiomyoma by magnetic resonance imaging prior to GnRH analogue therapy:

MR Imaging of Adnexal Pathology

383

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

correlation between T2-weighted images and effect of GnRH analogue. J Obstet Gynaecol 1995; 21(2):10717. Mittl Jr RL, Yeh IT, Kressel HY. High-signalintensity rim surrounding uterine leiomyomas on MR images: pathologic correlation. Radiology 1991;180(1):813. Karasick S, Lev-Toaff AS, Toaff ME. Imaging of uterine leiomyomas. AJR Am J Roentgenol 1992; 158(4):799805. Silverberg SG, Kurman RJ. Tumors of the uterine corpus and gestational trophoblastic disease. Atlas of tumor pathology. 3rd series. Washington, DC: Armed Forces Institute of Pathology; 1992. Wentz AC. Endometriosis. In: Jones HW, Wentz AC, Burnett LS, editors. Novaks textbook of gynecology. Baltimore (MD): Williams & Wilkins; 1988. p. 30327. Sugimura K, Okizuka H, Imaoka I, et al. Pelvic endometriosis: detection and diagnosis with chemical shift MR imaging. Radiology 1993;188(2): 4358. Yamashita Y, Hatanaka Y, Torashima M, et al. Magnetic resonance characteristics of intrapelvic haematomas. Br J Radiol 1995;68(813): 97985. Togashi K. Ovarian cancer: the clinical role of US, CT, and MRI. Eur Radiol 2003;13(Suppl 4): L87104. Anaf V, Simon P, Fayt I, et al. Smooth muscles are frequent components of endometriotic lesions. Hum Reprod 2000;15(4):76771.

[33] Tanaka YO, Yoshizako T, Nishida M, et al. Ovarian carcinoma in patients with endometriosis: MR imaging findings. AJR Am J Roentgenol 2000;175(5):142330. [34] Varma R, Rollason T, Gupta JK, et al. Endometriosis and the neoplastic process. Reproduction 2004;127(3):293304. [35] Brinton LA, Gridley G, Persson I, et al. Cancer risk after a hospital discharge diagnosis of endometriosis. Am J Obstet Gynecol 1997;176(3): 5729. [36] Fruscella E, Testa AC, Ferrandina G, et al. Sonographic features of decidualized ovarian endometriosis suspicious for malignancy. Ultrasound Obstet Gynecol 2004;24(5):57880. [37] Sammour RN, Leibovitz Z, Shapiro I, et al. Decidualization of ovarian endometriosis during pregnancy mimicking malignancy. J Ultrasound Med 2005;24(9):128994. [38] Bloomfield TH. Benign cystic teratomas of the ovary: a review of seventy-two cases. Eur J Obstet Gynecol Reprod Biol 1987;25(3):2317. [39] Stephenson WM, Laing FC. Sonography of ovarian fibromas. AJR Am J Roentgenol 1985;144(6): 123940. [40] Kurtz AB, Tsimikas JV, Tempany CM, et al. Diagnosis and staging of ovarian cancer: comparative values of Doppler and conventional US, CT, and MR imaging correlated with surgery and histopathologic analysisreport of the Radiology Diagnostic Oncology Group. Radiology 1999; 212(1):1927.

385

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 385414

Saline Infusion Sonohysterography


Ruth B. Goldstein,
&

MD

Saline infusion sonohysterography: clinical indications Unexplained uterine bleeding endometrium suboptimally imaged with transvaginal sonography Abnormal bleeding and abnormal endometrial appearance on transvaginal sonography (too thick, cystic, suspect mass) Transvaginal sonography and biopsy findings are inconsistent Patient has symptomatic fibroids and desires resection: hysteroscopic versus open resection

& & & & & & &

Suspected congenital anomaly (eg, septate uterus), synechiae, or unexplained infertility/habitual abortion Risks of saline infusion sonohysterography Scheduling the saline infusion sonohysterography Contraindications Technique Technical success/difficulties Benefits of saline infusion sonohysterography References

Saline infusion sonohysterography (SIS) is a diagnostic procedure that combines simultaneous vaginal sonography with infusion of sterile saline into the endometrial cavity. The infusion of saline serves as a contrast medium and distends the endometrial canal allowing exquisite display of the inner lining of endometrium during real-time imaging [Fig. 1]. This procedure has been known by many names including sonohysterography, hysterosonography, utlrasonohysterography, and SIS. The latter term, recommended by a consensus panel held recently on the topic of postmenopausal bleeding (PMB) [1,2], is used in this article. SIS was introduced 25 years ago. First described by Nannini and coworkers in 1981 [3], transabdominal sonographic imaging was used in conjunction with saline infusion to study infertile patients. The technique did not catch on. Several years later, however, vaginal probes were intro-

duced. This single addition to sonographic technology, which became widely available by 1990, led to a revolutionary change in gynecologic imaging and renewed interest in saline infusion. In 1993, Parsons and Lense [4] published encouraging results for a technique they termed sonohysterography as a means for further evaluating endometrial abnormalities detected on unenhanced sonograms. Shortly thereafter, Goldstein [5] became a public advocate for using this technique to triage women who had unexplained perimenopausal uterine bleeding to the best next step in the diagnostic evaluation (ie, directed versus nondirected biopsy). Since then, many reports of the benefits of SIS have touted the improved sensitivity for the detection of endometrial abnormalities compared with unenhanced vaginal sonography [612]. Today, SIS has evolved as a useful, safe, and minimally invasive examination for women who have abnormal uter-

Ultrasound Section, Department of Radiology, University of California San Francisco, 505 Parnassus Avenue, L-374, San Francisco, CA 94143-0628, USA E-mail address: ruth.goldstein@radiology.ucsf.edu
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.02.001

ultrasound.theclinics.com

386

Goldstein

Fig. 1. Normal saline infusion sonogram. Smooth endometrial cavity is distended by saline.

ine bleeding, infertility, and congenital uterine anomalies. The clinical indications (and contraindications) have been broadly endorsed by the American College of Obstetrics and Gynecology and the American Institute in Ultrasound in Medicine [2].

Unexplained uterine bleedingendometrium suboptimally imaged with transvaginal sonography


In women who have abnormal vaginal bleeding (after cervical or other etiologies are excluded), the uterus and endometrium are imaged with TVS. In premenopausal women, abnormal bleeding is often anovulatory or dysfunctional. A normal-appearing endometrium is reassuring but not sensitive for detection of small masses that may cause bleeding [Fig. 4], especially if the sonogram is performed in the secretory phase when the endometrium is thickest and most echogenic [Fig. 5]. SIS may help to detect or confirm small polyps [Fig. 6] or more accurately determine the location of leiomyomata in relation to the central uterine cavity (submucosal versus intramural). When the endometrium is incompletely imaged due to shadowing from fibroids [Fig. 7], adenomyosis, position of the uterus, or other reasons in premenopausal women, SIS can be helpful for more complete evaluation. In postmenopausal women who have abnormal bleeding, complete imaging of the endometrium is even more important because the risk of endometrial cancer is highest in this group. In women who have PMB, endometrial cancer is the paramount concern. Using TVS, a thin endometrium (bilayer endometrial thickness [ET] <5 mm) [Fig. 8] is associated with a high negative predictive value and a negligible risk of endometrial cancer (0.07%) [19], a risk so low that many believe that tissue sampling may be safely deferred [1,20,21]. Although small benign lesions may be missed using this 5-mm threshold in women who have PMB [10], more than a decade of published observations support the observation that cancers are extremely unlikely when the ET is less than 4 or 5 mm [22,23]. Thus, the author does not believe that SIS is necessary when the endometrial stripe is visualized com-

Saline infusion sonohysterography: clinical indications


The single most common clinical indication for SIS is abnormal vaginal bleeding, which is largely due to the fact that anatomic (as opposed to hormonal) causes of bleeding are often focal and more sensitively detected with SIS compared with unenhanced transvaginal sonography (TVS) [7,10,13]. Some believe that SIS should be performed in any woman who has abnormal bleeding; however, the author believes that patients would be well served if clinicians were to adopt a low threshold for performing this examination in women who have abnormal bleeding, especially for those in whom the endometrium is incompletely imaged or there is a suspected abnormality in the endometrium on the transvaginal sonogram [Fig. 2]. It has been convincingly shown that SIS detects focal endometrial masses more sensitively than TVS alone (over 95% versus 67%) [10,12,1416]. It is important to identify a focal lesion before biopsy because these lesions may be missed with nondirected tissue sampling (office endometrial biopsy or dilation and curettage [D & C]) [17,18]. In some cases, three-dimensional sonography may be helpful before and during saline infusion [Fig. 3], although the benefits over conventional two-dimensional SIS have not been proven. The following discussion focuses on five clinical situations in which the author has found SIS to be particularly useful.

Saline Infusion Sonohysterography

387

Fig. 2. Enlarged fibroid uterus obscures endometrium and prevents endometrial evaluation. Longitudinal (A) and coronal (B) images.

pletely, when it is homogenous, and when the bilayer thickness is less than 5 mm. It is critical, however, that these criteria be used clinically to defer endometrial biopsy (EMB) only when the endometrium can be visualized in its entirety (cervical canal to fundus) [Fig. 9]. If part of the endometrium is obscured for any reason (fibroids, refractive shadowing due to position of the uterus or poorly defined endometrial margins, adenomyosis, and so forth) [Fig. 10], then further evaluation with SIS or tissue sampling is recommended [1]. A small amount of endometrial fluid may be detected in the postmenopausal woman [see Fig. 8B] [24]. This finding is not necessarily abnormal and

does not increase suspicion for cancer of the endometrium or cervix. On the contrary, this fluid may be useful because it allows detailed display of the inner surface and thickness of the endometrium without saline infusion, and SIS may not be necessary.

Abnormal bleeding and abnormal endometrial appearance on transvaginal sonography (too thick, cystic, suspect mass)
In premenopausal patients, the thickness of the endometrium alone is usually not discriminating for the cause of the abnormal bleeding. The ET

388

Goldstein

Fig. 3. Three-dimensional sonography of endometrial polyps. (A) Unenhanced coronal transvaginal image of endometrium (arrowheads). Polyps are not evident. (B) Saline infused sonogram (two-dimensional). Calipers highlight an endometrial polyp. (C ) Planes for volume sampling for three-dimensional SIS. (D) Three-dimensional display of SIS shows polyps (arrows). (Courtesy of Dr. Beryl Benacerraf, MD, Boston, MA.)

may increase to 16 to 17 mm in the second half of the menstrual cycle in normal patients. When the endometrium is thickest and most echogenic, small masses can easily be missed [Fig. 5]. Compared with TVS, SIS is more sensitive for focal masses that may cause abnormal bleeding (polyps, fibroids, and hyperplasia). Because nonguided D & C or office endometrial sampling may miss these focal abnormalities [25,26], SIS should be considered despite a normal or heterogenous endometrial stripe [Fig. 11] [27]. In one study, 24% of endometrial polyps evident with SIS were not detected on the preliminary TVS [14]; however, detection of an obvious mass outlined by lucent proliferative endometrium before saline infusion can be considered reliable, and the patient should proceed directly to hysteroscopy [Fig. 12].

In postmenopausal women who have abnormal bleeding, an ET of 5 mm or greater should be interpreted as abnormal [Fig. 13A]. The risk of endometrial cancer is 100 times greater when the ET is greater than 5 mm than when the ET is 5 mm or less (7.3% versus 0.07%) [19], and some form of tissue sampling is advised [21]. Although endometrial cancer accounts for the minority (7%10%) of the bleeding after menopause, there is a reasonably high likelihood that some anatomic abnormality (other than atrophy) is present. These abnormalities include endometrial polyps, hyperplasia, and fibroids. In one prospective Swedish study in which 105 women who had PMB and an ET of 5 mm or greater underwent D & C and hysteroscopy, 84 (80%) had a pathologic diagnosis (19 endometrial cancer, 43 endometrial polyps, and 10 focal

Saline Infusion Sonohysterography

389

Fig. 3 (continued ).

390

Goldstein

Fig. 4. (A) Pre-SIS proliferative endometrium (calipers) appears normal. (B) During SIS, a small polyp (arrow) is detected.

hyperplasia) [26]. Not only did most of the women have real pathology but a focal lesion was also found in 98% of those who had pathology. Thus, gynecologists should proceed to some form of tissue sampling when the ET is greater than 5 mm (and the patient is bleeding). Whether this sample is obtained by way of simple, nondirected office EMB or hysteroscopic D & C (greater risk, cost, and so forth) is determined by whether the abnormality is focal or diffuse. As mentioned earlier, it is now

well established that focal endocavitary lesions may be missed by office aspiration biopsy and nondirected D & C [17,28,29]. In one study, nondirected D & C missed 58% of polyps, 50% of hyperplasia, 60% of complex atypical hyperplasias, and 11% (2/19) endometrial cancers [26]. TVS can show that a woman should undergo tissue sampling (ET >5 mm) but cannot always show whether the abnormality is diffuse or focal [Fig. 14]. This is the setting in which SIS can be very helpful [see

Saline Infusion Sonohysterography

391

Fig. 5. Small masses can be missed when the endometrium is in secretory phase. (A) Bilayer thickness of sagittal secretory endometrium (calipers) is 15 mm. Polyp is not detected. (B) Coronal secretory endometrium (arrowheads). Polyp is not visible. (C ) Repeat sonogram done during proliferative phase. A small mass (polyp) (arrow) is now evident.

392

Goldstein

Fig. 6. SIS can clarify suspicious findings on TVS. (A) Preliminary TVS shows suspicious heterogeneous but nonspecific endometrium (arrows). (B) SIS shows an endocavitary mass (polyp or pedunculated fibroid) delineated by calipers.

Fig. 13B] and has been shown to correlate well with results obtained with hysteroscopic resesections [8]. If SIS shows the lesion to be diffuse, then nondirected biopsy may be appropriate and hysteroscopy may not be necessary. Otherwise, hysteroscopic biopsy or D & C is recommended. There is one potential relative contraindication to SIS after detection of a thickened endometrium. When endometrial cancer is strongly suspected based on the sonographic findings on the TVS [Figs. 15 and 16], the author defers SIS due to the theoretic risk of spreading cancerous cells into the peritoneum through the tubes during saline

instillation. Alcazar and colleagues [30] found malignant cells in the peritoneal washings of one of the 14 patients (7%) with stage I adenocarcinoma of the endometrium who had undergone SIS prior to hysterectomy.

Transvaginal sonography and biopsy findings are inconsistent


In some cases, the ET appears thick on the sonogram yet tissue sampling returns a pathologic diagnosis of atrophy. There are only a few reasons for this discordance. The first is a case of cystic atrophy

Saline Infusion Sonohysterography

393

Fig. 7. Endometrial canal not well imaged due to illdefined margins and position of the uterus. Endometrium (calipers) appears heterogeneous but difficult to evaluate.

of these masses is often desired. If possible, hysteroscopic resection of fibroids is the method of choice so that a laparotomy or hysterectomy can be avoided. Generally accepted criteria for successful hysteroscopic resection include (1) mass greater or equal to 50% intracavitary, (2) size less than 5 cm, and (3) outer myometrial mantle 5 cm or greater. These criteria are well evaluated with SIS. The intracavitary location is more sensitively determined with SIS than with TVS [Fig. 20] [31]. In some cases, the vaginal sonogram during saline infusion is limited by the large fibroids. In these cases, the author has found that the anatomy can be beautifully displayed using transabdominal sonography by merely changing the mode of scanning from transvaginal to transabdominal during the installation of saline transvaginally [Fig. 21]. This technique allows a large field of view with the benefit of saline contrast enhancement of the cavity. When the patient has known or suspected endometriosis, an echogenic mass in the myometrium adjacent to the endometrium may fill during saline infusion, revealing itself to be an endometriumlined cavity [Fig. 22]. These masses are thought to represent a manifestation of adenomyosis (adenomyomatic cavity) [32].

misinterpreted as thick cystic endometrium on the TVS. Although endometrial atrophy (hormonally inactive endometrium) generally appears thin, occasionally cystic atrophy (cystic degeneration) of the endometrium [Fig. 17] can produce a multicystic appearance that can be confused for cystic hyperplasia. In cystic atrophy, the intervening visible endometrial tissue is scant. Similarly, if the patient is taking a drug like tamoxifen, she may develop subendometrial cysts, thought to be related to endometriosis and benign in clinical importance. SIS can be helpful in determining whether the cysts are subendometrial or intraendometrial. If the endometrial tissue is thick but contains cysts, then a diffuse of focal endometrial abnormality should strongly be suspected [Fig. 18]. Endometrial polyps often contain small cysts. Some can get quite large. If the EMB sampled the adjacent thin atrophic endometrium, then the findings of the pathologist and sonologist will be discordant [Fig. 19]. In these cases, SIS or hysteroscopy should be performed.

Suspected congenital anomaly (eg, septate uterus), synechiae, or unexplained infertility/ habitual abortion
Anatomic abnormalities including Mullerian anomalies, submucosal fibroids, and uterine synechiae are thought to account for 15% to 55% of recurrent pregnancy loss [3335]. Using SIS for endometrical cavity imaging, investigators found intrauterine abnormalities in 17 of 34 women evaluated for recurrent pregnancy losses. Although contrast hysterosalpingography has traditionally been used to evaluate for these anatomic abnormalities in infertile patients, SIS has been shown to perform better than hysterosalpingography [36]. SIS can be particularly helpful for identifying synechiae and distinguishing the septate uterus from the bicornuate uterus [34,37]. As a screening tool, SIS is twice as accurate as hysterosalpingography and TVS [36].

Risks of saline infusion sonohysterography


The risks (and benefits) of SIS should be discussed with the patient before the procedure. Risks are small but include discomfort from the speculum and balloon inflation, bleeding (usually trivial), uterine perforation (almost unheard of), and infection. Mild discomfort is common, but severe pain is rare. Infection (endometritis) is a serious complication but the risk is low, estimated to be less than 1% [12,17,38]. Risk of infection is probably

Patient has symptomatic fibroids and desires resection: hysteroscopic versus open resection
Fibroids may be associated with excessive, often painful vaginal bleeding and infertility. Resection

394

Goldstein

Fig. 8. Normal postmenopausal endometrium. (A) Endometrium is thin, measuring less than 5 mm in bilayer thickness on a sagittal transvaginal image (arrowheads). (B) Occasionally, a small amount of fluid outlines the atrophic endometrium (arrow).

increased if there is a prior history of pelvic inflammatory disease (PID), if the fallopian tubes are dilated, or if there is current (indolent) PID. Routine prophylaxis with antibiotics is not recommended but should be considered in patients at increased risk and in those who take dental prophylaxis [2]. An unenhanced vaginal sonogram before saline infusion in each patient should include gentle probing of the adnexa to test for unusual tenderness suggestive of indolent PID. If suspected, then the SIS should be postponed until after treatment has been administered. Pre- or post-

procedure treatment with ibuprofen or acetaminophen is optional.

Scheduling the saline infusion sonohysterography


The patient should be scheduled when there is the best chance of making the most specific/sensitive diagnosis. Ideally, the procedure should be performed when the endometrium is thinnest, most homogeneous, and if the patient is premenopausal, lucent.

Saline Infusion Sonohysterography

395

Fig. 9. Endometrium (small arrows) should be visualized in its entirety down to the cervix (white arrows). A small nabothian cyst (white arrowhead) confirms the cervical area is on this image.

In premenopausal women, the best time to perform SIS is in the proliferative phase (first half) of the menstrual cycle. In the second half of the cycle (secretory phase), the endometrium is thicker, wavier, and more difficult to interpret and SIS is not recommended [39] (there is also a greater chance that the patient could be pregnant). If possible, it is optimal to wait until the bleeding has stopped because blood clots can be confused for focal lesions. In postmenopausal women, timing is dependent on whether the patient is taking cyclic hormone replacement therapy (HRT). If she is taking HRT, then the SIS should be scheduled approximately 6 days after the last progestin pill when the endometrium is thinnest. Postmenopausal patients who are not taking HRT can be scheduled at any time. At the authors institution, the pre- or perimenopausal patient is instructed the to call to schedule the SIS on the first day of her menstrual cycle to ensure correct scheduling. The SIS should be scheduled for day 6 to 10 of the cycle, preferably after cessation of bleeding and before ovulation. Proper scheduling of the patient is important to avoid potential misinterpretations and lost work (if the patient has to be rescheduled). In the authors radiology suite, the procedure is scheduled for 30 to 45 minutes to allow for the preliminary vaginal sonogram, restroom stops, and consent. The procedure (saline infusion while imaging), how-

ever, is usually completed in approximately 10 minutes or less.

Contraindications
Strict contraindications to SIS include pregnancy and current pelvic infection. A pregnancy test should be performed if pregnancy status is uncertain. If the patient has had prior PID or takes dental prophylaxis, then premedication with antibiotics should be considered, but there is little data on the efficacy of this practice. Relative contraindications include current vaginal bleeding, but this may be unavoidable. Patients should be asked about latex allergy before the procedure and nonlatex condoms and gloves should be used in case of allergy.

Technique
Several available catheters can be used for SIS, including Meditech 5 or 7 French catheters (Medical Device Technologies, Gainesville, Florida), Ackrad Hysterocath 5 or 7 F catheters (Ackrad, Laboratories, Cranford, New Jersey), the Goldstein catheter (Cook, Spencer, Indiana), and the Soules 5.3 F insemination catheter (Cook). Although straight catheters and balloon-tipped catheters are reported to be equally efficacious [40], the author prefers using a balloon-tipped catheter because it prevents

396

Goldstein

Fig. 10. Endometrium is poorly seen on this sagittal image. Imagers should resist the temptation to interpret it as normal despite the calipers.

leakage of infused saline during infusion or dislodging of the catheter during speculum removal and vaginal probe insertion. The catheter should be flushed (and locked) before the procedure to prevent introduction of air during the procedure. The balloon should also be inflated/tested before the procedure to make sure it is functional and to allow easier distension during the procedure. SIS trays are commercially available [Fig. 23] (Medical Device Technologies, Gainesville, Florida), but you can easily make your own. The equipment includes an SIS catheter, a sterile saline and bowl, provodine/iodine solution, sterile gauze/pledgets/sponges, sterile swabs, a sterile speculum, a light source, a condom for the vaginal probe (preferably sterile) and sterile jelly, two syringes (one for saline infusion, the other for balloon inflation), and a tenaculum (optional). The author almost never finds it necessary to use a tenaculum during this procedure, although the speculum may have to be adjusted a bit to cannulate the cervix in some patients. The author finds that a ring forceps is handy, with pledgets for cleansing or absorbing excess betadine that may pool in the vagina and occasionally for better control of the catheter tip during insertion into the cervix. The following order of events works best: 1. Preliminary TVS is performed to assess the appearance of the endometrium, myometrium, and adnexae and to assess for adnexal tenderness.

2. Risks and benefits of procedure are explained and consent is obtained for SIS. 3. Patient is sent to restroom to empty bladder (again). 4. It is best to perform this procedure in stirrups. The perineum is cleaned with a provodine/ iodine solution and the speculum is inserted. Illuminated disposable speculums are very helpful. 5. The cervix is swabbed with the provodine/ iodine solution, the catheter is placed in the cervix, and the balloon is gently inflated with saline until a purchase is achieved and gentle pulling under direct vision does not displace the catheter. The author prefers placement of the balloon within the cervix (to allow more complete visualization of the endometrial cavity as infusion takes place), but this is potentially slightly more uncomfortable. 6. The speculum is removed, and the vaginal probe, covered by a condom, is reinserted. 7. Gentle infusion of sterile saline is completed during real-time sonography. 8. Coronal and sagittal imaging is done of the entire uterine canal. 9. The balloon is deflated and the catheter is gently removed to just above the endocervical os, removing it slowly with continued gentle saline infusion and continuous real-time imaging to allow imaging of the endocervix and potential polyps. 10. The catheter and the vaginal probe are removed.

Technical success/difficulties
SIS is successful in 90% to 95% cases [1012,41], perhaps slightly lower in postmenopausal women (86.5%) compared with premenopausal women (95%) [41]. A number of potential technical difficulties can occur, however, mainly during introduction of the catheter or installation. Cervical stenosis (in pre- or postmenopausal women) can result from prior instrumentation or delivery and prevent cannulation of the cervix. When difficulty is encountered, the author has found it helpful to change the position of the speculum to reorient the cervix to allow passage. Scarring of the endocervix is irregular, and often, merely gently repositioning the uterus with a change in the position/ orientation of the speculum allows successful passage of the catheter. Gently probing with a sound or sterile swab may also help. If air is inadvertently introduced into the cavity, then the examiner must scan around it or wait for it to dissipate to complete the examination. If the balloon is inflated above the internal cervical os, then the balloon may ob-

Saline Infusion Sonohysterography

397

Fig. 11. Heterogeneous endometrium (arrowheads) on sagittal (A) and coronal (B) images in a premenopausal woman who had abnormal bleeding. (C ) SIS with color Doppler imaging shows multiple polyps (nonvascular) (small arrows) that were resected hysteroscopically (benign).

scure the lower endometrium and must be moved or deflated to see the lower uterine segment and cervical canal.

Benefits of saline infusion sonohysterography


The major added value of SIS over unenhanced vaginal sonography is largely due to the improved

imaging of the inner surface of both sides of the endometrium after saline distends the uterine cavity. One recent meta-analysis suggested 95% sensitivity and 88% specificity for SIS [41]. Focal and diffuse abnormalities can be distinguished, and in most cases, endometrial polyps can be distinguished from submucosal fibroids based on the imaging characteristics. Color and duplex Doppler imaging should be used to improve specificity [42].

398

Goldstein

Fig. 11 (continued ).

Saline Infusion Sonohysterography

399

Fig. 12. Proliferative endometrium reveals endometrial polyps without SIS. (A, B) Proliferative endometrium shows two small echogenic masses (arrows). A small feeding vessel (small arrow) to one of the polyps is seen using power Doppler (A). (C ) Unenhanced sonogram shows at least two small polyps (small arrows) in the proliferative phase of the menstrual cycle outlined by a tiny amount of endometrial fluid (arrow). SIS is not necessary.

400

Goldstein

Fig. 12 (continued ).

Saline Infusion Sonohysterography

401

Fig. 13. Postmenopausal bleeding and endometrial thickening. (A) Preliminary TVS shows 10-mm endometrium (calipers) (is the abnormality diffuse or focal?). (B) SIS shows focal mass (arrow). Hysteroscopy is recommended.

402

Goldstein

Fig. 14. Postmenopausal bleeding and heterogeneous endometrium. ( A) Preliminary TVS shows that the ET is 8 mm, but it is unclear whether there is a focal or diffuse abnormality. (B) SIS shows a focal abnormality (long arrow) best removed hysteroscopically. SIS catheter balloon is shown in lower uterine segment (short arrow).

Saline Infusion Sonohysterography

403

Fig. 15. Elderly woman who had postmenopausal bleeding. Highly suspicious sonogram. SIS is not necessary or recommended. (A) Large mass (arrows) in endometrial cavity seen on transvaginal image (coronal view). Calipers mark the edges of the uterus. (B) Sagittal image (transvaginal) shows trace fluid (short arrow) outlining large mass (m). Arrowheads indicate the massmyometrial interface. (C ) Coronal T2-weighted MRI shows large mass (long arrow) outlined by fluid (short arrow) on T2-weighted image. (D) Sagittal T2-weighted MRI shows large endometrial mass (long arrow) with possible myometrial invasion (arrowhead). (MRI was ordered because hysteroscopy was technically difficult and sampling was unsuccessful.)

404

Goldstein

Fig. 15 (continued ).

Saline Infusion Sonohysterography

405

Fig. 16. Single sagittal image from TVS in an elderly woman who had postmenopausal bleeding. Endometrium is thickened (20 mm) (arrowheads), heterogeneous, and irregularly marginated. Endometrial cancer suspected. SIS not advised.

Fig. 17. Cystic atrophy. Sagittal image. Calipers indicate a very thin (<2 mm) bilayer ET. Small cysts are detected in the lower uterine segment (arrows).

406

Goldstein

Fig. 18. Cystic endometrium (calipers). Differential diagnosis is hyperplasia, polyp, or cancer. The intervening endometrium is too thick to consider cystic atrophy.

Saline Infusion Sonohysterography

407

Fig. 19. Endometrial biopsy missed a polyp (this figure represents three patients). ( A) Patient sent for sonogram because she continued to bleed despite endometrial biopsy showing atrophy. TVS shows an endometrial polyp (arrow). The surrounding endometrium is thin (short arrow) and compatible with atrophy. (B) Color Doppler image of same patient as in (A) shows vascular stalk in polyp (arrow). (C ) SIS shows a large endometrial polyp (p). Calipers outline the single-layer endometrium. (D) Coronal image shows cystic endometrium (arrowheads). (E ) Sagittal image during SIS of same patient as in (D) shows that cysts are within a large polyp (p). SIS catheter balloon is indicated by a small arrow.

408

Goldstein

Fig. 19 (continued ).

Saline Infusion Sonohysterography

409

Fig. 19 (continued ).

410

Goldstein

Fig. 20. Hysteroscopically resectable fibroid. (A) Coronal transvaginal image shows a probable shadowing mass centrally in the uterus (arrows). (B) SIS shows the 4-cm fibroid is submucosal and probably resectable through the hysteroscope. The SIS catheter balloon (b) is in the cervix.

Saline Infusion Sonohysterography

411

Fig. 21. Occasionally, transabdominal imaging during SIS can be of use to clarify findings. (A) Sagittal transvaginal imaging is limited during SIS. A small amount of endometrial saline is seen (arrow), but the relationship of the large fibroid (f) to the canal cannot be delineated. (B) Transabdominal image shows the large fibroid (f) is intramural and not protruding into the endometrial cavity (arrow). The inflated SIS balloon is seen in the cervix (small arrow).

412

Goldstein

Fig. 22. Adenomyotic cavities may fill with saline during SIS. (A) Preliminary sagittal image from vaginal sonogram shows myometrial masses (calipers) thought to be fibroids. Endometrium is seen inferior to mass (arrows). (B) During SIS, these cavities fill with saline (arrows), and multiple polyps (arrowheads) become evident.

Fig. 23. SIS kit including drape disposable speculum (D), sound (s) (arrowhead ), SIS catheter (arrow), basin (B), and cleansing sponges.

Saline Infusion Sonohysterography

413

References
[1] Goldstein RB, Bree RL, Benson CB, et al. Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in Ultrasound Sponsored Consensus Conference statement. J Ultrasound Med 2001;20:102536. [2] AIUM standard for the performance of saline infusion sonohysterography. J Ultrasound Med 2003;22:1216. [3] Nannini R, Chelo E, Branconi F, et al. Dynamic echohysteroscopy: a new diagnostic technique in the study of female infertility. Acta Eur Fertil 1981;12:16571. [4] Parsons AK, Lense JJ. Sonohysterography for endometrial abnormalities: preliminary results. J Clin Ultrasound 1993;21:8795. [5] Goldstein SR. Use of ultrasonohysterography for triage of perimenopausal patients with unexplained uterine bleeding. Am J Obstet Gynecol 1994;170:56570. [6] Mihm LM, Quick VA, Brumfield JA, et al. The accuracy of endometrial biopsy and saline sonohysterography in the determination of the cause of abnormal uterine bleeding. Am J Obstet Gynecol 2002;186:85860. [7] Williams CD, Marshburn PB. A prospective study of transvaginal hydrosonography in the evaluation of abnormal uterine bleeding. Am J Obstet Gynecol 1998;179:2928. [8] OConnell LP, Fries MH, Zeringue E, et al. Triage of abnormal postmenopausal bleeding: a comparison of endometrial biopsy and transvaginal sonohysterography versus fractional curettage with hysteroscopy. Am J Obstet Gynecol 1998; 178:95661. [9] Widrich T, Bradley LD, Mitchinson AR, et al. Comparison of saline infusion sonography with office hysteroscopy for the evaluation of the endometrium. Am J Obstet Gynecol 1996;174: 132734. [10] Bree RL, Bowerman RA, Bohm-Velez M, et al. US evaluation of the uterus in patients with postmenopausal bleeding: a positive effect on diagnostic decision making. Radiology 2000;216: 2604. [11] Lev-Toaff AS, Toaff ME, Liu JB, et al. Value of sonohysterography in the diagnosis and management of abnormal uterine bleeding. Radiology 1996;201:17984. [12] Bonnamy L, Marret H, Perrotin F, et al. Sonohysterography: a prospective survey of results and complications in 81 patients. Eur J Obstet Gynecol Reprod Biol 2002;102:427. [13] Timmerman D, Deprest J, Bourne T, et al. A randomized trial on the use of ultrasonography or office hysteroscopy for endometrial assessment in postmenopausal patients with breast cancer who were treated with tamoxifen. Am J Obstet Gynecol 1998;179:6270. [14] Laifer-Narin S, Ragavendra N, Parmenter EK, et al. False-normal appearance of the endome[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

trium on conventional transvaginal sonography: comparison with saline hysterosonography. AJR Am J Roentgenol 2002;178:12933. Nanda S, Chadha N, Sen J, et al. Transvaginal sonography and saline infusion sonohysterography in the evaluation of abnormal uterine bleeding. Aust N Z J Obstet Gynaecol 2002;42:5304. Kazandi M, Aksehirli S, Cirpan T, et al. Transvaginal sonography combined with saline contrast sonohysterography to evaluate the uterine cavity in patients with abnormal uterine bleeding and postmenopausal endometrium more than 5 mm. Eur J Gynaecol Oncol 2003;24:18590. Dubinsky TJ, Parvey HR, Gormaz G, et al. Transvaginal hysterosonography: comparison with biopsy in the evaluation of postmenopausal bleeding. J Ultrasound Med 1995;14:88793. Van den Bosch T, Vandendael A, Van Schoubroeck D, et al. Combining vaginal ultrasonography and office endometrial sampling in the diagnosis of endometrial disease in postmenopausal women. Obstet Gynecol 1995;85:34952. Smith-Bindman R, Weiss E, Feldstein V. How thick is too thick? When endometrial thickness should prompt biopsy in postmenopausal women without vaginal bleeding. Ultrasound Obstet Gynecol 2004;24:55865. Smith-Bindman R, Kerlikowske K, Feldstein VA, et al. Endovaginal ultrasound to exclude endometrial cancer and other endometrial abnormalities. JAMA 1998;280:15107. Gull B, Carlsson S, Karlsson B, et al. Transvaginal ultrasonography of the endometrium in women with postmenopausal bleeding: is it always necessary to perform an endometrial biopsy? Am J Obstet Gynecol 2000;182:50915. Karlsson B, Granberg S, Wikland M, et al. Transvaginal ultrasonography of the endometrium in women with postmenopausal bleedinga Nordic multicenter study. Am J Obstet Gynecol 1995;172:148894. Ferrazzi E, Torri V, Trio D, et al. Sonographic endometrial thickness: a useful test to predict atrophy in patients with postmenopausal bleeding. An Italian multicenter study. Ultrasound Obstet Gynecol 1996;7:31521. Goldstein SR. Postmenopausal endometrial fluid collections revisited: look at the doughnut rather than the hole. Obstet Gynecol 1994;83:73840. Word B, Gravlee LC, Wideman GL. The fallacy of simple uterine curettage. Obstet Gynecol 1958; 12:6428. Epstein E, Ramirez A, Skoog L, et al. Dilatation and curettage fails to detect most focal lesions in the uterine cavity in women with postmenopausal bleeding. Acta Obstet Gynecol Scand 2001;80:11316. Kamel HS, Darwish AM, Mohamed SA. Comparison of transvaginal ultrasonography and vaginal sonohysterography in the detection of endometrial polyps. Acta Obstet Gynecol Scand 2000; 79:604.

414

Goldstein

[28] Guido RS, Kanbour-Shakir A, Rulin MC, et al. Pipelle endometrial sampling. Sensitivity in the detection of endometrial cancer. J Reprod Med 1995;40:5535. [29] Goldchmit R, Katz Z, Blickstein I, et al. The accuracy of endometrial Pipelle sampling with and without sonographic measurement of endometrial thickness. Obstet Gynecol 1993;82: 72730. [30] Alcazar JL, Errasti T, Zornoza A. Saline infusion sonohysterography in endometrial cancer: assessment of malignant cells dissemination risk. Acta Obstet Gynecol Scand 2000;79:3212. [31] Leone FP, Lanzani C, Ferrazzi E. Use of strict sonohysterographic methods for preoperative assessment of submucous myomas. Fertil Steril 2003;79:9981002. [32] Reeves MF, Goldstein RB. Communication of adenomyosis with the endometrial cavity: visualization with saline-infusion sonohysterography. Ultrasound Obstet Gynecol 2006 [in press]. [33] March CM, Israel R. Hysteroscopic management of recurrent abortion caused by septate uterus. Am J Obstet Gynecol 1987;156:83442. [34] Buttram Jr VC, Gibbons WE. Mullerian anomalies: a proposed classification (an analysis of 144 cases). Fertil Steril 1979;32:406. [35] Keltz MD, Olive DL, Kim AH, et al. Sonohysterography for screening in recurrent pregnancy loss. Fertil Steril 1997;67:6704. [36] Soares SR, Barbosa dos Reis MM, Camargos AF.

[37]

[38]

[39]

[40]

[41]

[42]

Diagnostic accuracy of sonohysterography, transvaginal sonography, and hysterosalpingography in patients with uterine cavity diseases. Fertil Steril 2000;73:40611. Pellerito JS, McCarthy SM, Doyle MB, et al. Diagnosis of uterine anomalies: relative accuracy of MR imaging, endovaginal sonography, and hysterosalpingography. Radiology 1992;183: 795800. Farquhar C, Ekeroma A, Furness S, et al. A systematic review of transvaginal ultrasonography, sonohysterography and hysteroscopy for the investigation of abnormal uterine bleeding in premenopausal women. Acta Obstet Gynecol Scand 2003;82:493504. Wolman I, Groutz A, Gordon D, et al. Timing of sonohysterography in menstruating women. Gynecol Obstet Invest 1999;48:2548. Dessole S, Farina M, Capobianco G, et al. Determining the best catheter for sonohysterography. Fertil Steril 2001;76:6059. de Kroon CD, de Bock GH, Dieben SW, et al. Saline contrast hysterosonography in abnormal uterine bleeding: a systematic review and metaanalysis. Br J Obstet Gynaecol 2003;110:93847. Timmerman D, Verguts J, Konstantinovic ML, et al. The pedicle artery sign based on sonography with color Doppler imaging can replace second-stage tests in women with abnormal vaginal bleeding. Ultrasound Obstet Gynecol 2003;22:16671.

415

ULTRASOUND CLINICS
Ultrasound Clin 1 (2006) 415424

Abnormal Uterine Bleeding: The Role of Ultrasound


Steven R. Goldstein,
& & &

MD

Evolution of endometrial assessment Transvaginal ultrasound Other considerations for sonohysterography Timing of the procedure Difficulty threading the catheter Anesthesia/analgesia Risk of infection

& & & &

Concern about spreading adenocarcinoma into the peritoneal cavity Inadequate distension of the cavity Summary References

Abnormal uterine bleeding accounts for up to 20% of gynecologic visits [1]. Any pregnancy event must first be excluded, and the use of inexpensive, rapid, monoclonal antibody urine human chorionic gonadotropin tests, readily available over-the-counter, makes this a relatively simple maneuver. When a pregnancy event has been excluded, the most likely cause of bleeding is dysfunctional anovulatory bleedingwhat patients are often told is a hormone imbalance. As women get older, however, organic pathology such as polyp, submucous myomas, hyperplasias, and even frank carcinoma become more likely. According to the SEER database [2], the incidence of endometrial carcinoma in women aged 30 to 34 years is 2.3/100,000, increases to 6.1/100,000 between ages 35 and 40 years, and rises dramatically to 36.2/100,000 in women aged 40 to 49 years. In postmenopausal women on no hormone replacement therapy, any bleeding is considered cancer until proven otherwise, although the incidence of malignancy in such patients ranges from 2% to 10% depending on risk factors [3]. The role of the clinician in a patient who presents with bleeding is twofold: first, to exclude endometrial carcinoma in women older than 40 years [4],

and second, to identify the source of bleeding so it can be stopped or managed. Most patients who have abnormal bleeding will have dysfunctional uterine bleeding in association with episodes of anovulation (premenopausal) or endometrial atrophy (postmenopausal) that can best be managed hormonally or expectantly with reassurance. The main goal is to distinguish such patients from those who have organic pathologic conditions in a safe, painless, convenient manner.

Evolution of endometrial assessment


Initially, curettage was the gold standard. First described in 1843 [5], its performance in the hospital became the most common operation performed on women in the world. As early as the 1950s, a review of 6907 curettage procedures [6] found the technique missed endometrial lesions in 10% of cases. Of these, 80% were polyps. In the 1970s, vacuum-suction curettage devices allowed sampling without anesthesia in an office setting. The most popular was the Vabra aspirator (Berkeley Medevices, Berkeley, California). This device was found to be 86% accurate in diagnosing cancer [7]. Subsequently, less expensive, smaller,

New York University School of Medicine, 530 First Avenue, Suite 10N, New York, NY 10016, USA E-mail address: steven.goldstein@med.nyu.edu
1556-858X/06/$ see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.cult.2006.01.007

ultrasound.theclinics.com

416

Goldstein

less painful plastic catheters with their own internal pistons to generate suction became popular. One of these, the Pipelle device (Unimar, Wilton, Connecticut), was found to have similar efficacy but better patient acceptance compared with the Vabra aspirator [8]. Rodriguez and colleagues [9] did a pathologic study of 25 hysterectomy specimens. The percentage of endometrial surface sampled by the Pipelle device was 4% versus 41% for the Vabra aspirator. In one widely publicized study [10], the Pipelle had a 97.5% sensitivity to detect endometrial cancer in 40 patients undergoing hysterectomy. The shortcoming of that study was that the diagnosis of malignancy was known before the performance of the specimen collection. In another important study, Guido and colleagues [11] also studied the Pipelle biopsy in patients who had known carcinoma undergoing hysterectomy. Among 65 patients, a Pipelle biopsy provided tissue adequate for analysis in 63 (97%), but malignancy was detected in only 54 patients (83%). Of the 11 with false-negative results, 5 (8%) had disease confined to endometrial polyps and 3 (5%) had tumor localized to less than 5% of the surface area of the cavity. The surface area of the endometrial involvement in that study was 5% or less of the cavity in 3 of 65 (5%); 5% to 25% of the cavity in 12 of 65 (18%), of which the Pipelle missed four cases; 26% to 50% of the cavity in 20 of 65 (31%), of which the Pipelle missed four; and greater than 50% of the cavity in 30 of 65 patients (46%), of which the Pipelle missed none. These results provide great insight about the way

endometrial carcinoma can be distributed over the endometrial surface or confined to a polyp. Because tumors localized in a polyp or a small area of endometrium may go undetected, the investigators in that study concluded that the Pipelle is excellent for detecting global processes in the endometrium. From these data, it seems that undirected sampling, whether through curettage or various types of suction aspiration, is often fraught with error, especially in cases in which the abnormality is not global but focal (polyps, focal hyperplasia, or carcinoma involving small areas of the uterine cavity).

Transvaginal ultrasound
Introduced in the mid-1980s, the vaginal probe uses higher frequency transducers in close proximity to the structure being studied. It yields a degree of image magnification that has been dubbed sonomicroscopy [12]. In the early 1990s, it was used in women who had postmenopausal bleeding to see if it could predict which patients lacked significant tissue and could avoid dilation and curettage or endometrial biopsy and its discomfort, expense, and risk [13,14]. Consistently, the finding of a thin, distinct endometrial echo 4 to 5 mm or less has been shown to effectively exclude significant tissue in women who have bleeding. It is unfortunate that the corollary is not nearly as helpful. The positive predictive value of an endometrial echo greater than 5 mm is not so useful, although in the authors experience, many clinicians have inappropriately used a thick echo on ultrasound as an indicator of pathology. Such inappropriate

Fig. 1. Long-axis transvaginal ultrasound image of a postmenopausal patient who had a history of uterine bleeding. A thin, linear, distinct endometrial echo here has a negative predictive value of 99%. Notice that it is clearly seen taking off from the endocervical canal.

Abnormal Uterine Bleeding

417

Fig. 2. Long-axis view of a patient who underwent previous uterine surgery in whom a meaningful distinct endometrial echo is technically inadequate. The clinician should never be afraid to state endometrial echo not well visualized.

application of transvaginal ultrasound is especially worrisome in patients who have no bleeding and in whom the finding is incidental. Endometrial thickness should be measured on a sagittal (long-axis) image of the uterus, and the measurement should be performed on the thickest portion of the endometrium, excluding the hypoechoic inner myometrium. It is a double-thickness measurement from basalis to basalis [15]. If fluid is present, then it is usually associated with cervical stenosis and atrophy [16]. The layers are measured separately and should be symmetric. It should be remembered that the endometrial cavity is a three-dimensional structure, and attempts

must be made to image the entire cavity. Recognizing the potentially pivotal role of transvaginal ultrasound in diagnostic evaluation, a statement should be included in the report regarding the technical adequacy of the scan. A well-defined endometrial echo should be seen taking off from the endocervical canal [Fig. 1]. It should be distinct. Often, fibroids, previous surgery, marked obesity, or an axial uterus may make visualization suboptimal. If so, it is acceptable and appropriate to conclude endometrial echo not well visualized [Fig. 2]. In these cases, ultrasound cannot be relied on to exclude disease. The next step for such patients who have bleeding should be hysteroscopy or saline infusion sonohysterography depending on the skill set and preference of the physician and patient. Although the use of fluid enhancement was described with abdominal ultrasound for uterine and tubal observations [17], it never gained widespread use. The introduction of the vaginal probe changed that practice considerably [18,19]. The use of fluid instillation into the uterus coupled with such highresolution transvaginal probes allows tremendous diagnostic enhancement with an inexpensive, simple, well-tolerated office procedure (see the article by R.B. Goldstein elsewhere in this issue). In a prospective pilot study, saline infusion sonohysterography was performed in 21 women who had abnormal perimenopausal uterine bleeding [20]. Of the 21 patients, 8 had obvious polypoid lesions [Fig. 3] and were triaged for operative hysteroscopic removal. The pathology report confirmed benign polyps in all 8 patients. Three patients had submucous myomas. Two had wireloop resectoscopic excision [Fig. 4]. The third, who

Fig. 3. Saline infusion sonohysterogram of a patient who had abnormal uterine bleeding. In this coronal view of the uterus, a polyp measuring 8.2 12.3 mm (calipers) is seen emanating from the posterior wall.

418

Goldstein

Fig. 4. (A) Long-axis view of a patient who had abnormal uterine bleeding. There is a central thickened uterine echo that is also heterogeneous. (B) Saline infusion sonohysterogram reveals an endoluminal mass (15.6 18.7 mm) outlined by calipers that represents an intraluminal myoma. Note the acoustic shadowing emanating from the myoma. Furthermore, note that the endometrial cavity itself is lined with thin endometrium compatible with the early proliferative phase in this perimenopausal patient. Finally, the distance from the back of the myoma to serosa is 12.5 mm (calipers).

had a submucous myoma that extended to the serosal edge of the uterus, received expectant management. Nine patients had no obvious anatomic lesion, and the endometrial thickness of the anterior or posterior wall was a maximum of 3.2 mm. The studies were purposely performed on days 4 to 6 of the bleeding cycle when early proliferative change would be expected if no anatomic abnormality existed. Biopsy in all 9 of these patients revealed early proliferative endometrium. Thus, these patients had dysfunctional (ie, anovulatory) uterine bleeding and were successfully treated with

progestational agents. One patient had an endometrial thickness along the anterior wall of 7.6 mm, although the posterior wall was thin (2.3 mm). Curettage with hysteroscopy revealed simple hyperplasia without atypia; this patient was also treated with progestational agents. Thus, it was concluded that endometrial fluid instillation (sonohysterogram) to enhance vaginal ultrasonography in perimenopausal women can reliably distinguish between patients who have minimal tissue (3 mm or less single-layer measurements) whose bleeding is anovulatory and best treated hormonally

Abnormal Uterine Bleeding

419

from patients who have significant tissue (3 mm or more single-layer thickness) in need of formal curettage and hysteroscopy. Furthermore, polyps can be distinguished from submucous myomas. This distinction allows appropriate triage for operative hysteroscopy in terms of skill required and length of time and equipment needed. Furthermore, this procedure eliminates the need for diagnostic hysteroscopy in patients whose bleeding is dysfunctional. As determined in this pilot study, the addition of saline infusion sonohysterography can reliably distinguish perimenopausal patients who have dysfunctional abnormal bleeding (no anatomic abnormality) from those who have globally thickened endometria or focal abnormalities. A clinical algorithm was proposed and studied in a large prospective trial of perimenopausal women who had abnormal bleeding using unenhanced transvaginal ultrasonography followed by saline infusion sonohysterography for selected patients and then no endometrial sampling, undirected endometrial sampling, or visually directed endometrial sampling depending on whether the ultrasonographically based triage revealed no anatomic abnormality, globally thickened endometrium, or focal abnormalities, respectively [Fig. 5] [21]. In that study, 280 patients (65%) displayed a thin, distinct, symmetric endometrial echo of 5 mm or less on days 4 to 6, and dysfunctional uterine bleeding was diagnosed. One hundred fifty-three (35%) had saline infusion sonohysterography. Of

these procedures, 44 (29%) were performed because of the inability to adequately characterize and measure the endometrium [Fig. 6] and 109 (71%) were done for an endometrial measurement of 5 mm or greater. Sixty-one of those patients then had anterior and posterior endometrial thickness that was symmetric and less than 3 mm, compatible with dysfunctional uterine bleeding. Fifty-eight patients (13%) had focal polypoid masses [Fig. 7] that were removed hysteroscopically and confirmed pathologically. Twenty-two patients (5%) had submucous myomas, although 148 (34%) had clinical and ultrasonographic evidence of fibroids. Ten patients had symmetric single-layer measurements of endometrium at saline infusion sonohysterography greater than 3 mm (range, 39 mm). Of these, histologic type was proliferative endometrium in 5 and hyperplastic endometrium in 5. Saline infusion sonohysterography was technically inadequate in 2 patients who then underwent hysteroscopy with curettage. Undirected office biopsy alone without imaging potentially would have missed the diagnosis of focal lesions such as polyps, submucous myomas, and focal hyperplasia in up to 80 patients (18%). Based on these results, it seems apparent that any blind endometrial sampling should be preceded by saline infusion sonohysterography if the endometrial thickness is greater than 5 mm. A process must be shown to be symmetrically pan uterine or global to justify a blind procedure. When changes are focal (eg, polyps, some hyper-

VAGINAL ULTRASOUND
(ASAP after bleeding ends)
Thin Distinct Endometrial Echo 5 mm (Bilayer)
DX: DUB (Premenopause)
OR

Central Uterine Echo Thickened (>5mm) OR Endometrial Echo Not Adequately Visualized

Inactive EM (PM)

SALINE INFUSION SONOHYSTEROGRAPHY (SIS)


Thin EM (3 mm Single Layer) With No Focal Abnormalities Detected

Symmetrically Thickened EM (>3 mm Single Layer)

Focal Lesion OR Asymmetric Thickening

DX: DUB (Premenopause) OR Inactive EM (PM): No Further Diagnostic Procedures Necessary

Blind Office Endometrial Biopsy

Visually Directed Endometrial Sampling (Hysteroscopy)

Fig. 5. Clinical algorithm for ultrasound-based triage of any patients who have abnormal uterine bleeding.

420

Goldstein

Fig. 6. (A) Transvaginal scan in long-axis view of a perimenopausal patient who had abnormal uterine bleeding. The endometrial echo is not sufficiently seen along its entirety to make an accurate diagnosis. (B) Saline infusion sonohysterography reveals a lack of any endoluminal mass. The anterior and posterior endometrium measure 1.5 and 2.0 mm, respectively (calipers).

plasias, some carcinomas), they can be appreciated as such with fluid-instillation sonohysterography, and then directed biopsies must be performed. In the pilot study [20], although 9 of 21 patients had obvious sonographic and clinical evidence of fibroids, only 3 had a submucous component. Six of 21 had intramural-subserosal myomas coexisting with dysfunctional uterine bleeding. In the large prospective study [21], 148 of 433 women had myomas but only 22 had a submucous component. Usually, polyps are clearly discernable, as are submucous myomas. Sometimes, however, a broadbased polyp is difficult to distinguish from a

submucosal myoma. This distinction may be important for preoperative triage, in that a truly pedunculated submucous myoma behaves more like a polyp in terms of skill and equipment required for its removal in the operating room, whereas a broad-based polyp may behave more like a myoma and require resectoscopic capability. A reliable assessment with ultrasonography requires that the endometrial echo be homogeneous, that it is surrounded by an intact hypoechoic junctional zone, and that the operator constantly remembers that the endometrial cavity is a threedimensional structure. This fact may account for

Abnormal Uterine Bleeding

421

Fig. 7. Saline infusion sonohysterogram of a perimenopausal patient who had abnormal uterine bleeding. A polypoid lesion is seen extending near the anterior fundal region. This lesion measures 10.2 6.7 (calipers). At the time of dilation and curettage and hysteroscopy, a polyp was identified that was confirmed by pathology.

why Dijkhuzien and colleagues [22] had four cases that supposedly measured less than 10 mm (some as little as 2 mm) yet displayed polyps at hysteroscopy. Such cases underscore the importance of the three-dimensional character of the endometrial cavity and the occasional propensity of the ultrasonographic operator to obtain a limited number of two-dimensional views and assume that these represent the entire endometrial cavity. Any one frozen ultrasonographic image is nothing more than a two-dimensional snapshot, and failure to

meticulously recreate three-dimensional anatomy results in error. New three-dimensional ultrasound equipment can eliminate errors that may occur when the operator does not pay meticulous attention to mentally recreating three-dimensional anatomy. Furthermore, the use of color flow or power Doppler imaging to identify the central feeder vessel pathognomonic of an endometrial polyp is an alternative to sonohysterography in the diagnosis of polyps [Figs. 8 and 9]. This methodology had a

Fig. 8. Long-axis view of the uterus in a patient who had abnormal uterine bleeding. Color flow Doppler imaging clearly identifies a central feeder vessel. Presence of such a feeder vessel can be used to make the diagnosis of polyp even in the absence of saline infusion.

422

Goldstein

Fig. 9. Transvaginal pelvic scan of a patient 19 days since her last episode of bleeding. The endometrial surface is irregular. The irregular surface is not unusual, especially in patients who have had previous dilation and curettage procedures, myomectomies, childbirth, and so forth. The irregular surface to the endometrium, here identified as moguls, can be misleading. Performing the procedure this long after the last bleeding episode can be fraught with error and should be avoided.

positive predictive value of 81.3% in the study by Timmerman and colleagues [23].

Other considerations for sonohysterography Timing of the procedure


The uterus is an organ that has had multiple procedures in many women including D&Cs, child birth, myomectomy, Caesarean sections, abortion, etc. As the endometrium proliferates it is not always a smooth homogeneous layer. Sonohysterography is best performed as soon as possible after the bleeding cycle has ended when the endometrium is as thin as it is going to be all month long. Otherwise focal irregularities in the contour of the endometrium may be mistaken for small polyps or focal areas of endometrial hyperplasia [Fig. 9]. This was supported in a prospective blinded study by Wolman and colleagues [24] in which there was a 27% false positive rate in sonohysterography performed from day 16 to 28, while there were none when the procedure was performed prior to day 10. Sometimes the patient has such irregular bleeding that she can not tell what is an actual menses. It may be helpful in such cases to use an empiric course of a progestogen such as medroxyprogesterone acetate 10mg daily for 10 days as a medical curettage and then time the ultrasound evaluation to the withdrawal bleed.

other hand to change the position of the speculum will often modify the angle of the cervix with the fundus sufficiently to allow successful completion. Use of a tenaculum is a last resort. A cervical stabilizer will be less painful, less traumatic and does not cause bleeding from the cervix.

Anesthesia/analgesia
Anesthesia or analgesia is not required. In more than 1000 cases, I have seen three cases of a vasovagal response reminiscent of those occasionally seen with a plastic intrauterine device insertion in a nulliparous patient. The sonohysterography catheter is 1.8 mm in diameter and is remarkably painless in its insertion. The procedure is extremely well tolerated with no pain in the overwhelming majority of patients and minimal cramping in a very few.

Risk of infection
Sonohysterography should be handled similarly to traditional HSG. Thus, the decision about whether to obtain gonorrhoea or Chlamydia cultures as well as whether to use antibiotics will depend very much on the patient population with which the physician normally deals. In my experience, I have not routinely obtained cultures for sexually transmitted diseases nor have prophylactic antibiotics been used. In more than 1000 cases, I have not experienced any infectious morbidity. Of 1,153 procedures performed [25] the incidence of infectious complications that require surgical resolution was 0.7% which is similar to diagnostic hysteroscopy [26] but less than hysterosalpingography [27].

Difficulty threading the catheter


Occasionally, there will be difficulty in threading the catheter into its desired position. Using the

Abnormal Uterine Bleeding

423

Concern about spreading adenocarcinoma into the peritoneal cavity


Concern about spreading adenocarcinoma is a question of the benefit outweighing any theoretic risk [2427]. It is no longer standard practice to tie the fallopian tubes with silk before a total abdominal hysterectomy and bilateral salpingo-oophorectomy for endometrial carcinoma. Furthermore, hysteroscopy with saline or other distending media would have the same theoretic concern. Survival rates of patients who had endometrial carcinoma and underwent standard HSG were not different between patients who demonstrated intraperitoneal spill of the contrast medium and patients who did not [28]. Alcazar and colleagues [29] performed sonohysterography on 14 consecutive patients who had stage I adenocarcinoma of the endometrium. It was done at the time of laparotomy just when the abdomen was opened but before the start of the surgical procedure. All 14 readily spilled saline from the fallopian tubes. The fluid was analyzed, as were the cell washings. Only 1 patient (7%) had malignant cells in the spilled fluid, causing the investigators to conclude that the risk of malignant cell dissemination exists but is small.

present, a thin, distinct endometrial echo excludes significant pathology, assuming it is performed at an appropriate time if the patient is cycling. If a thin, distinct endometrial echo is not visualized (inadequate visualization or presence of thickened echo), then saline infusion sonohysterography can help to triage patients to (1) no anatomic pathology, (2) globally thickened anatomic pathology that may be evaluated with blind endometrial sampling, or (3) focal abnormalities that must be evaluated under direct vision. Such an ultrasound-based approach not only helps to exclude endometrial carcinoma but also identifies the source of any bleeding for better clinical management.

References
[1] Awwad JT, Toth TL, Schiff I. Abnormal uterine bleeding in the perimenopause. Int J fertile 1993; 38:2619. [2] SEER Cancer Statistics Review, 18731996 [serial online]. Available at: http://seer.cancer.gov/ csr/1973_1996/index.html. Accessed August 26, 2005. [3] Iatrakis G, Diakakis I, Kourounis G, et al. Postmenopausal uterine bleeding. Clin Exp Obstet Gynecol 1997;24:157. [4] ACOG practice bulletin: management of anovulatory bleeding. ACOG Committee on Practice BulletinsGynecology. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet 2001;72:26371. [5] Ricci JV. Gynaecologic surgery and instruments of the nineteenth century prior to the antiseptic age. In: Ricci JV, editor. The development of gynaecological surgery and instruments. Philadelphia: Blakiston; 1949. p. 3268. [6] Word B, Gravlee LC, Widemon GL. The fallacy of simple uterine curettage. Obstet Gynecol 1958; 12:6425. [7] Vuopala S. Diagnostic accuracy and clinical applicability of cytological and histological methods for investigating endometrial carcinoma. Acta Obstet Gyneocl Scand Suppl 1977;70:172. [8] Kaunitz AM, Masciello AS, Ostrowsky M, et al. Comparison of endometrial Pipelle and Vabra aspirator. J Reprod Med 1988;33:42731. [9] Rodriguez MJ, Platt LD, Medearis AL, et al. The use of transvaginal sonography for evaluation of postmenopausal size and morphology. Am J Obstet Gynecol 1988;159:8104. [10] Stovall TG, Photopulos GJ, Poston WM, et al. Pipelle endometrial sampling in patients with known endometrial cancer. Obstet Gynecol 1991;77:9546. [11] Guido RS, Kanbour A, Ruhn M, et al. Pipelle endometrial sampling sensitivity in the detection of endometrial cancer. J Reprod Med 1995;40: 5535. [12] Goldstein SR. Endovaginal ultrasound. 2nd edition. New York: Wiley Liss; 1991.

Inadequate distension of the cavity


In some patients, a patulous cervix results in a great deal of fluid running out transcervically. Other patients have fluid going out through fallopian tubes, even with slow injection and minimal pressure. As in hysteroscopy, some cavities are more difficult to distend than others. The clinician should check the position of the catheter, looking for its acoustic shadow most of the way to the uterine fundus. Unlike hysteroscopy (which requires distension for visualization), however, this procedure requires very little fluid to outline the cavity. Even a small ribbon of fluid acts as a sufficient interface to distinguish anterior and posterior endometrial surfaces and to outline endometrial pathology.

Summary
Abnormal uterine bleeding, whether it occurs in peri- or postmenopausal patients, is an important clinical concern and accounts for much medical intervention. When bleeding occurs in women older than 40 years (and in any postmenopausal woman), endometrial assessment is mandatory. In the past and even currently, many clinicians prefer to begin such assessment with blind endometrial sampling. This article presents an ultrasound-based approach to such patients. When

424

Goldstein

[13] Goldstein SR, Nachtigall M, Snyder JR, et al. Endometrial assessment by vaginal ultrasonography before endometrial sampling in patients with postmenopausal bleeding. Am J Obstet Gynecol 1990;163:11923. [14] Granberg S, Wikland M, Karlsson B, et al. Endometrial thickness as measured by endovaginal ultrasound ultrasonography for identifying endometrial abnormality. Am J Obstet Gynecol 1991;164:4752. [15] Goldstein RB, Bree RL, Benson CB, et al. Evaluation of the woman with postmenopausal bleeding: Society of Radiologists in Ultrasound Sponsored Consensus Conference statement. J Ultrasound Med 2001;20:102536. [16] Goldstein SR. Postmenopausal endometrial fluid collections revisited: look at the doughnut rather than the hole. Obstet Gynecol 1994;83:73840. [17] Randolph JR, Ying YK, Maier DB, et al. Comparison of realtime ultrasonography, hysterosalpingography, and laparoscopy/hysteroscopy in evaluation of uterine abnormalities and tubal patency. Fertil Steril 1986;46:82832. [18] Parsons AK, Lense JJ. Sonohysterography for endometrial abnormalities: preliminary results. J Clin Ultrasound 1993;21:8795. [19] Syrop C, Sahakian V. Transvaginal sonographic detection of endometrial polyps with fluid contrast augmentation. Obstet Gynecol 1992;79: 10413. [20] Goldstein SR. Use of ultrasonohysterography for triage of perimenopausal patients with unexplained uterine bleeding. Am J Obstet Gynecol 1994;170:56570. [21] Goldstein SR, Zelzter I, Horan CK, et al. Ultra-

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

sonography-based triage for perimenopausal patients with abnormal uterine bleeding. Am J Obstet Gynecol 1997;177:1028. Dijkhuzien FPHLJ, Brolmann HAM, Potters AE, et al. The accuracy of transvaginal ultrasonography in the diagnosis of endometrial abnormalities. Obstet Gynecol 1996;87:3459. Timmerman D, Verguts J, Konstantinovic ML, et al. The pedicle artery sign based on sonography with color Doppler imaging can replace second-stage tests in women with abnormal vaginal bleeding. Ultrasound Obstet Gynecol 2003; 22:16671. Wolman I, Groutz A, Gordon D, et al. Timing of sonohysterography in menstruating women. Gynecol Obstet Invest 1999;48:2548. Dessole S, Farina M, Rubattu G, et al. Side effects and complications of sonhystyerosalpingography. Fertil Steril 2003;80:6204. Cooper JM, Brady RM. Intraoperative and early postoperative complications of operative hysteroscopy. Obstet Gynecol Clin North Am 2000; 27:34766. Tuveng JM, Vold I, Jerve F, et al. Hysterosalpingography: value in estimating tubal function, and risk of infectious complications. Acta Eur Fertl 1985;16:1258. DeVore GR, Schwartz PE, Morris J. Hysterography: a 5-year follow-up in patients with endometrial carcinoma. Obstet Gynecol 1982;60: 36972. Alcazar JL, Errasti R, Zornoza A. Saline infusion sonohysterography in endometrial cancer: assessment of malignant cells dissemination risk. Acta Obstet Gynecol Scand 2000;79:3212.