Available via license: CC BY-NC
Content may be subject to copyright.
Received: 13 January 2021
-
Revised: 18 May 2021
-
Accepted: 30 May 2021
DOI: 10.1002/pd.5987
ORIGINAL ARTICLE
Myelomeningocele–Chiari II malformation–Neurological
predictability based on fetal and postnatal magnetic
resonance imaging
Farjad Khalaveh
1
|Rainer Seidl
3
|Thomas Czech
1
|Andrea Reinprecht
1
|
Gerlinde Maria Gruber
6
|Angelika Berger
3,4
|Herbert Kiss
5
|
Daniela Prayer
2
|Gregor Kasprian
2
1
Department of Neurosurgery, Medical
University of Vienna, Vienna, Austria
2
Department of Radiology, Division of Neuro‐
and Musculoskeletal Radiology, Medical
University of Vienna, Vienna, Austria
3
Department of Pediatrics and Adolescent
Medicine, Medical University Vienna, Vienna,
Austria
4
Department of Pediatrics and Adolescent
Medicine, Comprehensive Center for
Pediatrics, Division of Neonatology, Pediatric
Intensive Care & Neuropediatrics, Medical
University of Vienna, Vienna, Austria
5
Department of Obstetrics and Gynecology,
Division of Obstetrics and Feto‐maternal
Medicine, Medical University of Vienna,
Vienna, Austria
6
Division of Anatomy and Developmental
Biology, Department of Anatomy und
Biomechanics, Karl Landsteiner University of
Health Sciences, Krems an der Donau, Austria
Correspondence
Gregor Kasprian, Associate Professor of
Radiology Department of Radiology Division
of Neuro‐and Musculoskeletal Radiology,
Medical University of Vienna Waehringer
Guertel 18‐20 1090 Vienna, Austria.
Email: gregor.kasprian@meduniwien.ac.at
Abstract
Objective: This systematic comparison between pre‐and postnatal imaging findings
and postnatal motor outcome assesses the reliability of MRI accuracy in the prog-
nostication of the future long‐term (mean, 11.4 years) ambulatory status in a his-
toric group of postnatally repaired myelomeningocele (MMC) cases.
Methods: A retrospective, single‐center study of 34 postnatally repaired MMC
patients was performed. We used fetal and postnatal magnetic resonance imaging
(MRI) to compare the fetal and postnatal radiological lesion level to each other and
to the postnatal ambulatory level as a standard of reference and analyzed Chiari II
malformation characteristics.
Results: In 13/15 (87%) and 29/31 (94%) cases, the functional level was equal to or
better than the prenatal and postnatal radiological lesion level. A radiological lesion
level agreement within two segments could be achieved in 13/15 (87%) patients. A
worse than expected functional level occurred in cases with Myelocele (2/3 pa-
tients), coexistent crowding of the posterior fossa (2/3 patients) and/or abnormal
white matter architecture, represented by callosal dysgenesis (1/3 patients). In all
patients (2/2) with a radiological disagreement of more than two segments, seg-
mentation disorders and scoliosis were observed.
Conclusion: Fetal and postnatal MRI are predictive of the long‐term ambulatory
status in postnatally repaired MMC patients.
Key points
What's already known about this topic?
�Fetal and postnatal magnetic resonance imaging (MRI) show a good correlation in identi-
fying the level of the myelomeningocele (MMC) lesion.
�Prenatal ultrasound (US) and fetal MRI show a comparable agreement, within two seg-
ments, in predicting the short‐term ambulatory status.
This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited and is not used for commercial purposes.
© 2021 The Authors. Prenatal Diagnosis published by John Wiley & Sons Ltd.
922
-
Prenatal Diagnosis. 2021;41:922–932. wileyonlinelibrary.com/journal/pd
What does this study add?
�Fetal and postnatal MRI have a good predictive value for the long‐term ambulatory status in
patients with postnatal MMC closure.
�MMC patients with worse than expected ambulatory status showed signs of vermian
displacement and corpus callosum dysgenesis.
�In MMC patients with spinal segmentation disorders and scoliosis, there was a major
disagreement between fetal and postnatal MRI, specifically with regard to the MMC lesion
level.
1
|
INTRODUCTION
The incidence of myelomeningocele (MMC) has decreased in the past
several years due to an increase in folate acid supplementation.
1,2
Nevertheless, with a prevalence rate between 0.2 and eight per 1000
live births, MMC is the most common form of spinal dysraphism and
is associated with Chiari II malformation.
3–10
For an appropriate treatment plan, an early prenatal diagnosis is
of the utmost importance.
11
As a result of the improvement in pre-
natal ultrasound (US) diagnostics, 88% of neural tube defects (NTD)
in Europe are diagnosed during a median gestational age of
17 weeks.
12
Consequently, the abortion rate in cases of NTD has
increased in most European countries, especially of MMC patients
who are not eligible for prenatal closure.
12–16
These profound
changes in evaluation and treatment measures have led to a decrease
of postnatal MMC repairs. The legitimate concern of a poor ambu-
latory status, in particular, leads to substantial parental uncertainty
with regard to the option against surgical treatment and for the
termination of pregnancy. A better definition of the imaging‐based
predictability of future motor outcome in MMC cases would help
to reduce this uncertainty and strengthen parents’ confidence in this
difficult decision. More recent studies have focused on the functional
prediction value of prenatal sonography in postnatally repaired MMC
patients. There is a wide variation in studies about the correlation
between ambulatory status and radiological lesion level assigned by
prenatal sonographic findings, with the neuromotor level reportedly
equal to or better than the anatomic level in 52.9%‐91%.
17–23
Many
centers have implemented MRI as an complementary prenatal ex-
amination method in cases of a suspected NTD. Depending on the
expertise and quality of fetal MRI and US, there is much variation
with regard to the superiority of either technique.
17,24
Fetal MRI
allows a complete examination of the neuroaxis by the second and
third trimester, regardless of technique‐related limitations in trans-
abdominal ultrasonography.
25–27
When fetal sonography is compared
to postnatal MRI for the identification of the MMC lesion level, a
good correlation within one level could be observed in 82% of
cases.
28
By predicting the ambulatory status, prenatal US and fetal
MRI showed an equal agreement, within two segments, in 79.4% of
cases.
29
However, the studies that analyzed the predictive value
of fetal and postnatal MRI with regard to the ambulatory status of
patients with postnatally repaired MMC lesions were limited by small
patient groups or a short‐term follow‐up usually assessed within one
week of postnatal MMC closure.
17,29,30
Furthermore, other Chiari II
malformation characteristics, such as corpus callosum (CC) de-
formities or vermian displacement, were not considered as additional
factors that could potentially explain a poorer functional outcome
than that based on the prediction according to the anatomical level
alone.
31,32
In this single‐center, follow‐up study, we aimed to retrospec-
tively assess fetal and postnatal MR examinations and correlate
these to the ambulatory status of postnatally repaired MMC patients.
The main goal of our study was to determine the anatomically pre-
dictive and functionally prognostic value of fetal and postnatal MRI in
postnatally repaired MMC patients, with respect to the long‐term
(mean follow‐up period of 11.4 years) ambulatory status as the
standard of reference. By examining the entire neuroaxis using fetal,
as well as postnatal MRI, we aimed to identify and neuro-
radiologically characterize factors that negatively affect the long‐
term motor outcome of postnatally repaired MMC patients. This
descriptive reference data, derived from a “historic” cohort may
serve as an important benchmark for the comparison of long‐term
outcomes in prenatal surgery trials.
2
|
MATERIALS AND METHODS
The study protocol was approved by the Ethics Committee of the
Medical University of Vienna (EK 1318/2019).
2.1
|
Patients
Patients with open spina bifida who underwent a postnatal MMC
closure at the Neurosurgery Department of the Medical University of
Vienna between the years 2000 and 2015 were included in this
study.
Patients were referred to fetal MRI in cases where an MMC was
suspected by prenatal US examination. Postnatal MRI examinations
were performed only in cases of ambulatory deterioration due to
suspected tethered cord or CSF shunt malfunction. Fetal MR images
were obtained in 16 patients (47%) and 32 patients (94%) underwent
postnatal MRI. Fetal and postnatal MR images were evaluated, in
consensus, by one neurosurgeon (T.C.) and one neuroradiologist
(G.K.), both with more than 10 years‘ experience in pediatric
KHALAVEH ET AL.
-
923
neuroradiology and fetal MRI. The level of the MMC lesion was
defined as the segmental level of the fetal neural placode and the
postnatal tethered spinal cord. We defined the most cranial part of
the lesion level as the border to the seemingly normal myelon
segment, and also used the last completely intact vertebral body as a
landmark (Figure 1). A myelon segment was defined as normal if a T2‐
hyperintense CSF signal was visible around the myelon. A vertebral
body was defined as completely intact only in cases of a complete
posterior arch and spinous process. The vertebral bodies were
counted beginning caudally or cranially, depending on the level of
lesion. In lumbar lesions, vertebral bodies were counted beginning
caudally using the promontory as landmark. In thoracic lesions,
vertebral bodies were counted beginning cranially, using the foramen
magnum as landmark. The method for counting the vertebral bodies
did not differ intraindividually between fetal and postnatal MRI. The
caudal part of the neural placode or tethered spinal cord was not
described. Furthermore, we evaluated Chiari II malformation char-
acteristics, particularly vermian displacement and the shape of the
CC. As previously emphasized by Harvey Sarnat,
33
herniation relates
to an active displacement of an organ through a local tissue defect
and since it is not clear if the vermis primarily develops in the spinal
canal or is secondarily displaced, we used the term vermian
displacement.
34
For assessing vermian displacement the foramen
magnum was used as reference as well.
Due to the great heterogeneity in the nomenclature and the
definition of callosal anomalies,
35
we decided to use a modified clas-
sification system based on Edwards et al.
36
and classified the
morphology of the CC into four classes (Figure 2). Some showed an
inconspicuous form and were thus defined as (1) “normal” (Figure 3).
Due to an MMC‐associated hydrocephalus, some CC showed a thinned
form and were specifically grouped as (2) “hydrocephalic” (Figure 4)
rather than hypoplastic, as originally described by Edwards et al.
36
The
CC was defined as (3) “dysgenetic” in cases of “small” abnormally
shaped or missing areas, especially in the region of the splenium and
rostrum. A combination of group (2) and (3) was observed as well, and
thus, were grouped as (4) “dysgenic +hydrocephalic”.
FIGURE 1 Segmental level of the neural
placode. Patient with a Myelomeningocele and
lesion level at L1 and fetal vermian
displacement at C3. Functional level at L3‐L5.
Left: EPI/T2* sequence. Right: T2‐weighted
sequence
FIGURE 2 Corpus callosum (CC) morphology. A,B, Normal form of the CC. C‐F, Hydrocephalic configuration of the CC. G, H, Dysgenetic
CC: the splenium of the CC is absent and the proportions of the different callosal segments abnormal
924
-
KHALAVEH
ET AL.
Thirty infants (88%) were delivered via Caesarean section and
four (12%) via vaginal birth between 31 +3 and 40 +1 week of
estimated gestational age, respectively (Table 1).
MMC repair was performed within 24 h in 31 (91%) and 48 h in
three (9%) patients.
Follow‐up was conducted at the University Hospital of Vienna,
an MMC‐specialized center, by a multidisciplinary team of pediatric
neurosurgeons, neuroradiologists, neuropediatricians, orthopedists
and urologists, with a mean follow‐up of 11.4 years (range, 3‐
18 years). Five patients were between three to five years of age at
the last follow‐up. Twenty‐nine patients were older than seven
years.
2.2
|
Functional level
The functional level of all 34 MMC patients was assessed by a neu-
ropediatrician based on patients' best ambulatory status during
follow‐up and was classified into three groups (Table 2).
Those who were incapable of knee extension and/or hip flexion,
and thus, were in need of a wheelchair, were assigned to
group ≤L2.
37
In cases of knee flexion and/or dorsal foot extension
deficits, patients were assigned to group L3‐L5. These patients were
ambulatory only with the help of various devices, such as a walking
FIGURE 3 Fetal magnetic resonance (MR) images at 37 gestational weeks of a Myelomeningocele patient depicting the normal shape of
the corpus callosum (CC). Left: Axial MR image shows a normal configuration of the CC at the genu (arrow 1) and splenium (arrow 2). Middle:
Sagittal MR image shows a normal configuration of the genu (arrowhead 1), body (arrowhead 2) and isthmus (arrowhead 3) of the CC. Right:
Coronal MR image shows a normal configuration of the body of the CC
FIGURE 4 Fetal magnetic resonance (MR) images at 33 gestational weeks of a Myelomeningocele patient with a “hydrocephalic”
configuration of the corpus callosum (CC) and functional lesion level of L3‐L5. Left: Axial MR image shows a thinned genu (arrow) and splenium
of the CC. Middle: Sagittal MR image shows a thinned genu (arrowhead 1), body (arrowhead 2) and posterior segment (arrowhead 3) of the CC.
Right: Coronal MR image shows a thinned body of the CC
TABLE 1Characteristics of the study patients
Variable Value
Total no. of patients 34 (100%)
Sex
M 13 (38.2%)
F 21 (61.8%)
Type of delivery
Vaginal 4 (11.8%)
C‐section 30 (88.2%)
Gestational age at birth (in weeks) 37.7 (31.4–40.1)
Time of MMC closure
≤24 h 31 (91.2%)
≤48 h 3 (8.8%)
Gestational age at fetal MRI (in weeks) 27.8 (19.4–37.0)
Gestational age at 2
nd
fetal MRI (in weeks) 29.8 (22.4–35.4)
Age at 1
st
postnatal MRI (in d) 75 (0–4087)
CSF Shunt 27 (79.4%)
Note: Values are expressed as numbers (%) or as median (range).
Abbreviations: CSF, cerebral spinal fluid; MMC, myelomeningocele;
MRI, magnetic resonance imaging.
KHALAVEH ET AL.
-
925
frame, crutches, or an ankle/knee/hip orthosis. Patients with a deficit
in plantar flexion and/or who were able to walk without any sup-
porting device were assigned to group ≥S1.
38
In cases of an asym-
metric functional neurological level, the worst functional level was
used.
2.3
|
Fetal MRI
Fetal MRI examination was performed at the Department of
Biomedical imaging and Image‐guided Therapy at the Medical Uni-
versity of Vienna, without any sedation.
The fetal MRI protocol consisted of T2‐TSE sequences in three
orthogonal planes (TE =100–140 ms, TR =variable, slice thick-
ness =3–4.4 mm, FOV =230–290 mm), echoplanar sequences, and
SSFPE/FIESTA/true‐FISP‐sequences (FOV =260–300 mm, over
continuous slices with 3 mm overlays).
In 16/34 MMC patients (47%), the fetal MRI examination was
performed after transabdominal US showed a developmental anom-
aly. In 6/16 patients (38%), a second fetal MRI examination was
performed. The first and second fetal MRI examinations were per-
formed at a median estimated gestational age of 27 +6 weeks (range,
19 +3–37 weeks) and 29 +6 weeks (range, 22 +3–35 +3 weeks),
respectively.
2.4
|
Postnatal MRI
Postnatal MRI examinations were performed at the patients' local
diagnostic center or at the Department of Biomedical imaging and
Image‐guided Therapy at the Medical University of Vienna.
The postnatal MRI protocol consisted of T2‐TSE‐sequences in
three orthogonal planes (TE =50–140 ms, TR =variable, slice
thickness =2–6.5 mm, FOV =180–1024 mm), T1‐sequences
(TE =2.1–23 ms, TR =20–1800 ms, slice thickness =0.9–6.5 mm,
FOV =128–1024 mm), and CISS sequences (TE =2.4–4.4 ms,
TR =5.3–8.9 ms, slice thickness =0.6–1 mm, FOV =128–
512 mm).
In 32/34 MMC patients (94%), a postnatal MRI examination was
performed after closure of the MMC lesion. The first postnatal MRI
examination was done within the first day and 11.3 years (median,
75 days) after delivery.
One spinal, one cranial, and two cranial +spinal MRI exami-
nations were not performed postnatally, since functional
neurological development either improved or did not deteriorate
during clinical follow‐up. According to the surgical records, the
MMC lesions were in the lumbosacral area; one patient had a
functional level of L3‐L5, and two patients were classified into the
≥S1 functional group.
2.5
|
Statistical analysis
Data were analyzed using IBM SPSS Statistics for Windows version
26.0 (IBM, Armonk NY). Data are described as absolute frequencies
and percentages. Due to the descriptive character of the study, no
statistical analyses were performed.
3
|
Results
All of the 34 postnatally repaired MMC patients were assigned to
one of the predefined functional level groups. Fetal MR images
were available in 16 patients (47%). However, the image quality
was limited in one case. Thus, fetal MR images of 15 patients
were diagnostically used. Based on fetal MR images, MMC was
diagnosed in nine of these patients and Myelocele in 5/15 pa-
tients. Myelocystocele was diagnosed in one patient who showed
mild Chiari II malformation characteristics with no vermian
displacement. However, according to the surgical report a neural
placode was visible at birth and thus the lesion was defined as
MMC.
Postnatal MRI examinations were done in 32/34 patients (94%).
The postnatal segmental level of the tethered spinal cord was
assigned to 31/34 patients (91%) since, in one patient, only a cranial
MRI examination was performed. Of 34 patients, 27 (79%) received a
CSF shunt.
There was one death at the age of 10 years, with pneumonia as
the cause of death.
3.1
|
MMC lesion level
There were 8/34 (24%), 19/34 (56%), and 7/34 (21%) patients who
had a functional level of ≤L2, L3‐L5, and ≥S1, respectively (Table 3).
The functional level, as the reference standard, was compared with
fetal and postnatal MR images. For better comparison, the radio-
logical lesion level was organized into groups as well.
TABLE 2Functional level assessed by clinical neurological
examination
Functional lesion level Mobility
Above or equal to L2 Wheelchair
L3‐L5 Supporting devices (walking frame,
crutches, ankle/knee/hip orthosis)
Below or equal to S1 Unassisted
TABLE 3Functional lesion level of study patients and rate of
shunted patients in relation to their functional level group
Functional level Patients (n=34) Shunted patients (n=27)
Above or equal to L2 8 (23.5%) 7/8 (87.5%)
L3‐L5 19 (55.9%) 15/19 (78.9%)
Below or equal to S1 7 (20.6%) 5/7 (71.4%)
Note: Values are expressed as numbers (%).
926
-
KHALAVEH
ET AL.
3.1.1
|
Functional lesion level versus prenatal
radiological lesion level
In 13/15 (87%) cases, the functional level was equal to or better than
the prenatal lesion level as determined by fetal MRI. There were 4/15
(26.7%) patients who showed a better than expected ambulatory
status. In two cases, the functional level was worse than expected
and both cases were diagnosed with a Myelocele. In one case (case
A), a prenatal genetic examination was performed, which showed a
4q‐17p translocation. This patient (case A) had an anatomical pre-
natal lesion level of L3‐L5 with a functional level of ≤L2. The second
patient (case B) was diagnosed with a lesion level of ≥S1 and had a
functional level of L3‐L5.
The fetal MR images of case A showed the last intact vertebral
body and a normal myelon segment at L2 and the neural placode at
L3. In the postnatal MR images at the age of 22 days, the lesion level
was diagnosed at L4 and the normal myelon segment at L3. Prenatal
vermian displacement was diagnosed at C3 and postnatally at C2.
The CC showed a dysgenetic form on fetal and postnatal MR images.
On the 12
th
day of life, the patient received a cerebral spinal fluid
(CSF) shunt, and five months after birth, a craniocervical decom-
pression was performed.
In case B (Figure 5), the fetal MR images showed the last intact
vertebral body and normal myelon segment at L5 and the neural
placode at S1. In the postnatal MR images at the age of four months,
the lesion level was diagnosed at L4/5. Prenatal vermian displace-
ment was diagnosed at C3/4 and postnatally at C0. The CC had a
hydrocephalic form on fetal and postnatal MR images. Furthermore,
one singular periventricular heterotopia was diagnosed on postnatal
MRI examination, which could not be detected by fetal MRI at a
gestational age of 22 +3 weeks. The patient received a CSF shunt on
the 18
th
day of life. On fetal and postnatal MR images, sacral seg-
mentation defects and scoliosis could be observed. During follow‐up,
two untetherings were performed due to neurological deterioration
with no signs of CSF shunt malfunction. After each operation at the
age of 1.5 and 4.8 years, the neurological function improved, and the
patient was ambulatory with the help of a supporting device.
3.1.2
|
Functional lesion level versus postnatal
radiological lesion level
In 29/31 (94%) cases, the functional level was equal to or better than
the postnatal radiological lesion level. There were 7/31 (22.6%) pa-
tients who showed a better than expected ambulatory status. The
functional level was worse than expected in only two cases. Both
patients with a postnatal radiological lesion level of L3‐L5 had a
functional level of ≤L2.
Case A was already mentioned above. In the second case (case C)
(Figure 6), the postnatal MR images showed the last intact vertebral
body at L3 and a tethering of the spinal cord.
On the eighth day of life, the patient received a CSF shunt
and, on the first postnatal MR images, a normal CC (circumscribed
secondary defect to shunt placement at the genu/body transition ‐
Figure 6) and no vermian displacement with bulging of the atlan-
tooccipital membrane could be observed.
After deterioration in hip flexion at the age of six months, MRI
showed tethering of the spinal cord at L3/4 with a syringomyelia at C5.
Based on the clinical deterioration with the concordant neuroradio-
logical results, an untethering was performed. Eight months later, a
neurological deterioration of the upper extremities and a progressive
dysphagia led to a craniocervical decompression at the age of 1.2 years.
3.1.3
|
Radiological lesion level comparison
When the fetal and postnatal radiological lesion level were
compared, an exact agreement could be achieved in 6/15 (40%)
FIGURE 5 Fetal and postnatal magnetic resonance (MR) images and a radiograph of a myelocele patient with a functional lesion level of
L3‐L5. Left: Fetal MR image at an estimated gestational age of 19 +3 weeks indicates the position of the neural placode at level S1. Center:
Four months after birth, the postnatal MR image shows the defect distally to the vertebral bodies L5 and tethering of the spinal cord. Right:
The radiograph at the age of four years shows scoliosis and segmentation defects of the lumbar spine
KHALAVEH ET AL.
-
927
patients. An agreement within two segments could be observed in
13/15 (87%) patients. In one case (case D), the prenatal radiological
lesion level was evaluated to be five segments more caudal. In
another case (case E), the prenatal radiological lesion level was
assumed to be three segments more cranial. Both patients had
distinctive thoracolumbar segmentation defects, as well as scoliosis,
and were classified into the functional group of ≤L2.
3.2
|
Chiari‐II‐malformation characteristics
3.2.1
|
Vermis
A vermian displacement between C1 and C6 could be observed by
fetal MRI in 14/15 (93%) patients (Table 4and Figure 7). All these
patients received a CSF shunt implantation. All patients with
Myelocele had a vermian displacement on fetal MRI. In one patient
with no vermian displacement and no CSF‐shunt implantation, a
Myelocystocele was diagnosed by fetal MRI. In 7/15 (46.7%) pa-
tients, the vermis prenatally appeared at the same cervical spinal
level as on postnatal MR images. In 5/15 (33.3%) patients, the
vermis “ascended” postnatally by one to four segments. In all of
those five patients, MMC closure was performed within 24 h after
birth (Table 4).
Of the 4/15 (26.7%) patients with a better than expected
ambulatory status determined by fetal MRI, 2/4 (50%) had a vermian
displacement at C3 and 1/4 (25%) at C6.
Of the 7/31 (22.6%) patients with a better than expected ambu-
latory status determined by postnatal MRI, vermian displacement at
C1 (1/7, 14%), C3 (1/7, 14%), C4 (3/7, 43%), and C6 (1/7, 14%) could be
FIGURE 6 Postnatal magnetic resonance
images of an myelomeningocele patient with a
functional lesion level of ≤L2. Left: T2‐images
three months after birth show the last intact
vertebral body at L3 and tethering of the spinal
cord. Right: At the age of 14 months, cranial T2‐
images indicate crowding of the hypoplastic
posterior fossa and obliterated subarachnoid
spaces at the craniocervical junction
TABLE 4Chiari II malformation characteristics on fetal and postnatal MR images
Chiari II Characteristics Fetal MRI (n=15) Postnatal Cranial MRI (n=31) Shunted Patients with Postnatal MRI Scans (n=27)
Vermian displacement
C0 1 (6.7%) 9 (29%) 4 (44.4%)
C1 1 (6.7%) 2 (6.5%) 2 (100%)
C2 2 (13.3%) 3 (9.7%) 2 (66.7%)
C3 6 (40%) 6 (19.4%) 6 (100%)
C4 3 (20%) 8 (25.8%) 8 (100%)
C5 0 2 (6.5%) 2 (100%)
C6 2 (13.3%) 1 (3.2%) 1 (100%)
C7 0 0 0
Corpus callosum
Normal 2 (13.3%) 8 (25.8%) 6 (75%)
Dysgenetic 4 (26.7%) 8 (25.8%) 7 (87.5%)
Hydrocephalic 9 (60%) 12 (38.7%) 10 (83.3%)
Hydrocephalic +dysgenetic 0 3 (9.7%) 2 (66.7%)
Note: Values are expressed as numbers (%) or as median (range). All fetal MRI patients needed a CSF shunt. In two patients with a CSF shunt, postnatal
MRI scans could not be obtained.
Abbreviation: MRI, magnetic resonance imaging.
928
-
KHALAVEH
ET AL.
observed. All of them received a CSF shunt. One patient (case F) with
no vermian displacement and a prenatally normal and postnatally hy-
drocephalic CC did not receive a CSF shunt.
3.2.2
|
Corpus callosum
On fetal MR images, the CC appeared normal, hydrocephalic, and
dysgenetic in 2/15 (13%), 9/15 (60%), and 4/15 (27%) patients,
respectively. In only one case did a prenatal hydrocephalic CC appear
dysgenetic on postnatal MR images (Table 4). In one prenatally
normal (case F) and one dysgenetic CC, an additional hydrocephalic
feature could be observed postnatally. One of these patients (case F)
was diagnosed with a Myelocystocele by fetal MRI. Of the 4/15
(26.7%) patients with a better than expected ambulatory status
determined by fetal MRI, a normal (1/4, 25%), dysgenetic (1/4, 25%),
and hydrocephalic (2/4, 50%) configuration of the CC could be
observed. Of the 7/31 (22.6%) patients with a better than expected
ambulatory status determined by postnatal MRI, a normal (2/7, 29%),
dysgenetic (1/7, 14%), hydrocephalic (2/7, 29%), and dysge-
netic +hydrocephalic (2/7, 29%) configuration of the CC could be
observed.
4
|
DISCUSSION
In this single‐center study, we retrospectively evaluated 34 post-
natally repaired MMC patients with a mean follow‐up of 11.4 years
(range, 3–18 years).
The neuroradiological lesion level showed a good correlation
between fetal and postnatal MRI within two segments. Lesion level
discrepancies of more than two segments occured in cases with
scoliosis and segmentation disorders (Figure 8). With regard to the
long‐term ambulatory status of postnatally repaired MMC patients,
fetal and postnatal MRI showed a good predictability. Patients with a
worse than predicted ambulatory status showed CC dysgenesis and
significant vermian displacement.
Similar results have been shown by other studies, where the pre-
dictability of the ambulatory status was examined using different
prenatal neuroimaging examinations, such as US or fetal MRI, with
good correlation between the prenatal lesion level and the ambulatory
status of MMC patients.
18–20,25,29
The most recent study was con-
ducted by Sherrod et al., in which a comparison between fetal MR im-
ages and functional neurological level showed a correlation within one
level in 79.4%.
29
This study was limited by the short follow‐up, with a
functional assessment within one week of postnatal MMC repair. The
duration of follow‐up and the motor level classification were in contrast
to our study, as each patient was assigned an exact motor level. We
believe that, by consulting parents about the long‐term ambulatory
status, it is more plausible to classify the ambulatory status of MMC
patients taking into consideration the ambulatory helping device, for
example, wheelchair, crutches, or without aid. However, disregarding
the differences between our study and the study by Sherrod et al., we
could confirm their results, showing that the ambulatory status was
equal to or better than predicted in 29/31 (94%) and in 13/15 (87%)
cases by postnatal and fetal MRI, respectively.
With regard to the three patients with a worse than predicted
ambulatory status, all received a CSF shunt due to hydrocephalus
and showed signs of vermian displacement, CC dysgenesis, massive
thoracolumbar scoliosis, as well as segmentation disorders. As re-
ported by Kollias et al.,
19
this complete spectrum of abnormalities
may explain the inadequate predictions of ambulatory status. Ver-
mian displacement, in particular, may have a certain influence on
the ambulatory status of patients with MMC and especially Mye-
locele.
39
Posterior fossa crowding, which leads to parenchymal
compression and to primary motor neuron lesions, results in a
worse than expected ambulatory status.
40
Furthermore, according
to Nagaraj et al., patients with Myelocele are associated with severe
Chiari II malformations.
39
This was also supported by our results,
which showed a vermian displacement in all of our Myelocele
FIGURE 7 Left: Fetal magnetic resonance (MR) image of a Myelocele patient at an estimated gestational age of 27 weeks showing a
vermian displacement (arrowhead) at C3. Right: Fetal MR image of a Myelocele patient at an estimated gestational age of 22 weeks showing
vermian displacement (arrowhead) at C4. The exact level of displacement was determined by counting the T2‐weighted hyperintense
intervertebral discs (arrow)
KHALAVEH ET AL.
-
929
patients, of which two had a worse than predicted ambulatory
status. All of our patients with a vermian displacement diagnosed
by fetal MRI received a CSF shunt and were ambulatory with a
supporting device (≤L2 or L3‐L5).
The only patient (case F) with no vermian displacement on fetal
MR images did not receive a CSF shunt and was ambulatory without
a supporting device (≥S1), even though the pre‐and postnatal
radiological lesion level was diagnosed at L3‐L5. In this patient, a
Myelocystocele with mild Chiari II malformation characteristics was
diagnosed by fetal MRI. However, since a neural placode was
observed at birth, this spinal cord malformation was defined as MMC
with no vermian displacement.
More recent studies suggest that the morphological configuration
of the CC may play a significant role in controlling gait symmetry.
31,32
Thus, callosal alterations in MMC patients may reflect a more wide-
spread white matter abnormality,
41,42
and consequently, cause
ambulatory impairments.
31,32
On the basis of the aforementioned
studies, our data indicate that the morphology of the CC could also
have an additional influence on ambulatory status. Recently, Kunpalin
et al. has shown that CC morphology could be assessed by prenatal US
with 71.7% of fetuses with spina bifida showing CC abnormalities.
43
Our MMC patients showed different CC forms, with 8/31 (25.8%) and
4/15 (26.7%) of the patients showing “dysgenetic” forms. In one patient
(case A) with a worse than expected ambulatory status, the CC
appeared dysgenetic on fetal and postnatal MR images. However, in 1/
4 (prenatally) and 1/7 (postnatally) evaluated cases, a callosal
dysgenesis was present, despite their better long‐term ambulatory
function, as predicted by their prenatally or postnatally described
anatomical lesion levels.
Usually, a normally shaped CC and normally positioned vermis in
MMC cases are associated with a postnatal motor development that
is better than predicted by the anatomical spinal lesion level alone.
However, vermian displacement and callosal dysgenesis are present
in up to 20% (prenatally) and 3% (prenatally) of cases with better
than expected development, and thus, are not necessarily associated
with a worse outcome. Finally, our data further indirectly point to-
ward a potential beneficial effect of prenatal surgery, as the reversal
of vermian displacement may prevent an aggravation of defective
motor function due to early brainstem decompression.
In summary, CC morphology and vermian displacement
40
need
to be considered when ambulatory prediction is made by fetal MRI.
Hence, to prevent overly optimistic assumptions of the ambulatory
status in postnatally repaired patients, it is imperative to consider
those alterations when consulting MMC‐affected families.
Early reversal of posterior fossa crowding by fetal surgery may
be especially beneficial in these patients. Due to the high accuracy of
fetal MRI in assessing the level of vermian displacement, this imaging
method appears to be ideally suited to identify this subgroup of MMC
patients.
5
|
STRENGTHS AND LIMITATIONS
This study is limited by the fact that not all of our cases systemati-
cally underwent fetal MRI at the time of data collection. In this re-
gard, some information was missing.
Since US is a widely accessible method for prenatal examination
purposes, it would have been interesting to compare our fetal MRI
results with well‐structured and systematically assessed ultrasound
data. However, due to the retrospective design of our study, the
original ultrasound images were not accessible in most of our patients.
Furthermore, as data from a single center were included in this
study, the small sample size allowed the use of only descriptive
statistics.
The strength of our study is the long‐term follow‐up at a single
center, where all fetal MRI and functional neurological examinations
were conducted, and the use of a “historic” cohort that received high‐
quality prenatal MRI and a high level of postnatal multidisciplinary
FIGURE 8 Fetal and postnatal magnetic resonance (MR) images and a radiograph of an myelomeningocele patient with a functional lesion
level above L2. Left: Fetal MR image at an estimated gestational age of 36 +4 weeks shows the neural placode at Th12. Center: Two months
after birth, the radiograph shows scoliosis and multisegmental thoracic vertebral body segmentation defects. Right: Two months after birth,
the postnatal MR image shows the tethered spinal cord at Th7
930
-
KHALAVEH
ET AL.
care. This offers an important benchmark for prenatal surgery trials,
of which long‐term follow‐up data is soon expected. Moreover, our
data may serve as a reference in counseling families affected by
MMC.
6
|
CONCLUSIONS
Fetal and postnatal MRI showed a good agreement in determining
the radiological MMC lesion level. However, major disagreement
between both examinations could be observed in patients with seg-
mentation disorders and scoliosis; thus, a prediction of the radio-
logical lesion level by either of these examinations is less reliable in
these cases.
Despite the fact that postnatal MRI is the gold standard in pre-
dicting the functional level of MMC patients, similar results were
provided by fetal MRI. Both examinations have a good predictive
value for the long‐term ambulatory status in MMC patients. “Out-
liers” could be detected that showed patients with a worse or better
than expected ambulatory status. The absence of vermian displace-
ment is generally associated with a better outcome than that pre-
dicted by the assessment of the anatomical lesion level alone. The
presence of vermian displacement and callosal dysgenesis negatively
impact the long‐term ambulatory status in some, but not all, cases.
ACKNOWLEDGMENTS
None
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the
corresponding author upon reasonable request.
ORCID
Farjad Khalaveh
https://orcid.org/0000-0001-6607-7770
Andrea Reinprecht https://orcid.org/0000-0002-3019-4104
Gerlinde Maria Gruber https://orcid.org/0000-0001-7403-0788
Angelika Berger https://orcid.org/0000-0001-8775-2405
Herbert Kiss https://orcid.org/0000-0002-0712-5282
Daniela Prayer https://orcid.org/0000-0003-4612-2356
Gregor Kasprian https://orcid.org/0000-0003-3858-3347
REFERENCES
1. Prevention of neural tube defects: results of the medical research
council vitamin study. MRC vitamin study research group. Lancet.
1991;338(8760):131‐137.
2. Czeizel AE, Dudas I. Prevention of the first occurrence of neural‐
tube defects by periconceptional vitamin supplementation. N Engl J
Med. 1992;327(26):1832‐1835.
3. Bowman RM, Boshnjaku V, McLone DG. The changing incidence of
myelomeningocele and its impact on pediatric neurosurgery: a re-
view from the children's memorial hospital. Childs Nerv Syst.
2009;25(7):801‐806.
4. Canfield MA, Annegers JF, Brender JD, Cooper SP, Greenberg F.
Hispanic origin and neural tube defects in Houston/Harris County,
Texas. II. Risk factors. Am J Epidemiol. 1996;143(1):12‐24.
5. Chan A, Robertson EF, Haan EA, Keane RJ, Ranieri E, Carney A.
Prevalence of neural tube defects in South Australia, 1966‐91:
effectiveness and impact of prenatal diagnosis. BMJ.
1993;307(6906):703‐706.
6. Frey L, Hauser WA. Epidemiology of neural tube defects. Epilepsia.
2003;44(3):4‐13.
7. Rodrigues AB, Krebs VL, Matushita H, Carvalho, de, WB. Short‐term
prognostic factors in myelomeningocele patients. Childs Nerv Syst.
2016;32(4):675‐680.
8. Saitsu H, Yamada S, Uwabe C, et al. Development of the posterior
neural tube in human embryos. Anat Embryol. 2004;209(2):107‐117.
9. Zheng XY, Song XM, Chen G, et al. [Epidemiology of birth defects in
high‐prevalence areas of China]. Zhonghua Liuxingbingxue Zazhi.
2007;28(1):5‐9.
10. Albright AL, Adelson PD, Pollack IF. Principles and Practice of Pediatric
Neurosurgery. Thieme; 2008.
11. Miller E, Ben‐Sira L, Constantini S, et al. Impact of prenatal magnetic
resonance imaging on postnatal neurosurgical treatment. J Neuro-
surg. 2006;105(3):203‐209.
12. Boyd PA, Devigan C, Khoshnood B, Loane M, Garne E, Dolk H.
Survey of prenatal screening policies in Europe for structural mal-
formations and chromosome anomalies, and their impact on detec-
tion and termination rates for neural tube defects and Down's
syndrome. BJOG. 2008;115(6):689‐696.
13. Miskin M, Rothberg R, Rudd NL, Benzie RJ, Shime J. Arthrogryposis
multiplex congenita‐‐prenatal assessment with diagnostic ultra-
sound and fetoscopy. J Pediatr. 1979;95(3):463‐464.
14. Pooh RK. Neurosonoembryology by three‐dimensional ultrasound.
Semin Fetal Neonatal Med. 2012;17(5):261‐268.
15. Trigubo D, Negri M, Salvatico RM, Leguizamón G. The role of in-
trauterine magnetic resonance in the management of myelomeni-
gocele. Childs Nerv Syst. 2017;33(7):1107‐1111.
16. Carreras E, Maroto A, Illescas T, et al. Prenatal ultrasound evalua-
tion of segmental level of neurological lesion in fetuses with mye-
lomeningocele: development of a new technique. Ultrasound Obstet
Gynecol. 2016;47(2):162‐167.
17. Appasamy M, Roberts D, Pilling D, Buxton N. Antenatal ultrasound
and magnetic resonance imaging in localizing the level of lesion in
spina bifida and correlation with postnatal outcome. Ultrasound
Obstet Gynecol. 2006;27(5):530‐536.
18. Coniglio SJ, Anderson SM, Ferguson JE, 2nd. Functional motor
outcome in children with myelomeningocele: correlation with
anatomic level on prenatal ultrasound. Dev Med Child Neurol.
1996;38(8):675‐680.
19. Kollias SS, Goldstein RB, Cogen PH, Filly RA. Prenatally detected
myelomeningoceles: sonographic accuracy in estimation of the spi-
nal level. Radiology. 1992;185(1):109‐112.
20. Biggio JR, Jr, Owen J, Wenstrom KD, Oakes WJ. Can prenatal ul-
trasound findings predict ambulatory status in fetuses with open
spina bifida? Am J Obstet Gynecol. 2001;185(5):1016‐1020.
21. Peralta CF, Bunduki V, Plese JP, Figueiredo EG, Miguelez J, Zugaib
M. Association between prenatal sonographic findings and post‐
natal outcomes in 30 cases of isolated spina bifida aperta. Prenat
Diagn. 2003;23(4):311‐314.
22. Van Der Vossen, S Pistorius LR, Mulder EJ, et al. Role of prenatal
ultrasound in predicting survival and mental and motor functioning
in children with spina bifida. Ultrasound Obstet Gynecol.
2009;34(3):253‐258.
23. Rintoul NE, Sutton LN, Hubbard AM, et al. A new look at myelo-
meningoceles: functional level, vertebral level, shunting, and the
implications for fetal intervention. Pediatrics. 2002;109(3):409‐413.
24. Levine D, Barnes PD, Madsen JR, Abbott J, Mehta T, Edelman RR.
Central nervous system abnormalities assessed with prenatal
magnetic resonance imaging. Obstet Gynecol. 1999;94(6):
1011‐1019.
KHALAVEH ET AL.
-
931
25. Aaronson OS, Hernanz‐Schulman M, Bruner JP, Reed GW, Tulipan
NB. Myelomeningocele: prenatal evaluation‐‐comparison between
transabdominal US and MR imaging. Radiology. 2003;227(3):
839‐843.
26. Egloff A, Bulas D. Magnetic resonance imaging evaluation of fetal
neural tube defects. Semin Ultrasound CT MR. 2015;36(6):487‐500.
27. Saleem SN, Said AH, Abdel‐Raouf M, et al. Fetal MRI in the evalu-
ation of fetuses referred for sonographically suspected neural tube
defects (NTDs): impact on diagnosis and management decision.
Neuroradiology. 2009;51(11):761‐772.
28. Nagaraj UD, Bierbrauer KS, Stevenson CB, et al. Spinal imaging
findings of open spinal dysraphisms on fetal and postnatal MRI. AJNR
Am J Neuroradiol. 2018;39(10):1947‐1952.
29. Sherrod BA, Ho WS, Hedlund A, Kennedy A, Ostrander B, Bollo
RJ. A comparison of the accuracy of fetal MRI and prenatal ul-
trasonography at predicting lesion level and perinatal motor
outcome in patients with myelomeningocele. Neurosurg Focus.
2019;47(4):E4.
30. Chao TT, Dashe JS, Adams RC, Keefover‐Hicks A, McIntire DD,
Twickler DM. Fetal spine findings on MRI and associated outcomes
in children with open neural tube defects. AJR Am J Roentgenol.
2011;197(5):W956‐W961.
31. Richmond SB, Fling BW. Transcallosal control of bilateral actions.
Exerc Sport Sci Rev. 2019;47(4):251‐257.
32. Brodoefel H, Ramachandran R, Pantol G, et al. Association between
linear measurements of corpus callosum and gait in the elderly. Eur
Radiol. 2013;23(8):2252‐2257.
33. Sarnat HB. Regional ependymal upregulation of vimentin in Chiari II
malformation, aqueductal stenosis, and hydromyelia. Pediatr Dev
Pathol. 2004;7(1):48‐60.
34. Sarnat HB. Disorders of segmentation of the neural tube: Chiari
malformations. Handb Clin Neurol. 2008;87:89‐103.
35. Mahallati H, Sotiriadis A, Celestin C, et al. Heterogeneity in defining
fetal callosal pathology: a systematic review. Ultrasound Obstet
Gynecol. 2020. https://doi.org/10.1002/uog.22179
36. Edwards TJ, Sherr EH, Barkovich AJ, et al. Clinical, genetic and im-
aging findings identify new causes for corpus callosum development
syndromes. Brain. 2014;137(Pt 6):1579‐1613.
37. Cochrane DD, Wilson RD, Steinbok P, et al. Prenatal spinal evalua-
tion and functional outcome of patients born with myelomeningo-
cele: information for improved prenatal counselling and outcome
prediction. Fetal Diagn Ther. 1996;11(3):159‐168.
38. McDonald CM, Jaffe KM, Mosca VS, et al. Ambulatory outcome of
children with myelomeningocele: effect of lower‐extremity muscle
strength. Dev Med Child Neurol. 1991;33(6):482‐490.
39. Nagaraj UD, Bierbrauer KS, Stevenson CB, et al. Myelomeningocele
versus Myelocele on fetal MR images: are there differences in brain
findings? AJR Am J Roentgenol. 2018;211(6):1376‐1380.
40. Woitek R, Prayer D, Weber M, et al. Fetal diffusion tensor quanti-
fication of brainstem pathology in Chiari II malformation. Eur Radiol.
2016;26(5):1274‐1283.
41. Ou X, Glasier CM, Snow JH. Diffusion tensor imaging evaluation of
white matter in adolescents with myelomeningocele and Chiari II
malformation. Pediatr Radiol. 2011;41(11):1407‐1415.
42. Miller E, Widjaja E, Blaser S, Dennis M, Raybaud C. The old and the
new: supratentorial MR findings in Chiari II malformation. Childs
Nerv Syst. 2008;24(5):563‐575.
43. Kunpalin Y, Deprest J, Papastefanou I, et al. Incidence and patterns
of abnormal corpus callosum in fetuses with isolated spina bifida
aperta. Prenat Diagn. 2021. https://doi.org/10.1002/pd.5945
How to cite this article: Khalaveh F, Seidl R, Czech T, et al.
Myelomeningocele–Chiari II malformation–Neurological
predictability based on fetal and postnatal magnetic
resonance imaging. Prenatal Diagnosis. 2021;41(8):922–932.
https://doi.org/10.1002/pd.5987
932
-
KHALAVEH
ET AL.
