Applied Radiology

Alpha-fetoprotein (AFP) screening was shown to be effective for detecting neural tube defects (NTDs) in the 1970s. ¹ In 1991, the American College of Obstetrics and Gynecology (ACOG) endorsed offering maternal serum (MS)-AFP testing to all pregnant women. Since then, screening in the United States became more wide-spread, and experience in this country and others has demonstrated considerable benefits from AFP screening, not only for the detection of NTDs but also for a number of other fetal abnormalities. Further, even in the absence of multiple gestations and discrete fetal defects, it is estimated that as many as 20% to 38% of women with unexplained high MS-AFP will suffer adverse pregnancy outcomes ^2-4 including fetal death,* growth retardation, premature birth, and pre-eclampsia. ^5,6 While the pathophysiology in these women is poorly understood, it has been hypothesized that placental ischemia is the underlying cause. ⁷

Currently, more than 300,000 pregnant women in California undergo AFP testing annually. Among the first 1.1 million women screened through the California AFP Screening Program, 1390 fetal anomalies (morphologic and chromosomal) were detected (prevalence of 1.3/1000). These included 710 NTD (417 anencephaly, 247 spina bifida, and 46 encephalocele), 286 ventral abdominal wall defects, 163 fetuses with Down syndrome, and 231 cases
of other chromosomal anomalies. Impressively, of all anomalies de-tected in this program, nearly three-quarters involved two organ systems: the neural axis (51%) and defects of the ventral abdominal wall (21%). This distribution of "likely" fetal anomalies is especially germane to the sonologist searching for the cause of elevated MS-AFP.

* These fetal deaths occur mainly in the second trimester, and the risk appears to be directly related to the degree of maternal serum alpha-fetoprotein elevation. ⁶

Alpha-fetoprotein: Where it comes from, how it gets there

AFP is a glycoprotein produced initially by the yolk sac and fetal gut, and later predominantly by the fetal liver. In the fetus, serum AFP level increases until approximately 14 to 15 weeks, and then falls progressively. In normal pregnancies, AFP from fetal serum enters the amniotic fluid (in microgram quantities) through fetal urination, fetal gastrointestinal secretions, and transudation across fetal membranes (amnion and placenta) and immature epithelium. Detectable quantities of AFP in the maternal serum (nanogram quantities) gradually increase during gestation, peaking at 30 to 32 weeks, and decline thereafter. Maternal serum levels are usually reported in multiples of the median (MoM) to standardize interpretation among laboratories.

Abnormal quantities of fetal AFP can enter the maternal serum in several ways. Among fetal defects, the most common mechanism is through fetal cutaneous defects, such as anencephaly or myelomeningocele. These defects result in leakage of fetal serum proteins into the amniotic fluid and, secondarily, into maternal serum. Intrinsic placental abnormalities and maternal-fetal hemorrhage, also allow fetal AFP to mix with maternal serum. In some cases, the precise mechanism for the fetomaternal transfer is not known (proximal gut obstruction, renal agenesis), and may be secondary to diminished fetal gut degradation or elevated fetal serum concentrations of AFP.

It would be ideal if MS-AFP levels were only elevated in abnormal fetuses or other pathologic conditions of pregnancy. Unfortunately, there is considerable overlap in MS-AFP levels between normal and abnormal pregnancies. Thus, the choice of a judicious cutoff value, which maximizes detection of anomalies and minimizes the number of false-positive results, must be chosen for this screening program to be effective. Many screening programs in the United States have settled on a serum a value of >= 2.5 MoM. Using this cut-off, approximately 90% of anencephalic fetuses, 75% to 80% of fetuses with an open spinal defect, 98% of fetuses with gastroschisis, and approximately 70% of fetuses with omphaloceles will be detected. ⁸ Further, using 2.5 MoM as the cut-off has resulted in a reasonably low screen-positive rate (approximately 4% to 5%). Other programs use a lower cut-off of >= 2.0 MoM, which results in a higher detection rate of these anomalies, with the trade-off of a higher false-positive rate.

MS-AFP screening programs: How patients are triaged

Maternal serum is tested between 16 and 18 weeks. If the MS-AFP is only marginally elevated (between 2.5 and 3.0 MoM) and the gestational age is below 17 weeks, the maternal serum may be retested. Accurate dating is critical for AFP screening because serum AFP levels rise approximately 15% per week during the 16- to 18-week window. MS-AFP values are also corrected for maternal weight, race, and the presence of diabetes (diabetes has a depressing effect on MS-AFP, so lower levels may be found in association with NTDs). ⁹

In California, approximately 2% of screened women have elevated MS-AFP levels (>= 2.5 MoM), and approximately 3% have MS-AFP levels ¾ 0.5 MoM. The latter will not be addressed here. Roughly 6% to 15% of women with high MS-AFP have some type of major congenital defect ^2,10,11 and this risk increases with the magnitude of MS-AFP elevation.

If MS-AFP is elevated, a non-targeted standard antepartum obstetrical ("Level 1") sonogram is performed for the purpose of identifying easily recognized causes of "false-positives" (gestational age >= 2 weeks more advanced than estimated clinically, multiple gestations, fetal death, and obvious fetal defects). The intent of a Level 1 sonogram is to provide a general assessment of fetal/pregnancy health and is performed according to the published guidelines endorsed by the AIUM and ACR. ¹² Impressively, 20% to 50% of the elevated MS-AFP levels will be explained by findings on this preliminary sonogram. ^13,14 If the cause of elevated MS-AFP is not explained, traditionally the next step has been to offer amniocentesis for measurement of amniotic fluid (AF)-AFP. Among women who choose to undergo amniocentesis following a normal Level 1 sonogram, the majority will have normal AF-AFP (< 2.0 MoM), and there is no further diagnostic evaluation. ¹⁵ If the AF-AFP is elevated (>= 2.0 MoM), acetylcholinesterase (an isoenzyme important in neurotransmission) is tested on the amniotic fluid sample. Acetylcholinesterase is present in association with exposed neural tissue (and occasionally with abdominal wall defects). High AF-AFP plus positive acetylcholinesterase is quite specific for a fetal defect. In most screening programs, karyotype testing is also routinely performed on the amniotic fluid specimen.

If the AF-AFP is elevated (>= 2.0 MoM), a targeted fetal sonogram ("Level 2") is offered, because approximately one-third of fetuses are anomalous. ¹⁵ Similar to MS-AFP, the likelihood of a fetal defect increases proportionately with the degree of AF-AFP. The Level 2 sonogram is performed in these cases to determine: 1) whether any fetal anomaly is present: (elevated AF-AFP may be false-positive); 2) if present, what the nature of the anomaly is (i.e., NTD versus omphalocele); and 3) if present, the severity of the anomaly (i.e., the spinal level of the myelomeningocele) and to search for other anomalies.

AF-AFP testing is a highly sensitive method for detecting or excluding NTDs. The negative predictive value of a normal AF-AFP is approximately 97% and elevated AF-AFP plus acetylcholinesterase allows > 99% accurate detection of NTD. ¹⁶ Targeted ultrasonography performed in conjunction with abnormal AF-AFP is also highly accurate in identifying anomalous fetuses (i.e., >99% accurate). ^15,17 However, there is a small but important procedural fetal loss rate associated with amniocentesis, generally quoted to be approximately 1/200 (0.5%), which is not present with the sonographic examination. As a result, women with elevated MS-AFP, in increasing numbers, opt to go directly from the serum AFP test to a targeted fetal sonogram (i.e., to skip the amniocentesis). Indeed, this method has been adopted in the United Kingdom and detailed, targeted sonograms are now routinely performed as the next step in women with high MS-AFP. This approach has become more popular in the last few years for two major reasons. First, sonographic detection of the "likely" anomalies associated with high MS-AFP is currently highly accurate. Expected rates of sonographic detection for neural tube and abdominal wall defects are currently >90% (several series report 100% detection rate for NTDs). ^17-21 It is estimated that a complete, detailed, normal sonogram can now reduce the MS-AFP-based risk of a neural tube or ventral abdominal wall defect by 95%. ^22,23 Second, going directly to Level 2 sonography circumvents the small but important procedural risk of fetal loss from amniocentesis. It has been suggested that the slightly higher diagnostic yield of amniocentesis (as compared to ultrasound) may not be justified, given both the higher cost and procedure-related fetal loss rate associated with this approach. ²⁴

Some have cautioned against adopting a routine policy of circumventing the amniocentesis ²⁵ because: 1) this approach will require a much larger number (i.e., ten times as many) of targeted sonograms, and the larger number of experienced examiners may not be available or patients may be required to travel a long distance for the targeted sonogram; 2) that even "experienced" examiners, scanning a population with a lower prevalence of defects may not detect as many defects as AF-AFP testing; ¹⁹ and 3) skipping amniocentesis will cause potentially detectable chromosomal abnormalities to be missed. The last issue remains controversial. Some favor a paradigm in which Level 2 sonography follows a high MS-AFP, arguing that there is only a very small risk of an abnormal karyotype in a fetus without morphologic defects. If the targeted sonographic fetal survey in a woman with elevated MS-AFP is normal, it has been estimated that the risk of a fetal chromosomal abnormality is only 0.6% to 1.1%, ^22,26-30 and of the chromasomally abnormal fetuses, sex chromosome aberrations (other than 45X) account for many (30% to 50%). ²⁶

There is no "right" choice, but all women facing the choice of Level 2 sonography versus amniocentesis should be fully informed of these controversies during their counseling. Decisions to do amniocentesis versus Level 2 sonography will vary according to patient (maternal age, other seriologic markers [i.e., hCG, estriol], and personal choice) and institution (depending on availability of experienced sonologists). Amniotic fluid testing should still be strongly considered in the following patients: 1) fetal position or maternal body habitus precludes an adequate sonographic fetal anatomic survey; 2) equivocal sonographic findings (i.e., abnormal posterior fossa but spinal defect not seen); 3) experienced sonographic examiner not available; and 4) non-lethal anomaly detected on Level 1 sonogram, for which karyotype testing is appropriate.

Increased MS-AFP: What should you look for?

The neural axis and ventral abdominal wall will be the most critical regions under scrutiny during the Level 2 sonogram. A focused examination of the neural axis in each fetus should include an assessment of overall cranial size and contour, ventricular size (transaxial diameter of ventricular atrium >10 mm is abnormal), ³¹ posterior fossa including cerebellar morphology, and cisterna magna. ^31-33 At UCSF, we also include images of the cavum septum pellucidum as a check for forebrain malformations. The spine should be carefully examined in each fetus, including segment by segment images in the transaxial and sagittal planes from the craniocervical junction through the sacrum. The normal curvature of the spine and intact dorsal skin line should be documented and the ossified posterior elements should be examined for abnormal splaying. The ventral abdominal wall of the fetus is examined with focused attention on the umbilical cord insertion. The examiner should maintain a heightened sensitivity to the presence of bowel loops within the umbilical cord or floating in the amniotic fluid distant from the cord insertion/abdominal wall.

The most commonly encountered individual defects are discussed below.

Anencephaly--Anencephaly accounts for approximately half of all NTDs (figure 1). On average, anencephaly is associated with the highest AF-AFP and MS-AFP values of all NTDs, and approximately 90% will be detected by an MS-AFP >= 2.5 MoM. This is a lethal anomaly in which the bony calvarium is absent above the orbits. Normal cerebral cortex is absent. Some dysplastic "brain tissue" histologically representing angiomatous stroma, may be observed above the orbits, apparently floating freely in the amniotic fluid. Owing to its irregular shape and the absence of recognizable normal morphology, it is unlikely to be confused for normal brain. One should be cautious, however, not to confuse an engaged fetal head (in which the convexity may not be well visualized) for anencephaly. This is accomplished by the observation of amniotic fluid above the orbits and the calvarial defect. It is critical that this diagnosis is accurate because most patients will electively terminate their pregnancies following this diagnosis. Anencephaly can be diagnosed in virtually all affected fetuses after 14 weeks gestation. ³⁴

Myelomeningocele--Myelomeningocele occurs in approximately 1:1000 live births in California (incidence is slightly higher in the southeastern United States). The myelomeningocele sac can be detected on sagittal or transverse views. Even in the absence of a sac, myelomeningocele is suggested by a defect in the normal smooth dorsal skin line and splayed posterior ossification centers on the transaxial image (figure 2). Widening of the posterior ossification centers can also be seen on coronal images of the spine. The majority of spinal dysraphisms occur in the lumbosacral region, so this area should be scrutinized with extra care. Very abnormal spine curvature may be associated with the amniotic band syndrome or limb body wall complex. When the fetus is in breech presentation, endovaginal scanning can be helpful.

Spinal defects can be small and inconspicuous. This is undoubtedly the reason that sonographic detection was only mediocre (50% to 80%) in reports from the early 1980s. ^9,35,36 Descriptions of several important cranial findings associated with "open" spina bifida have been tremendously beneficial for improving the sensitivity with which spina bifida is detected antenatally. Cranial findings associated with "open" (non-skin covered) fetal myelomeningoceles include the "lemon sign," ^22,37-39 "banana sign," ³⁹ effaced cisterna magna, ³² ventriculomegaly, ^31,40 and small biparietal diameter (BPD). ⁴¹ At least one of these findings is present in >99% of affected fetuses. ²²

The lemon sign (figure 3A) describes an inward scalloping of the frontal cranial bones seen in nearly all second-trimester fetuses with open spina bifida, but tends to disappear in affected fetuses in the third trimester. Importantly, the lemon sign may also be seen in as many as 1% of normal fetuses ^38,42 in addition to a number of other neural axis anomalies diminishing its positive predictive value for spina bifida. ⁴³ Thus, the diagnosis of spina bifida should never be based solely on the observation of a lemon sign.

The banana sign (figure 3B) and the effaced cisterna magna occur secondary to the hindbrain malformation known as the Chiari II malformation, which is present in almost all (>95%) fetuses who have open spinal lesions. The bony posterior fossa is small in the Chiari II malformation and the developing cerebellum is cramped. The crowded cerebellum appears to wrap around the brain stem (creating a transaxial cerebellar configuration akin to the shape of a banana) or, at the minimum, the cisterna magna is completely or nearly obliterated. The banana sign is highly specific for the Chiari II malformation but not quite as sensitive as effacement of the cisterna magna (some of the posterior fossa deformities of the Chiari II malformation are not severe enough to produce a banana cerebellum). Effacement of the cisterna is more sensitive for detection of the Chiari II malformation, but less specific (it can be seen in association with hydrocephalus). The cisterna magna (3 to 10 mm) can be visualized in 97% of normal fetuses at 15 to 25 weeks gestation. Because the Chiari II malformation (and myelomeningocele) is nearly always associated with an abnormal-appearing posterior fossa, a small or absent cisterna should raise suspicion for a spinal lesion. As a corollary, a normal-appearing cerebellum and cisterna magna, has a high (>98%) negative predictive value for the Chiari II malformation. Thus, a normal-appearing posterior fossa reduces the risk of an open spinal lesion by >98%.

Other cranial findings associated with spina bifida include a BPD which is small for dates (second trimester) and ventricular enlargement. The degree of ventricular dilatation in fetuses with myelomeningoceles tends to increase with gestational age. ⁴⁰ In a series of 51 fetuses with spina bifida aperta (nonskin-covered), we found ventriculomegaly (atrium >10 mm) was present in only 44% of myelomeningocele fetuses examined before 24 weeks, but present in 94% of fetuses scanned in the third trimester. ⁴⁰ The degree of ventriculomegaly is also related to the degree of visualized posterior fossa deformity, ⁴⁰ but not to the spinal level of the lesion. ⁴⁴ It should be emphasized that even though the cranial findings have so greatly improved the sensitivity with which we can detect myelomeningocele (currently reported to be >90% and >95% in many centers), ^15,23 the final diagnosis of a myelomeningocele should only be made after direct observation of the spinal defect.

If a myelomeningocele is detected

Outcome of fetuses with mye-lomeningocele is influenced by the presence of associated malformations, chromosomal abnormalities, the level of the spinal lesion (children with higher lesions have more severe motor handicaps), and childhood shunt infections. ⁴⁵ Prenatal sonography can offer important and accurate information regarding the presence of ventriculomegaly, the level of the spinal lesion, and the presence of associated malformations. The bony level of the defect can be accurately estimated (± one spinal level) sonographically in 79% of fetuses. ⁴⁶ This is accomplished in most cases by "counting up" from the last sacral ossification center (assumed to be S4 in the second trimester and S5 in the third). Associated malformations (in addition to the Chiari malformation and hydrocephalus) are present in 13% to 24% of fetuses with myelomeningocele. ^18,40,47 Multiple malformations increase the likelihood of fetal karyotype abnormalities, ^48-50 but chromosomal abnormalities are reported in 10% to 15% ^49,51 of fetuses with "isolated" myelomeningocele. Thus, if the parents plan to carry the pregnancy, it is prudent not only to perform a complete, detailed, fetal anatomic survey but also to offer fetal karyotype testing.

Cephalocele--Cephaloceles are relatively rare (1.2/10,000 births), midline cranial defects that contain meninges, cerebrospinal fluid (meningocele), ± neural tissue (encephalocele). ⁵² These lesions only account for approximately 3% of fetal anomalies detected with MS-AFP screening and 6% of detected NTDs. ^25,53 In the United States, most (80% to 85%) of these occur in the occipital location, ⁵⁴ a small percentage occur in the frontal (10% to 15 %) or parietal (10 to 15%) area. Because encephaloceles occur in the midline, off-midline cranial defects should suggest the presence of the amniotic band syndrome. Most occipital lesions are associated with abnormalities of the posterior fossa and parietal lesions may also be associated with the Chiari malformation. The face and orbits should be examined carefully. The interorbital distance is usually widened in association with a frontal encephalocele.

Prognosis of fetuses with cephalo-celes is generally poor (only 21% live-born in our series) ⁵⁴ and outcome is related to the presence of associated neural and non-neural malformations (common), as well as the size and content of the lesion. Poorer outcome is associated with a large volume of herniated brain. Associated brain malformations include the Dandy-Walker malformation, agenesis of the corpus callosum, cerebellar hypoplasia, and migrational abnormalities. Karyotype abnormalities are common, found in 44% of tested fetuses in one report. ⁵⁴ It is important to remember that many encephaloceles are skin-covered (60% in one series) ⁵³ and, therefore, may elude detection with MS-AFP screening. ⁵⁵ Occipital cephaloceles may occur as part of the heritable (autosomal recessive) Meckel's syndrome (encephalocele, cystic dysplastic kidneys, and polydactyly). ⁵⁶ Affected pregnancies may be terminated without adequate pathologic diagnosis. Therefore, prenatal recognition of this potential syndromic association is important for counseling regarding future pregnancies, because the 25% risk of recurrence associated with Meckel syndrome greatly exceeds the recurrence risk of other cephaloceles (3%). ¹⁶

Ventral abdominal wall defects--Ventral abdominal wall defects include omphalocele, gastroschisis, and abdominal wall defects associated with the amniotic band syndrome or limb body wall complex. Scrutiny of the fetal umbilical cord insertion and ventral abdominal wall allows sonographic detection of omphalocele and gastroschisis in >90% of fetuses. ^15,23 Omphaloceles occur in approximately 1:4,000 live births, and include a spectrum of midline defects that range from large (usually containing liver and bowel) (figure 4) to small (which may contain only 1 or 2 bowel loops). The exteriorized viscera are contained by an anu-amnioperitoneal membrane, and the umbilical cord inserts midline into the sac. Features most predictive of prognosis are other serious malformations (expected in 50% to 75% of affected fetuses, including cardiac malformations in 30% to 35%) and chromosomal abnormalities (approximately 10% to 20%), mainly trisomies 18 and 13. Although "bowel-only" omphaloceles are generally smaller, less conspicuous sonographically, and often easier to repair postnatally, the rate of chromosomal abnormalities (perhaps 70% to 80%) is 8 to 10 times higher than that found in fetuses in whom the omphaloceles contain liver within herniated sac. ^37,57 Be aware that small bowel-only omphaloceles may contain only one or two loops of bowel that have migrated into the cord so that the abnormality may not be recognized solely by examination of the cord insertion into the fetal abdomen. Thus, examination of the umbilical cord beyond the fetal abdomen for several centimeters is prudent.

Gastroschisis is a full-thickness, paramedian, abdominal wall defect, usually occurring to the right of the fetal umbilical cord insertion, through which bowel is exteriorized. Importantly, there is no covering membrane (figure 5). Associated malformations other than gut malrotation and atresia are rare, and the prevalence of chromosomal abnormalities is not increased. Fetal growth retardation is seen in up to 40%. ⁵⁸ Bowel dilatation and mild thickening is common as gestation progresses and loosely correlates with seriously damaged bowel requiring resection postnatally. ⁵⁸ Early in the second trimester (<20 weeks), gastroschisis may be difficult to observe. The defect in the abdominal wall is small (1 to 3 cm) and, early on, the bowel is usually nondilated.

A large population-based study involving 72,782 consecutively screened pregnancies was used to establish distributions of AFP in pregnancies with gastroschisis and omphalocele. ⁸ Based on a cut-off of 2.5 MoM, all fetuses with gastroschisis (20/20) and approximately 70% (10/18) of omphaloceles were detected during MS-AFP screening.

Less commonly observed fetal defects associated with elevated MS-AFP--A number of other important fetal anomalies are associated with elevated MS-AFP, and these potential defects should also be sought on the targeted sonogram. Less common defects include fetal teratoma (pharyngeal, sacral), defects caused by the amniotic band syndrome (asymmetric cephaloceles, gastropleuralschisis), cystic hygroma, lesions that alter the placentomaternal barrier (i.e., placental chorioangioma, lakes, and abruption/ hemorrhage), proximal fetal gut obstructions (i.e., esophageal and duodenal atresias), some renal abnormalities ⁵⁹ (including multicystic dysplastic kidney, pelviectasis, congenital or Finnish nephrosis), and oligohydramnios. Thus, careful examination of the face, posterior neck, oropharynx, thorax, and abdomen (including a normally filled stomach) should be performed. ⁶⁰ The limbs and digits should be assessed for abnormalities suggesting the anu-amniotic band syndrome or VACTERL (vertebral, anorectal, cardiac, tracheoesophageal fistula, renal, and limb anomalies) association. Amniotic fluid volume should be qualitatively or semiquantitatively assessed in addition to careful examination of the placenta.

The AF- and MS-AFP may be elevated in fetuses with cystic hygroma (CH). While the precise mechanism is not known, it is speculated that fetal serum proteins may leak through the membrane/integument covering the CH, or perhaps enter the maternal blood through an intrinsic placental abnormality associated with an abnormal karyotype (present in 60% to 80% of second- and third-trimester fetuses with CH).

Teratomas (most commonly sacral [figure 6], but also oropharyngeal and lingual) can grow to a very large size in fetal life. These tumors often ulcerate, allowing leakage of fetal protein into the amniotic fluid and, secondarily, into maternal serum. In many cases, they are not completely skin-covered. Transverse axial views of the oropharynx, coronal and axial views of the face (to exclude oropharyngeal and lingual teratomas), and transverse and longitudinal views of the sacral area (sacrococcygeal teratomas are most common) should be obtained in patients referred for elevated MS-AFP. How sensitively teratomas are detected by MS-AFP screening is not known.

Esophageal atresia and duodenal atresia have been associated with elevated AFP. Some have speculated that a smaller than average degradation of swallowed AFP might account for the AFP elevation. A normally filled fetal stomach and the absence of a persistently filled or dilated duodenum should be sought. The normal fetal duodenum empties immediately and a persistently filled duodenum (even if it does not appear "over-distended") is always abnormal. The presence of a fetal "double bubble" suggests duodenal obstruction (usually atresia, but can be due to stenosis, Ladd's bands, or annular pancreas). Importantly, nearly one-third of fetuses with duodenal atresia have Down syndrome. Thus, if a double bubble is detected, a focused examination of the fetal heart is performed and karyotype testing is offered to the parents.

Esophageal atresia is suggested by an absent/unfilled stomach and polyhydramnios; but this constellation of observations is insensitive (<50%) for the sonographic detection of fetal esophageal atresia before the third trimester. This is due to the fact that the proximal esophageal pouch is only rarely seen in fetuses, and a fistula exists between the lower esophagus and bronchial tree in >90%, allowing passage of some fluid into the fetal stomach. A small, not absent stomach was observed in 5 of 12 fetuses with proven esophageal atresia by McKenna et al. ⁶⁰ In addition, frank polyhydramnios, is typically not seen before 20 to 24 weeks gestation.

Renal abnormalities including congenital (Finnish) nephrosis, multicystic dysplastic kidney, renal agenesis, and pelviectasis have been associated with elevated MS-AFP. ^59,61 In some cases, the AFP is elevated secondary to abnormal leakage of proteins into fetal urine. Congenital nephrosis results in a dramatic fetal proteinuria in utero and, because there are no renal morphologic features, this is a very difficult diagnosis to make with certainty antenatally. The clue to the diagnosis is that both the MS- and AF-AFP levels are extremely high (i.e., typically >= 10 MoM!) with negative amniotic fluid acetylcholinesterase and without evidence of maternal-fetal hemorrhage or other fetal morphologic defects. ⁶² In cases of renal agenesis, the mechanism for MS-AFP elevation is not known, but it is speculated that these fetuses may have higher serum protein levels owing to diminished excretion.

Finally, placental abnormalities including chorioangioma, placental abruption, periplacental hemorrhages (i.e., subchorionic hemorrhage), and placental lakes may result in elevated MS-AFP. A careful examination of the placenta should be performed. ^63-65 Relatively minor placental abnormalities (i.e., placental lakes, large marginal veins) are seen commonly in pregnancy patients. Although the placenta should be carefully examined in all women referred for high MS-AFP, this placental lesion should be the diagnosis of exclusion (after morphologic defects have been excluded), as the cause of increased MS-AFP. AR