Obstetric imaging

Introduction

 

Diagnostic imaging in pregnancy fulfils two major requirements. The first and numerically most frequent is in screening and monitoring the pregnancy and the second is in assisting in diagnosing the complications of pregnancy.

Screening and monitoring the progress of pregnancy

Ultrasound is without doubt the technique of choice in this aspect of obstetrics. It has naturally evolved into three relatively distinct areas of application with differing degrees of involvement of radiologists:
1. Routine antenatal screening
2. Fetal morphology and abnormality
3. Fetal well-being

Routine antenatal screening
This screening is most commonly carried out by radiographers, technicians and midwives, and usually consists of establishing fetal number, measurements to confirm or establish gestational age, placental site localisations and varying degrees of fetal morphology check, depending on the skill and experience of the operator.
The measurements usually carried out are Crown Rump Length (CRL), Biparietal Diameter (BPD), Abdominal Circumference (AC), and Femur Length (FL). These are too well known to need further elaboration. But there are many others: Amniotic Fluid Index (AFI) which is becoming increasingly popular for documentation of amniotic fluid volume; Occipito Frontal Diameter (OFD), Head Circumference (HC), which may be used when the BPD is not measurable, and when the head shows unusually marked degrees of dolichocephaly or brachycephaly; Transverse Cerebellar Diameter (TCD), which is also useful for gestational age estimation and is independent of head shape. Virtually, anything measurable in the fetus has been measured, but many of these measurements have specific rather than general usage.

The use of routine antenatal screening has recently been criticized as being ineffective in reducing adverse perinatal outcome and adding unnecessary cost to antenatal care. However, the conclusions of these investigations do no more than support the general principle of the law of diminishing returns. In any selected low risk population the chances of abnormal findings are diminished and the cost of finding them will be high. However, one cannot extend these obvious conclusions to the whole obstetric population which includes high risk and uncertain risk mothers. For instance, one of the major functions of routine screening is to establish gestational age (GA). Without having an accurate GA how could you identify postmaturity, premature labour or premature rupture of membranes if you do not know what time to call "maturity"? How could you monitor intrauterine growth retardation? Furthermore, could you interpret the results of laboratory data such as AFP and HCG with sufficient accuracy?

Up to 30% of pregnant women may not be accurate in their reported LMP because of irregular cycle, first trimester spotting, breast feeding, oral contraception or just plain fecklessness. Fundal height may be confused by multiple pregnancy, obesity or fibromyomas.

A CRL below 35 mm or a BPD between 12-18 weeks has a 90% accuracy predicting GA compared with 70% based on menstrual age. Combined measurements are even more accurate.

To suggest that one can practice modem obstetrics without the use of routine US is to recommend a return to the dark ages!
Ideally, all pregnancies should have routine screens at least twice. First at 18 weeks to establish GA and exclude major anomalies and the second at 32 weeks for fetal growth and placental site. Pregnancies with identifiable risk factors may need more.

Fetal abnormality
The demonstration of fetal abnormality involves radiologists, as well as radiographers and obstetricians. Radiologists should be able to contribute significantly in this field with their experience in imaging and anatomy and with MRI looming on the horizon as a powerful supplementary technique for the fetus. However, as diagnosis of fetal abnormality is generally less useful towards the end of pregnancy there has been great pressure to make the diagnoses earlier. This is not surprising, as all major congenital malformations originate in the embryonic period (six to ten weeks) with the exception of the external genitalia which take up to 14 weeks to achieve fetal maturity. There is increasing effort to make the diagnosis can be made as early as the first trimester. In this use a TVUS is essential to visualize the embryonic and early fetal anatomy. This has been given much impetus in the assisted reproduction programmes and can be put to good use in early threatened abortion.

Fetal well-being
The majority of findings in routine antenatal US tend to show the end result, of what has happened to the fetus in its intrauterine environment up to that point of time. The aim of fetal well-being studies is more to predict what is likely to happen in the future of that pregnancy based on mainly physiological parameters. Such studies of fetal physiology involve the considerable use of Doppler wave form studies, and available time. As radiology is a shortage speciality and this territory appears to be the prerogative of the obstetrician it will not be mentioned further here.

Complications of pregnancy

The complications of pregnancy that are most likely to require the services of the radiologist, and the techniques that may be potentially useful, are:

- Threatened abortion                    (US, TVUS, CFI)
- Ectopic pregnancy                       (US, TVUS, CFI)
- Trophoblastic disorders              (US, TVUS, CFI)
- Fetal abnormality                          (US, TVUS, CFI, MRI)
- Placenta praevia                          (US, TVUS, CFI, MRI)

- Pelvis mass discovered 
        during pregnancy                    (US, TVUS, CFI, MRI)
- Deep vein thrombosis                  (US, CFI)
- Eclampsia                                     (MRI)

Threatened abortion
The radiologist may be requested to investigate PV bleeding with or without pain in early pregnancy. The common causes will be pseudomenstruation with a viable pregnancy, ectopic and molar pregnancy (see below)

Figure 1. A 3D US of a living embryo within its gestation sac next to an aborting twin (asterisk). The umbical cord (arrowhead), limb buds (small arrows) and yolk sac (large arrows) of the living twin are clearly visible.

aborting twin (Fig. 1), missed abortion, and incomplete abortion.
Crucial to the management of such cases is the necessity to demonstrate a living embryo or fetus in a healthy gestation sac. This is achieved by demonstration of what is usually termed fetal heart motion (strictly speaking, it should be embroynal heart motion). If this can be demonstrated the chances of a successful ongoing pregnancy will be between 90-97%. It is usually clearly visible and there is no need for M-mode scan or Doppler other than for documentary proof.

Signs that the gestation has failed are:
- an irregular shaped sac
- poor decidual reaction with poor delineation between decidua capsularis and parietalis giving a "thin-walled" sac appearance
- a low position of the sac (just above the internal os)
- no embryo visible in a sac of mean diameter in excess of 25 mm
- no heart activity in an embryo with a CRL greater than 5 mm (ca. 6.5 weeks)
- no embryonic body activity with a CRL greater than 15 mm (ca. 8 weeks)
- no embryo within a gestation sac containing a yolk sac of 5 mm

Ectopic pregnancy

The presence of an ectopic pregnancy can now be very confidently established even without the necessity to measure serum RCG although this is still a useful adjunct.

Transabdominal US is no longer considered to be sufficient. The best results are obtained with TVUS and CFI.

The signs to seek are:
- empty uterine cavity
- up to 14 % may show a decidual cyst deep in the endometrium not reaching the endometrial interface
- blood in the uterine cavity may be mistaken for a sac, but it will not contain either an embryo or yolk sac
- differential blood flow demonstrable between the uterine blood vessels in the two cornua. This is difficult to quantify by Doppler wave forms but relatively easy to see
- a cystic mass in the adnexae surrounded by a hypervascular ring, described as being a "ring of fire" on CFI. Unusually, the mass may not be cystic. The waveform of the vessels in the ring is the characteristic low impedance, pregnancy type, flow.
- the cyst may contain a recognizable embryo and in about 10% even a demonstrable heart beat if the embryo's CRL measures beyond 10 mm (7 weeks)

If there is still doubt, measure the serum RCG. If the RCG level exceeds 1,800 MIU per ml there should be an intrauterine sac visible with TVUS. Go back and recheck. If still negative a follow-up of RCG and TVUS must be performed. Copious blood in the cul de sac is a late sign! If doubt still remains, the gynaecologist/obstetrician will get fed up and do a laparoscopy!

Early diagnosis of ectopic pregnancy allows for non-surgical interventional treatment with TVUS guided direct injection of Methotrexate into the ectopic sac.

Trophoblastic disorders: hydatidiform mole and choriocarcinoma
The diagnosis of hydatidiform mole is easy with conventional ultrasound. The presence of a fetus or fetal parts suggests the possibility of a partial mole. The fetus will show congenital malformations and karyotyping is often triploid. These moles are always benign.

The imaging diagnostic problem arises with the complete moles. These have a 46XX karyotype and about 20% progress to choriocarcinoma. When the uterus has been evacuated but the RCG fails to return to normal in 12 weeks or still remains high or begins to rise again, and if there is a persistent corpus luteum cyst, TVUS and CF! is again essential to seek the site of remaining trophoblast. A chest x-ray is necessary to exclude lung metastases.

The conventional grey scale image may appear to be normal or only mildly abnormal but CFI invariably reveals a wealth of unsuspected vessels indicating the presence and location of persistent trophoblast/choriocarcinoma.

Fetal abnormality
There are now excellent reference texts concerned with fetal abnormality and virtually half of the subject matter in obstetric ultrasound texts is devoted to this subject. It would be impossible to do justice to the subject in just a few pages, but some general observations and comments are pertinent.

1. The diagnosis of fetal anomaly is difficult and considerable expertise is required.
The German protocol in which suspected fetal anomalies have to be sent to specialist centres is an excellent model.

2. Karyotyping facilities are an essential complementary feature for fetal anomaly diagnosis.
Karyotyping is a boon for the problem of the suspected abnormal fetus in which only soft signs can be demonstrated. Any ultrasound service without this back up is incomplete.

3. The frequency of abnormalities within any particular ethnic or racial group is subject to variation.
For instance, urinary tract anomalies are relatively rare in the Chinese, common in Europeans and very common in the Middle Eastern races. Spina bifida is also relatively rare among Chinese but common in Europeans. But Down's syndrome and anencephaly appear to be fairly steady in all races.

Figure 2. Acrania. Free floating brain is seen above a microopthalmia (arrow).

4. Significance of finding a fetal abnormality
When a fetal abnormality is discovered one must carefully categorise it. Is it:
a) Incompatible with fetal survival?
Example: the majority of fetuses with severe cystic hygromadie in utero. Early diagnosis will not make a major difference to the outcome. It may enable an earlier termination and alert to possible risks in subsequent pregnancies.
b) Incompatible with survival post neonatal or early infant age.
Example: renal agenesis, bilateral multicystic kidneys, hypoplastic left heart, acrania (Fig. 2), trisomy 18, hydranencephaly, lethal dwarfism. Early diagnosis will enable early termination and avoid considerable parental distress and distress of the medical staff that would otherwise be left to care for the doomed baby.
c) Compatible with extrauterine life but incompatible with an independent existence?
Example: microcephaly, spina bifida, Down's syndrome. Early diagnosis is essential to make an informed decision regarding possible termination.
d) Compatible with a normal independent life following rep air of the abnormality?
Examples: posterior urethral valve with well pre serve d kidneys, pelviureteric junction obstruction, choledochal cyst (Fig. 3 a, b), cleft lip, mesoblastic nephroma.
Prenatal diagnosis is beneficial because the parents and clinicians can be forewarned and appropriate treatment can be planned ahead, including

a	Figure 3. a) Transverse scan of a fetal abdomen (19 weeks GA) with a choledochal cyst (arrow). b) Longitudinal scan showing the sinus venosus anterior (arrowhead) and IVC posterior (arrow) to the choledochal cyst.
b

transfer to a specialist centre if considered necessary. The discovery of urinary tract abnormalities at this stage may prevent children presenting in later life in an irreversibly damaged state.
e) Compatible with normal life as the abnormality is self-limiting and
not essential?
Examples: ovarian cyst, fetal adrenal haemorrhage, multicystic kidney. Following diagnosis, the parents require reassurance and perhaps follow-up as in the case of the multicystic kidney, where there may be a 25 % chance of developing a pelviureteric junction obstruction on the contralateral side.
f) Compatible with a normal life despite a permanent deformity such as
achondroplasia?
g) The last group is those in whom the echographically suspected diagnosis is uncertain.
Not diagnosing or missing abnormalities of groups a), b), e) and f) will not alter the long term outcome even though it could affect the quality of care provided. Missing group d) lesions may delay necessary corrective treatment. Missing group c) abnormalities is the most serious error - committing the unfortunate family to decades of unhappiness in the majority of such cases. In group g) there is scope for further delineation if karyotyping and MRI are available.

5. Ultrasound alone has not achieved as high a degree of sensitivity as might be hoped for.
It is reasonably good at identifying the more gross abnormalities (60-90 %) but may miss the more subtle ones, either when these are solitary or when they make up essential parts of a syndrome where the abnormalities may not be functionally related. For instance, a choledochal cyst is a solitary abnormality, but posterior urethral valve may produce functionally related bladder hypertrophy, hydroureter and hydronephrosis. The inexperienced examiner may then miss the bladder changes and report only hydronephrosis. Spina bifida is accompanied by a Chiari II malformation but this may be difficult to demonstrate with US. In conditions such as Zellweger's syndrome, the brain abnormality or the renal abnormality may be identified alone and the lack of appreciation of the other leads to a failure of accurate diagnosis.

In some conditions signs are "soft" apart from some disagreement of measurements such as head circumferences vs. abdominal circumference in Trisomy 18, in which there otherwise may be little else to see. It may be that 3D US will enable better visualization of these soft signs. The well known paediatric diagnostic category of "funny looking kid" could well become "funny looking fetus" in the future (Fig. 4). In adult and paediatric medicine one has a whole armamenterium of investigations. With fetal medicine these have been limited until karyotyping and MRI became available.

MRI of fetal abnormality presents unique problems. In addition to the thermal problems already alluded to, fetal motion requires fetal paralysis. The most commonly used technique currently is ultrasound guided umbilical vein injection of pancuronium bromide. But this is a technique not without significant hazard of fetal demise! Rapid imaging techniques,

Figure 4. 3D US of a fetal face.

such as echo planar sequences suffer from a relatively limited spatial resolution and therefore may not be suitable for some organs. Because of the rapid rise times and high gradients heat deposition will be high but this can be balanced by the shorter scan time. Transmitted maternal aortic pulsation can be minimised by placing the mother in a decubitus position. The fetal brain has a high water content providing some further constraints to MRI diagnosis, especially as there has been a greater interest in the fetal brain than other parts.

The reported experience does appear to be clustered into the last trimester and is invariably supplementary to an echographically disclosed abnormality. In some of these cases there were doubts and difficulties (consistent with group g) above). There is generally reasonable agreement between MRI and US but in certain cases MRI does add further information of clinical value, and is likely to become an important supplementary technique.

6. Non-specific signs of fetal abnormality
In this section attention is drawn to some non-specific signs that should alert to the possibility of a significant fetal abnormality.

Nuchal translucency in the first trimester
A unilocular cystic collection over the fetal posterior neck and upper back containing an estimated volume of fluid greater than 3 mm3 has an approximated 30% chance of having an abnormal karyotype. A multiloculated or multitrabeculated nuchal cystic hygroma with a midline septum has similar chance of abnormal karyotype but also a 20% chance of having another abnormality but normal karyotype.

Nuchal thickening disappears as the embryo matures into fetalhood, but remains as a soft sign in a small number of Down's syndrome babies, and progresses to full blown cystic hygroma in many cases of Turner's syndrome (46XO).

Choroid plexus cysts (CPC)
These are mentioned because there is still confusion and controversy regarding their presence. Undoubtedly, in late pregnancy persistence of CPC is strongly associated with trisomy 18. However, in early pregnancy CPCs are relatively common and unless a careful scan reveals another abnormality they are highly likely to be benign. Therefore, it is not necessary to karyotype all fetuses found to have CPC in early pregnancy, and one would consider it is also not necessary if additional abnormalities are found. However, it will still be very useful in borderline cases.

Single umbilical artery (SUA)
The umbilical cord is easy to visualize. There is a high association with fetal anomaly which may be found in between 15-48 % of all cases with SUA.

Amniotic fluid abnormalities: polyhydramnios and oligohydramnios
In the second and third trimesters of pregnancy amniotic fluid is predominantly fetal urine. The prerequisites for fetal urine production are good renal artery perfusion, functioning normally formed kidneys (or at least one kidney), and a competent urinary tract allowing normal expulsion of urine.

Production of amniotic fluid (fetal urine) has to be balanced by its reduction which requires competent fetal swallowing and propulsion through stomach and doudenum into small bowel where the fluid can be reabsorbed and homeostasis maintained via the placenta.

This balance can be upset by any abnormalities of the urinary tract, cardiovascular system or upper gastrointestinal tract, or any generalized abnormalities that make the fetus "sick" and indirectly affect the above, e.g. most causes of fetal hydrops and diabetes mellitus.

The presence of an adequate quantity of amniotic fluid is essential for fetal lung and musculoskeletal development. Therefore, the observation of abnormal quantities of amniotic fluid is an excellent non-specific sign that there may be underlying fetal abnormalities or fetal sickness. Abnormalities are present in between 30-60% of polyhydroamnios and about 40% of oligohydramios. However, these figures are somewhat meaningless as there has not been a rigid standard adhered to in classifying poly- and oligohydramnios.

Measurement of the amniotic fluid index (AFI) has emerged as a simple popular and reproducible technique of estimating amniotic fluid. The AFI is obtained by dividing the uterus into four quadrants, measuring the vertical deepest pocket in that quadrant and summing the four measurements. Currently values above 25 are taken to indicate the presence of polyhydramnios and values below 5 are taken to indicate oligohydramios.

One anticipates that with future universal use of technique we will be able to have a better measure of the relative risks of abnormality with a given amniotic fluid index at a specific gestational age.

Placenta praevia
Placenta praevia occurs in about 0.5 % of all pregnancies. US is already very accurate in determining placental site and the presence of placenta praevia and is therefore the technique of first choice.

The concept of placental migration and possible confusion caused by an overfilled or underfilled bladder are well known but there are about 5% pregnancies in which it may be difficult to establish placental location relative to the external os with confidence. These are: the obese mother, posterior location of placenta and intervening fetal head, vasa praevia, and the rare but potentially fatal placenta accreta and percreta.

In these situations further investigation prior to delivery are now available. The most easily available would be TVUS.

The probe is introduced gently with on-screen guidance to avoid impacting the cervix. The uterus is examined from positions in the vaginal fornices. With the addition of CFI it is easy to identify placenta covering the internal os, and marginal vessels covering the os. Obliteration of the placental myometrial line and high flow venous lakes within the placenta suggest placenta accreta.

Figure 5. Transverse scan of an early viable pregnancy with a large complex mass with thick septae behind the uterus.

Claims have also been made for the value of MRI in such situations, but as yet there has not been any study comparing the value of TVUS and CFI with MRI.

On cost and availability it would seem that US, TVUS and CFI is sufficient.

Discovery of parauterine mass during pregnancy
A parauterine mass may be discovered during pregnancy by palpation but more frequently it is revealed during careful routine antenatal ultrasonography (Fig. 5). "Careful" indicates that the pelvis has been scanned properly and that the routine study was not tunnel-visioned onto measuring the BPO and little else!

The differential diagnosis includes uterine fibromyoma, functional lute al cysts, endometrial chocolate cysts, cystic teratomas, and ovarian malignancy.

It is the last diagnosis which is most feared and for which radiologists will be recruited to test their powers of clairvoyance. The five year survival rate of stage I ovarian carcinoma is 60-90% but for stage IV 0-4 %, and therefore early diagnosis is crucial. Early diagnosis of ovarian cancer is usually incidental such as discovery in pregnancy during antenatal US or caesarean section.

The use of serum tumor markers, such as CA125, may be limited as these are also raised during pregnancy and with endometriosis.

Ovarian tumors may be found in 1/2000 to 1/600 pregnancies. The commonest tumor reported is benign cystic teratoma. Malignancy, fortunately, is relatively rare constituting up to 5 % of the ovarian masses found.

Surgical removal is recommended in all cases. It will not only remove possible malignancy but also lessen the chance of torsion (about 25 %) and haemorrhage. Supportive progestational therapy reduces the subsequent abortion rate form 85% to 10%.

Radiologists have used US, TVUS and MRI with success in demonstrating the presence and origin of the lesion although it is often difficult to distinguish an endometrial from an ovarian cyst. The diagnosis of malignant potential is more difficult.

CFI has been used successfully to identify tumor neovascularity which can then be interrogated with pulsed Doppler. Tumors show characteristic low impedence waveform. This holds true for the non-pregnant woman. Unfortunately, it is not clear whether this also applies in pregnancy as the characteristic of pelvic vascular flow during pregnancy also shows low impedence and it may be that any low impedence vascular changes discovered in an ovarian mass are purely the effect of pregnancy.

The conventional wisdom still holds that pure cysts are more likely to be benign and any solid components raise the risk level of possible malignancy and indicate need for surgery.

Deep vein thrombosis (DVT)
Suspicion of DVT can be easily allayed or confirmed by the use of US and CFI without need for radiological contrast venography. Worries regarding irradiation of the fetus are removed and pregnant women can be examined at a lower threshold of clinical suspicion.

Eclampsia
MRI has not only provided another non-ionising radiation type of imaging diagnosis, but has its own unique information to reveal. In obstetrics this has been very true in studies of eclampsia and preeclampsia. Pregnancy induced hypertension can result in headache, confusion, raised intracranial pressure and coma, visual loss and paralysis from cerebral hemorrhage with long term morbidity. Early diagnosis and appropriate treatment is essential in these previously healthy women.

Women with preeclampsia may show hyperintense signals on T2weighted images in the deep cerebral white matter. Those with eclampsia show increased signal at grey matter junctions, cortical oedema and haemorrhage especially in the region of the posterior cerebral circulation.

Current MRI is more sensitive than CT in demonstrating abnormalities in eclamptic women who have had seizures but still does not show abnormality in all studies. As newer more sensitive sequences become available one can anticipate that MR! will be more sensitive and may become a screening tool for mothers at risk. It may be that the "at risk" criteria will be altered as mothers without all the accepted signs such as oedema and proteinuria may still develop cerebral problems.

The severe vasospasm that accompanies eclampsia is demonstrable by magnetic resonance angiography. This information is similar to that obtained by more invasive modalities such as angiography or radionuclide studies using SPECT.

In addition to eclamptic changes, MRI has shown physiological changes of the pituitary during pregnancy consisting of hyperintensity on T1-weighted images, upward bowing and increase in size to 10 mm vertical height which should not be confused with pathology.

Pathological conditions of the pituitary such as pituitary haemorrhage and tumors have also been reported as here other intracranial pathologies: intracranial aneurysm, venous thrombosis and venous angiomas.

 

Con Metreweli