Low-field dedicated magnetic resonance imaging - Wiley Online Library

6 downloads 10 Views 111KB Size Report
Service d'Anatomie Pathologique, Hôpital Civil, Hôpitaux Universitaires de Strasbourg, Strasbourg ... obtained with a low-field (0.1 T) dedicated magnetic reson-.

Ultrasound Obstet Gynecol 1998;12:271–275

Low-field dedicated magnetic resonance imaging: a potential tool for assisting perinatal autopsy B. Langer, P. Choquet*, S. Ravier*, B. Gasser†, G. Schlaeder and A. Constantinesco* Service de Gynécologie-Obstétrique II, Hôpital de Hautepierre; *Laboratoire de Biomécanique, Hôpital de Hautepierre; †Service d’Anatomie Pathologique, Hôpital Civil, Hôpitaux Universitaires de Strasbourg, Strasbourg, France

Key words:



ABSTRACT earlier discovery of malformations and thus to autopsies of smaller fetuses which are always more difficult. To overcome these difficulties, postmortem magnetic resonance imaging (MRI) has been proposed for general4 and, more specifically, for fetal5–7 necropsy studies. However, in all these studies, MRI was conducted with wholebody MRI systems using the standard coils provided for head, wrist or knee examinations. MRI has already been used for imaging embryos at a microscopic level8, but the specifically adapted magnetic resonance research systems used are not widely available. We report the postmortem observation of two second-trimester fetuses, one normal and one abnormal, using a dedicated low-field MRI system9.

Although the practice of perinatal autopsy has increased in recent years, examination of the fetus and especially of the fetal brain during the first trimester or the beginning of the second trimester remains difficult. Postmortem highresolution images of the brain of a normal and an abnormal fetus of the same gestational age (22 weeks) were obtained with a low-field (0.1 T) dedicated magnetic resonance imaging (MRI) system. We demonstrated that a small MRI machine supplemented data from classical necropsy and may help in the interpretation of in utero ultrasound and magnetic resonance images for the antenatal diagnosis of fetal malformations.


During the past 15 years, perinatal autopsy has contributed to a better understanding of fetal and perinatal pathology1. Gross autopsy generally allows the diagnosis of major malformations, even when there is a high degree of maceration. Conventional radiography is classically used to obtain additional information. To improve postmortem diagnosis and confirm in utero findings, other imaging techniques have recently been proposed, for example, the use of magnifying glasses or dissecting microscopes2, angiographic ultrasound1 and endoscopy. Unfortunately, examination of the fetal brain by these techniques is problematic. Its soft structure, owing to its high water content, leads to early maceration and liquefaction which can render an adequate examination impossible2. Prolonged formalin preservation prior to dissection facilitates study of the brain, but even under such circumstances dissection remains difficult3. Moreover, progress in antenatal diagnosis had led to the

A dedicated (i.e. small field of view) low-field MRI system generating a vertical 0.1 T B0 field, previously developed in our laboratory9, was used (Figure 1). It is based on an open, small, water-cooled resistive magnet (Drusch, Poissy, France), with a homogeneous imaging zone (± 5 ppm) of about 10 × 10 cm in area and 6 cm in height and an MR 3030 imaging console (SMIS, Guilford, UK). The acquisition and reconstruction software as well as the MRI sequences were developed in our laboratory10. Owing to the perpendicular orientation of the main magnetic field with respect to the sample access, we developed solenoidal coils, ranging from 1.5 cm to 12.5 cm in diameter. They are known for their high quality factor and good B1 homogeneity11. For this study, we chose a coil with a 5.5-cm diameter, close to that of the sample, allowing a good filling factor. We obtained three-dimensional T1-weighted

Correspondence: Dr B. Langer, Service de Gynécologie-Obstétrique II, Hôpital de Hautepierre, Hôpitaux Universitaires de Strasbourg, 67098 Strasbourg CEDEX, France




AMA: First Proof

Received 25–11–97 Revised 5–8–98 Accepted 7–8–98

Postmortem low-field MRI

Langer et al.



Figure 1 Low-field magnetic resonance imaging system generating a vertical 0.1 T B0 field with a solenoid coil inside the magnet

images with a rapid gradient echo sequence (FLASH12) and three-dimensional T2-weighted images with a rapid spin echo train sequence (RARE12). The combination of small fields of view and small acquisition matrix sizes (square or rectangular) led to spatial resolution images of up to 0.5 × 0.5 mm per pixel with a slice thickness of 2 mm.

CASE REPORTS Case 1 A 28-year-old nulliparous woman was admitted to our department with premature rupture of the membranes at 22 weeks’ gestation. There was no indication of infection and an ultrasound scan at 20 weeks’ gestation had been normal. Given the pregnancy’s unfavorable prognosis, the parents opted for termination of the pregnancy. MRI was performed 1 day after expulsion, and 0.1 T MRI clearly allowed the observation on a sagittal view (Figure 2a) of the lateral ventricle with its choroid plexus, the thalamus, the cerebral peduncle and the pons with the fourth ventricle and the cervical vertebrae. The axial view (Figure 2b) showed the two cerebral hemispheres with the frontal and the inferior horns of the lateral ventricle, the cerebral peduncles and the cerebellum.

Figure 2 Normal fetal head at 22 weeks’ gestation. (a) Sagittal view. Acquisition parameters: FLASH 500/25/80° nex 4, St 2 mm. In plane resolution: 0.5 × 0.5 mm/pixel. Acquisition time (16 slices): 51 min 12 s. BM, brain mantle; LV, lateral ventricle; CP, choroid plexus; Th, thalamus; C, cerebellum; IV, fourth ventricle; CM, cisterna magna; P, pons; CV, medulla and the cervical vertebrae. (b) Axial view. Acquisition parameters: FLASH 500/9/80° nex 2, St 2 mm. In plane resolution: 0.5 × 0.5 mm/pixel. Acquisition time (16 slices): 25 min 36 s. Fv, frontal and Iv, inferior horns of the lateral ventricle; BM, brain mantle; Th, thalamus; C, cerebellum

272 Ultrasound in Obstetrics and Gynecology AMA: First Proof


Postmortem low-field MRI

Langer et al.

An autopsy performed 48 h after expulsion did not reveal any macroscopic malformations and this case was therefore considered normal.

Case 2 A 32-year-old woman, gravida 4, para 3, with an unremarkable medical history, was referred to our department at 21 weeks’ gestation, because of suspected holoprosencephaly. Ultrasound examination confirmed the diagnosis of alobar holoprosencephaly associated with polydactyly. Karyotyping showed trisomy 13. After extensive counselling, the parents requested termination of the pregnancy. MRI was performed 1 day after expulsion. The sagittal magnetic resonance image showed the proboscis above the orbit, the brain mantle (cup type) on the thalamus and the dorsal sac posteriorly (Figure 3a); the thalamus and the cerebellar hemispheres were also visible. The axial MRI view clearly showed the dorsal sac around the fused thalami and the fused orbit within one globe (Figure 3b). On the coronal MRI view, the fused orbits, the proboscis and the absence of a nose could all be visualized (Figure 3c and d). An autopsy was performed 72 h after expulsion. The female fetus (22 weeks) weighed 212 g and measured 14 cm. The autopsy findings were an alobar holoprosencephaly with a proboscis, behind a single orbit containing the two fused ocular globes, absence of the olfactory bulbs and a monoventricular brain cavity with fusion of the thalami (Figure 4).

DISCUSSION These two cases demonstrate the capability of low-field dedicated MRI to obtain high-quality pictures of the postmortem fetal brain. An autopsy provides vital information for documenting fetal malformations, determines the cause of death and allows correlation with prenatal diagnostic findings. The last point is of increasing importance, owing to the potential medicolegal consequences of the echographer’s decision. One problem is the lack of specially trained pathologists familiar with fetal specimens. Even for the specialist, the perinatal autopsy of the brain, especially in early pregnancy, remains difficult and less informative. It has already been shown that in utero MRI can improve the ultrasound diagnosis of fetal central nervous system anomalies such as holoprosencephaly13. There is no doubt that MRI has the potential to provide postmortem images of embryos and to confirm in vivo findings. Two studies using standard whole-body systems (at 1.5 T with standard radio-frequency coils) showed that there was a relatively good agreement between the necropsy examinations and MRI6,7. One study enhanced the anatomical structures (except the head) that could be recognized on magnetic resonance images5. In addition, extensive MRI work on mouse embryos has already been done, describing the different stages of fetal development8.

We believe that MRI cannot replace classical necropsy and that the ‘gold standard’ will remain gross autopsy and subsequent histological analysis of tissue sections14–17. However, it is likely that MRI will be useful for supplementing tissue information by means of the ‘proton stain’18. As already stated18, several advantages of MRI should be emphasized: its non-destructive nature leaves the specimen intact for further studies; multiple contrast capabilities at exactly the same location can be obtained by varying the pulse sequences and their characteristics; inherent three-dimensional properties show the relationships between organs; and, finally, multiple axial, coronal or sagittal slices and also multiple oblique or even curvilinear sections are possible. These aspects are of particular importance in cases of autopsy refusal by parents6. Moreover, postmortem MRI could be used to guide histopathologists to areas of interest for microscopic evaluation19. Multiple-plane sections without destruction of embryos allow one to choose the most informative section for autopsy. In this respect, Brookes and Hall-Craggs17 suggested the possibility of performing biopsies under magnetic resonance guidance. This procedure can be proposed as an alternative means of obtaining pathological information when an autopsy is refused. Our MRI system uses a low magnetic field in a widely open magnet: this represents a great advantage, since the size of the MRI needle artifact is directly dependent on the main magnetic field value20. The further development of MRI as a complementary examination to necropsy is, in our opinion, dependent on the use of special small machines similar to the one used in this study. Clinical whole-body machines already used in different studies4–7 are not intended for this purpose. First, they are heavily used for clinical investigations; second, their size is too great in relation to that of embryos. Two problems not tackled in previous studies are the availability and the cost of MRI: it may be difficult to find time slots, except at night, to carry out perinatal postmortem MRI examinations, and MRI is a costly procedure. Small dedicated systems for magnetic resonance imaging of the extremities in adults are becoming increasingly available for use in fetal imaging. The smaller space requirement and the lower purchase and functioning costs than for wholebody systems make it more likely that a small dedicated MRI system could be made accessible to the pathologist close to the necropsy room. The image acquisition time is relatively short. For instance, the total image acquisition time for the examination of the abnormal fetus (Case 2 – three-dimensional acquisitions in the three planes; 16 slices with T1- or T2-ponderation) was less than 40 min. This time is in the same range as that reported by other authors using higher magnetic fields4,7. For the normal fetus, the highest resolution was accompanied by the longest acquisition time. One should keep in mind that the signal-to-noise ratio (S/N) is dependent on the acquisition time8 t: S/N ∝ √t. In the case of postmortem studies, there is no justification for using the fastest acquisition sequences. Instead, the signal-to-noise ratio and contrasts could be improved by carefully choosing the parameters of the sequences.

Ultrasound in Obstetrics and Gynecology 273 97/205

AMA: First Proof

Postmortem low-field MRI


Langer et al.


d b

Figure 3 Alobar holoprosencephaly. Fetal head at 22 weeks’ gestation. (a) Sagittal view. Acquisition parameters: RARE 2000/400/90° nex 4, St 2 mm. In plane resolution: 0.7 × 0.7 mm/pixel. Acquisition time (16 slices): 12 min 48 s. BM, brain mantle (cup type); DS, dorsal sac; P, proboscis; Th, thalami; C, cerebellum. (b) Axial view. Acquisition parameters: RARE 2000/400/90° nex 4, St 2 mm. In plane resolution: 0.7 × 0.7 mm/pixel. Acquisition time (16 slices): 12 min 48 s. O, fused orbit within one globe; Th, fused thalami; DS, dorsal sac. (c) Coronal view with the two fused orbits and the proboscis. Acquisition parameters: RARE 2000/400/90° nex 4, St 3 mm. In plane resolution: 0.8 × 0.8 mm/pixel. Acquisition time (16 slices): 12 min 48 s. P, proboscis; O, fused orbit; L, lips. (d) Macroscopic view of the cyclop’s face (compare with Figure 3c)

274 Ultrasound in Obstetrics and Gynecology AMA: First Proof


Postmortem low-field MRI

Langer et al.


Figure 4 Alobar holoprosencephaly. Macroscopic view of dorsal sac (Case 2)

Another point that we should stress is the need for a standardized set-up procedure for low-field dedicated MRI which could be performed everywhere and give similar contrast results. The studies already published have shown that a correlation should be made between abnormal fetuses and autopsy findings; moreover, as for other medical applications of MRI21, correlations should be made between normal dead fetuses and magnetic resonance images, in order to depict the appearance of normality clearly before attempting to solve the problem of abnormalities seen on magnetic resonance images. Normal fetus MRI references are needed for perinatal necropsy and for the use of MRI as a diagnostic procedure during pregnancy. In conclusion, MRI of the fetus remains a technique for obtaining complementary information and cannot replace autopsy. In this study, we have shown that a small lowfield dedicated magnetic resonance device with carefully designed coils could be used for postmortem imaging of fetuses, without drawbacks in terms of resolution, or contrast between different tissues.

1. Chambers HM. The perinatal autopsy: a contemporary approach. Pathology 1992;24:45–55 2. Curry CJR, Honoré LH. A protocol for the investigation of pregnancy loss. Clin Perinatol 1990;17:723–42 3. Kier EL, Fulbright RK, Bronen RA. Limbic lobe embryology and anatomy: dissection and MR of the medial surface of the fetal cerebral hemisphere. Am J Neuroradiol 1995;16: 1847–53 4. Ros PR, Li KC, Vo P, Baer H, Staab EV. Preautopsy magnetic resonance imaging: initial experience. Magn Reson Imaging 1990;8:303–8 5. Roberts MD, Lange RC, McCarthy SM. Fetal anatomy with magnetic resonance imaging. Magn Reson Imaging 1995;13: 645–9 6. Brookes JAS, Hall-Craggs MA, Sams VR, Lees WR. Noninvasive perinatal necropsy by magnetic resonance imaging. Lancet 1996;348:1139–41 7. Woodward PJ, Sohaey R, Harris DP, Jackson GM, Klatt EC, Alexander AL, Kennedy A. Postmortem fetal MR imaging: comparison with findings at autopsy. Am J Roentgenol 1997; 168:41–6 8. Smith BR, Linney E, Huff DS, Johnson GA. Magnetic resonance microscopy of embryos. Comput Med Imaging Graph 1996;20:483–90 9. Gries P, Constantinesco A, Brunot B, Facello A. MR imaging of hand and wrist with a dedicated 0.1-T low-field imaging system. Magn Reson Imaging 1991;9:949–53 10. Arbogast-Ravier S, Xu F, Choquet P, Brunot B, Constantinesco A. Dedicated low-field MRI: a promising lowcost technique. Med Biol Eng Comput 1995;33:735–9 11. Hoult DI, Richards RE. The signal-to-noise ratio of the nuclear magnetic resonance experiment. J Magn Reson 1976; 24:71–85 12. Parikh AM. Magnetic Resonance Imaging Techniques. New York: Elsevier, 1991;76–86, 224–33 13. Resta M, Greco P, D’Addario V, Florio C, Dardes N, Caruso G, Spagnolo P, Clemente R, Vimercati A, Selvaggi L. Magnetic resonance imaging in pregnancy: study of fetal cerebral malformations. Ultrasound Obstet Gynecol 1994;4:7–20 14. Berry PJ, Keeling JW, Wigglesworth JS. Perinatal necropsy by magnetic resonance imaging [Letter]. Lancet 1997;349:55 15. Rushton DI. Perinatal necropsy by magnetic resonance imaging [Letter]. Lancet 1997;349:56 16. Niermeijer MF. Perinatal necropsy by magnetic resonance imaging [Letter]. Lancet 1997;349:56 17. Brookes JS, Hall-Craggs MA. Postmortem perinatal examination: the role of magnetic resonance imaging [Editorial]. Ultrasound Obstet Gynecol 1997;9:145–7 18. Johnson GA, Benveniste H, Black RD, Hedlund LW, Maronpot RR, Smith BR. Histology by magnetic resonance microscopy [Review]. Magn Reson Q 1993;9:1–30 19. Lamont P, Sachinwalla T, Pamphlett R. Magnetic resonance imaging of postmortem infant brains. J Child Neurol 1994; 9:59–62 20. Arbogast-Ravier S, Gangi A, Choquet P, Brunot B, Constantinesco A. An in vitro study at low field for MR guidance of a biopsy needle. Magn Reson Imaging 1995;13: 321–4 21. Drapé JL, Constantinesco A, Arbogast S, Sick H, WolframGabel R, Brunot B. High resolution MRI of the normal finger at 0.1 T: anatomic correlations. Surg Radiol Anat 1992;14: 349–60

ACKNOWLEDGEMENT The authors thank Gérard Vetter for his technical assistance.

Ultrasound in Obstetrics and Gynecology 275 97/205

AMA: First Proof

Suggest Documents