mri studies of the fetal brain and craniumuu.diva-portal.org/smash/get/diva2:469024/fulltext02.pdf14...

MRI Studies of the Fetal Brain and Cranium

Nuno Maria Canto Moreira Figueira de Almeida

Dissertation presented at Uppsala University to be publicly examined in Grönwallsalen,Akademiska Sjukhuset, Uppsala. Friday, February 17, 2012 at 09:15 for the degree of Doctorof Philosophy (Faculty of Medicine). The examination will be conducted in English.

AbstractCanto Moreira, N. 2012. MRI Studies of the Fetal Brain and Cranium. Department ofRadiology. 53 pp. Uppsala. ISBN 978-91-506-2268-3.

Ultrasound is the primary modality for fetal imaging, but Magnetic Resonance Imagingnowadays has a valuable complementary role as it often reveals findings that alter pregnancymanagement.

Knowledge on some clinically relevant areas of the normal fetal development is still lacking,and this was the aim of this project. We wanted 1) to obtain reference MRI data of normalbrain measurements before 24 gestation weeks (GW), 2) to study the development of thehippocampus, 3) to study the development of the ear and 4) to test the ability of MRI forevaluating the lip and palate.

For this, we retrospectively analysed a database with 464 in vivo and 21 post mortem fetalMRI examinations.

Study I evaluated a series of 70 normal fetuses. A table of normal brain measurements from17 to 23 GW was built, the first in the literature that includes ages below 20 GW.

Study II focused on the evolution of the hippocampus from 18 to 38 GW by evaluating 3 postmortem and 60 in vivo MRI examinations. Our results suggested this area to develop later andmore asymmetrically than previously thought.

Study III analysed a series of 122 normal MRI in vivo and 16 MRI post mortem. We describedthe development of the fetal ear in vivo for the first time in the literature, realizing that the valueof MRI is limited by the size of the structures evaluated.

In study IV, 60 brain-targeted MRI examinations of 55 normal fetuses and 5 fetuses withorofacial clefts were blindly reviewed by two readers, focusing on the lips and palates. Ourresults suggest a high accuracy of MRI in the evaluation of this area, regardless of fetal age orprevious ultrasound findings.

This thesis brings new knowledge on the normal development of the fetal brain and cranium.

Keywords: Fetal MRI, Brain, Cranium, Development, Normal, Ear, Lip, Palate, Biometry,Hippocampus, Measurements

Nuno Canto Moreira, Uppsala University, Department of Radiology, Oncology and RadiationScience, Radiology, Akademiska sjukhuset, SE-751 85 Uppsala, Sweden.

© Nuno Canto Moreira 2012

ISBN 978-91-506-2268-3urn:nbn:se:uu:diva-164685 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-164685)

Ao meu Pai

Tillägnad min Far

5

LIST OF PAPERS This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I. Moreira NC, Teixeira J, Themudo R, Amini H, Axelsson O, Raininko R, Wikström J.: Measurements of the normal fetal brain at gestation weeks 17 to 23: a MRI study. Neuroradiology. 2011 Jan; 53(1): 43-48.

II. Bajic D, Moreira NC, Wikström J, Raininko R.: Asymmetric development of the hippocampal region is common. A fetal magnetic resonance study. AJNR - American Journal of Neuroradiology. (Epub ahead of print 2011 Nov 24).

III. Moreira NC, Teixeira J, Raininko R, Wikström J.: The ear in fetal MRI: what can we really see? Neuroradiology. 2011 Dec; 53(12): 1001-1008.

IV. Moreira NC, Ribeiro V, Teixeira J, Raininko R, Wikström J.: Visualisation of the fetal lip and palate: is brain-targeted MRI reliable? (Manuscript)

Reprints were made with permission from the respective publishers.

7

TABLE OF CONTENTS ABSTRACT ............................................................................................................................................. 2LIST OF PAPERS .................................................................................................................................... 5TABLE OF CONTENTS ......................................................................................................................... 7ABBREVIATIONS .................................................................................................................................11INTRODUCTION .................................................................................................................................. 13

How important is fetal MRI? ........................................................................................................................... 13Historical overview .......................................................................................................................................... 13Indications for fetal MRI .................................................................................................................................. 14MR technique ................................................................................................................................................... 15Safety concerns ................................................................................................................................................ 15Evaluation of the normal development of the brain and cranium .................................................................... 16

1) Gyration ................................................................................................................................................... 162) Cerebral layering .................................................................................................................................... 173) Myelination .............................................................................................................................................. 184) Posterior fossa ......................................................................................................................................... 185) Hippocampus ........................................................................................................................................... 185) Cranium ................................................................................................................................................... 19

a) Ear ........................................................................................................................................................ 19b) Lip and palate .......................................................................................................................................20

AIMS OF THE STUDY ........................................................................................................................ 21General aim ...................................................................................................................................................... 21Specific aims .................................................................................................................................................... 21

MATERIAL ........................................................................................................................................... 23Study I ..............................................................................................................................................................23Study II .............................................................................................................................................................23Study III ...........................................................................................................................................................23Study IV ...........................................................................................................................................................23

METHODS ............................................................................................................................................ 24Study I ..............................................................................................................................................................24Study II .............................................................................................................................................................25Study III ...........................................................................................................................................................26Study IV ...........................................................................................................................................................27

RESULTS .............................................................................................................................................. 29Study I .............................................................................................................................................................. 29Study II .............................................................................................................................................................30Study III ........................................................................................................................................................... 32Study IV ...........................................................................................................................................................34

8

DISCUSSION ......................................................................................................................................... 37CONCLUSIONS .................................................................................................................................... 43FUTURE DIRECTIONS ....................................................................................................................... 43RESUMO EM PORTUGUÊS ................................................................................................................ 44SAMMANFATTNING PÅ SVENSKA ................................................................................................ 44ACKNOWLEDGEMENTS ................................................................................................................... 46REFERENCES ...................................................................................................................................... 49

11

ABBREVIATIONS ADC Apparent Diffusion CoefficientBPD Biparietal Diameter (Cerebral)CNS Central Nervous SystemdB DecibelDWI Diffusion-Weighted ImagingEAC External Auditory CanalFLAIR Fluid Attenuated Inversion RecoveryFOD Fronto-Occipital DiameterFSE Fast Spin-EchoGW Gestation WeeksLSC Lateral Semicircular CanalMR Magnetic ResonanceMRI Magnetic Resonance ImagingRF Radio FrequencySSFP Steady-State Free PrecessionSSFSE Single-Shot Fast Spin-EchoSSH GRE EPI Single-Shot Half Fourier Gradient-Recalled Echo Planar ImagingT1W T1-WeightedT2W T2-WeightedTCD Transverse Cerebellar DiameterUS UltrasoundVAP Vermian Antero-Posterior DiameterVH Vermian HeightWK Weighted Kappa

13

Introduction

INTRODUCTION

How important is fetal MRI? Ultrasound is the first line modality for fetal im-aging. Its accuracy upon detecting anomalies in the low-risk population is nevertheless quite vari-able, with detection rates that may range from 0 to 100% depending on the type of anomaly to be screened, operator skills, equipment, gestation age, national pregnancy management policy or level of diagnostic centre 1. In a large European prospective study (Eurofetus) evaluating more than 200 000 fetuses from 14 countries, an aver-age sensitivity of 61% for anomaly detection was reported 2.

However, some special ultrasound methods - as neurosonography 3 - have high detection rates that have been stated to be equivalent to MRI 4 and there has been a debate concerning the capabili-ties of the two methods 5.

It is clear nowadays that MRI and ultrasound do not compete with each other, as they do not oc-cupy a similar place in fetal management 6. MRI is a tertiary method that is most of the times per-formed to clarify ultrasound findings, often re-vealing additional information that may alter pa-tient counselling and case management 7-11.

In fact, just by being performed by distinct med-ical groups (radiologists vs obstetricians), MRI and US provide complementary standpoints that,

when part of a team work, contribute to a wider diagnostic capability.

Historical overview MRI in pregnancy was first described in 1983 12. By not using ionizing radiation, the potential val-ue of this method as a tool for evaluating the fe-tus was immediately recognized 13, 14. However, the long acquisition times of the MRI sequences available at that time rendered the observation of the fetal anatomy very difficult in practice due to motion artefacts. Studies in that decade were mostly T1-weighted and tended to be limited to older, insinuated fetuses, or to be targeted for vol-umetric measurements 15, 16 that used faster echo-planar sequences. Even if fetal motion control could be achieved in some instances by invasive procedures, such as curarization of the umbilical cord 17, 18, all these limitations contributed then to restrain the technique to a small number of uni-versity centres.

The availability of single-shot MR sequences in the 90’s made it possible to obtain T2-weighted images in 1 second or even less. These sequences allow each image to be acquired separately, thus highly benefiting a movement-prone method such as fetal MRI 19, 20 that became increasingly more popular. The number of available sequences that

Figure 1.Examples of sequences commonly used in fetal MRI a) T2-weighted SSFSE, b) T1-weighted FLAIR, c) DWI and d) SSFP thick-slab dynamic sequence.

14


Table 1. Major indications for fetal MRI versus ultrasound. Modified from Pistorius et al 6.

can efficiently cope with fetal motion is neverthe-less quite limited even today (fig. 1) and we cher-ish the recent development of a very promising single-shot T1-weighted sequence 21, that was still lacking in daily practice.

Indications for fetal MRI Indications for performing a fetal MRI may be related to the mother, the fetus, or both. Mater-nal causes include obesity or other aspects that might otherwise compromise an US examination, whereas fetal causes most of the times involve a) the clarification of previous US findings, b) the screening for genetic pathology or c) to rule out diseases that might not be detectable by US 22.

Traditionally fetal MRI has been mostly targeted for the central nerv-ous system, as this field is particu-larly challenging for US and the fast T2-weighted sequences were ideal for providing contrast within the unmy-elinated fetal brain and towards the surrounding cerebrospinal fluid. MRI is now currently acknowledged to be also important in body pathology 22, namely of the lungs, kidney and liver, but this is beyond the scope of our work.

Major fetus-related indications for performing a fetal MRI in neuroradi-ology include all neuronal migration

disorders, corpus callosum dysgenesis and poste-rior fossa malformations 8, 10, 23. However, in or-der to obtain the best diagnostic result, a precise knowledge of the diagnostic capabilities of US is also needed, or when to combine both methods 6, as summarized in table 1.

Attemptive guidelines for the use of MRI in fetal medicine have been published 6, 24, but the precise indications and timing for this procedure still vary from country to country, in relationship to factors as local legislation for pregnancy management 25, differences in the number of fetal MRI centres, or the existence of specialised ultrasound techniques like neurosonography 3, 4, 26, 27.

Pathology US MRI both eitherScreening X

MRI contra-indicated or failed X

Early diagnosis (until 19 GW?) X

Assessment of fetal movements (up to 3rd trimester) X

Assessment of cerebral blood flow X

Evaluation of associated abnormalities X

Oligohydramnios X

Engaged fetal head and ruptured membranes X

Assessment of fetal movements (late 3rd trimester) X

Posterior fossa abnormalities X

Detecting and determining age of intracranial bleeding X

Detecting intracranial tuberous sclerosis X

Schizencephaly X

Acute asphyxia X

Postmortem brain imaging X

Severe microcephaly X

Corpus callosum and pericallosal abnormalities X

Fetal brain death X

Cytomegaloviral infection X

Intracranial tumors X

Trauma X

Vein of Galen abnormalities X

Germinal matrix and intraventricular bleeding X

Suspected hemimegalencephaly X

Septo-optic dysplasia X

Ventriculomegaly X

Holoprosencephaly (after 20 weeks) X

Corpus callosum abnormalities (after 20 weeks) X

Craniosynostosis X

15

Introduction

MR technique In modern MRI units, there are no special hard-ware requisites for performing a fetal examination. Mothers lie supine or in right lateral decubitus (if late in pregnancy), care being taken to maximize comfort and reduce noise. Fetal motion and peri-staltism may be reduced by fasting 28, or by means of a mild sedative given to the mother 29, but these methods are not in use in our institutions. Instead, we try to provide a calm environment for the mother and to perform the examinations outside the circadian peaks of fetal movement 30, although realizing that this is not always possible in daily practice.

Cardiac or flexible abdominal coils are placed around the maternal abdomen, trying to obtain the maximum signal-to-noise ratio for the organ to be studied. This attemptive targeting is mostly meant for older fetuses, since the precise position of younger ones that can move more freely inside the abdominal cavity is harder to predict.

The imaging protocol in our institution in Portu-gal (table 2) changes somewhat according to dif-

ferences in fetal size or the precise clinical prob-lem at stake but, routinely, ultrafast single-shot T2-weighted images are obtained sequentially in three orthogonal planes of the fetus, followed by axial T1W, axial DWI and a sagittal thick-slab dy-namic sequence. Most CNS-targeted studies take less than 30 minutes to perform but, as a safety precaution, they are usually interrupted when 45 minutes of examination time are reached.

Safety concerns Fetal MRI at 1.5 Tesla is generally accepted to be a safe procedure per se, also with no evidence of long-term harmful consequences 31, 32. There are, however, three potential areas of concern regard-ing fetal safety: the static field, the field gradients and the radiofrequency pulses.

The possibility that the static magnetic field of MRI by itself could have a an effect on fetal growth in mice was reported in 1988 33. Even if

Table 2. Standard brain Fetal MR Imaging protocol at Dr. Campos Costa (Philips Intera 1.5T). (1) Matrices for the T2W SSFSE sequence have been evolving over time, from 256*256 to 512*512.

Order of acquisition

Sequence Plane FOV (mm)

Slice thick.(mm)

Gap(mm)

Matrix

1 SSFP Survey 3 - plane 400 8 0 224*209

2-4 T2W SSFSE 3 - plane 200-300 3-4 0.3 256*256 (1)

5 DWI (b=1000+ADC) axial 230-250 4 1 432*104

6 FLAIR/T1W axial 270-330 4 1 512*256

7 SSFP Dynamic scan sagittal 350 10-15 0 256*192

Optional

8 T2* GRE EPI axial 230-250 5 1 512*256

9 FLAIR /T1W coronal 200-240 4 1 512*256

16


there is no evidence to date of such an effect in human fetuses when undergoing MRI studies at 1.5 Tesla units 24, 34, MR examinations are usually only performed upon the completion of organo-genesis, after 16 gestational weeks 35.

Gradients currently used in MRI may reach 120 dB in noise intensity. As such, hearing damage to the fetus has been considered a potential hazard, but this was not confirmed in practice 36, 37.

The fast imaging sequences that are used to re-duce the impact of fetal motion on image qual-ity usually induce a high radiofrequency specific absorption rate (SAR) and therefore depose heat on the mother and fetus. As heat is a known muta-genic 38, potential increases in fetal body tempera-ture during MR examinations have been investi-gated 39, 40, leading to guidelines on the maximum amount of exposure to radiofrequency radiation that is considered to be safe for the fetus 41.

It is known however that in fetal MRI the actual RF exposure period does not normally exceed a third of the examination time 42 and a potential harmful heating of the fetus has not been proven in clinical practice, but the issue is at the centre of the present limitation to scan fetuses at mag-nets no stronger than 1.5 Tesla. Otherwise, the ex-pected better signal-to-noise ratio of higher field strengths could be highly beneficial in this area that deals with very small subjects, even if early

reports suggested that increasing artefacts could in fact reduce the quality of fetal MRI at 3T 43.

Evaluation of the normal development of the brain and cranium To rule out pathology, knowledge about the nor-mal development of the fetal brain and cranium in vivo is needed. MRI is an excellent tool for that purpose, in particular for the evaluation of 1) the gyration pattern, 2) the layering of the cerebral mantle, 3) premyelinating and myelinating pro-cesses and 4) posterior fossa. Of interest for our project, we must also refer to it’s role for the study of 5) the hippocampus and 6) the cranium/base of the skull.

1) Gyration The value of recognizing a specific gyration pat-tern for each gestational age in order to assess maturation has long been recognized in vitro 44. This has also been attempted by ultrasound 45, but it was only with MRI that the chartering of the sulcal and gyral development of the human brain in vivo was conclusively obtained 46.

The evolution of brain sulcation can be used to evaluate fetal age after approximately 24 GW, as the brain is essentially agyric until then (fig. 2). In

Figure 2.Fetuses of 20, 24 and 34 GW, from left to right. Note the progressive deepening of the Sylvian fissures with age (black arrows). At 24 GW, the calcarine sulci start to be delineable (white arrow). By 34 GW, the primary and secondary sulcation is complete.

17

Introduction

75% of fetuses, the parietooccipital sulcus and the callosal sulcus can be detected by 23 GW, where-as the calcarine sulcus is visible by 24–25 GW and the central sulcus by 26 GW 46. All primary and some secondary sulci are supposed to be visible on fetal MRI by 34 GW 46. To be noticed, how-ever, that the gyration pattern on MRI appears to lag by an average of two GW when compared with autopsy specimens 46.

As gyration is not usable as a marker for brain maturation in young fetuses, measurements of brain growth are needed for that purpose. This approach has been the gold-standard for fetal age evaluation with ultrasound 47, 48, but data for MRI are also needed, as this method partly evaluates different structures and takes measurements in a different manner than US.

Research has been made with MRI on volumet-ric assessment of the brain growth by using seg-mentation methods 49-52, but these are time-con-suming and difficult to perform in daily practice. More clinically-applicable tables with linear brain measurements are still scarce 29, 46, 53-55, in partic-ular data for young ages, as only one published study exists engaging fetuses before 23 GW 54, and none before 20 GW.

2) Cerebral layering A layering pattern of the cerebral mantle can al-ready be seen at histology and in vitro MRI at 10 GW 56. This is constituted by the T2-hypointense ventricular zone and cortical plate, both separated by the cell-sparse intermediate zone.

By the time clinical MRI is performed, after the end of embryogenesis, a more complex 5-layer pattern is definable, as seen in figure 3. This tran-sient appearance fades out on MRI by approxi-mately 28-29 GW, even if it is still recognizable on histology later in pregnancy 56, 57.

The first group of future cortical neurons and migrating glia is produced at the ventricular zone from 8 to 16 GW 58. Later on, the outer subven-tricular zone becomes the major source of glial cells, and also of the more superficially located neurons of the developing six-layered cortex 59.

The intermediate zone - that will become the white matter proper - contains growing axonal pathways, but also an intense proliferation of oli-godendrocytes and astrocytes to guide neuronal migration, this cell density being the responsible for the comparatively low signal the layer has on T2W images 60.

The subplate is also an important transient structure of the mid-fetal life. This is where most neuron apoptosis occurs 58, but it is also a “wait-ing” area for growing cortical afferents that are surrounded by a very large extracellular space 61. This T2-hyperintense layer becomes progressively less evident as synaptogenesis evolves at the third trimester, totally fading on MRI after 28-29 GW.

The outermost layer, the cortical plate, receives the afferents from the subplate and it remains the only laminar compartment that is well-delineated in MRI scans until birth 56.

Figure 3. Fetuses of 25 (left) and 29 GW (right). The ventricular zone (VZ), the subventricular zone (SVZ), the intermediate zone (IZ), subplate (SP) and cortex (CP) can be identified at the younger fetus, but the layering is much less distinct at the older one.

18


Other transient brain structures of importance are the ganglionic eminences - that will later form the basal ganglia and thalamus - and the recently described periventricular crossroads 62. The latter are major areas of intersection of fiber systems, situated at predilection sites for hypoxic-ischemic injuries 62, thus suggesting a selective local vul-nerability by means that are not yet fully under-stood.

3) Myelination Ultrasound, on the contrary to MRI, is unable to assess myelination. This progressive process that spans from mid-gestation to adulthood, has known milestones on MRI that can be trustfully used to evaluate the maturation of the brain 63. To know these milestones is very important clini-cally, since disturbances in myelination are a fair-ly common component of hundreds of diseases, from hypoxic-ischemic to metabolic processes.

In fetal MRI in vivo (fig. 4), a T2-hyposignal in relationship to myelination can already be seen at 28 GW in the posterior fossa, namely at the grac-ile and cuneate nuclei, vestibular nuclei, cerebellar vermis, inferior and superior cerebellar peduncles, dentate nucleus, medial longitudinal fasciculus, inferior olivary nuclei, medial lemnisci, lateral lemnisci and inferior colliculi, but also at parts of the diencephalon as the medial geniculate bodies, subthalamic nuclei and ventrolateral nuclei of the thalamus 64. From this gestational age on, myeli-

nation is not visualized at any new site until 36 GW, when myelin can be seen in the corona ra-diata, posterior limb of the internal capsule, cor-ticospinal tracts and lateral geniculate bodies 64.

4) Posterior fossa MRI is excellent for the evaluation of the poste-rior fossa, as it allows a direct visualization of the cerebellar hemispheres, vermis and brainstem in three orthogonal planes. In a recent study 65 fe-tal MRI was able to rule out pathology in 28% of 90 fetuses referred for suspected posterior fossa anomalies on ultrasound. However, in the same study, only a 60% agreement existed between the pre- and postnatal MRI findings and an early ges-tational age at the time of examination was con-sidered an important factor for such a disparity. Therefore, opinion has built 23, 66 that MRI find-ings should be taken with caution when evaluating the posterior fossa early in gestation, especially in the case of the vermis or brainstem.

5) Hippocampus Morphological studies of the fetal hippocampus have been focusing more on the cellular types and organization than on its general shape. Three stud-ies dealt with MRI evaluations of formalin-fixed specimens 56, 67, 68, but with this type of approach the anatomical shape and proportions may change when the brain is fixed in formalin or prepared for the histological examination. Therefore, studying

Figure 4. Fetuses of 28 GW (left side) and 37 GW (right side).The T2-hyposignal of myelina-tion (arrows) can be seen at the colliculi, dorsal pons and gracile nucleus in the younger fetus, and at the posterior limb of the internal capsule (arrow) in the older one.

19

Introduction

the hippocampal development in vivo is pertinent.The hippocampal development begins at 8 GW

(fig. 5) and the hippocampal sulcus becomes vis-ible at 10 GW 69. This sulcus progressively deep-ens as a part of the hippocampal inversion pro-cess, that has been reported to be completed at approximately 21 GW 67-71, even if some data sug-gest it might take longer 72. The fully developed hippocampus is oval in coronal slices, and Baker and Barkovich have suggested that this oval shape is preceded by a round or pyramidal shape 73. In-deed, this non-oval shape has been shown to be more common in younger than older premature neonates in an ultrasound study 72 but it may per-sist throughout life in 19% of the general popula-tion, most often on the left side 74, 75.

On MRI in vivo, the hippocampal sulcus can be seen at 22-23 GW and the collateral sulcus at 23-26 GW 76. There is one in vivo fetal MR study focusing on the hippocampal infolding process, from 20 GW until 37 GW 77, but developmental changes in its shape have not been assessed before with this technique.

5) Cranium Detailed in vitro MRI studies regarding the devel-opment of the fetal cranium have been published 78, 79, but there is paucity in fetal MRI literature in vivo in what concerns this area, probably due to the fact that the bone and muscle prevailing here do not provide the same high T2 contrast that is available for the endocranium. The exceptions are reports on some structures with a high water con-tent, as the eye globes 80, 81 or the lacrimal ducts 82.

In particular, prenatal data in vivo for the evalu-ation by MRI of the ear and the normal lip and palate are scarce, but they would be of great clini-cal relevance:

a) Ear A vast amount of congenital syndromes are ac-companied by ear involvement 83, 84. 3D ultrasound is useful to delineate the pinna, it helps to rule out local malformations and can give evidence of ane-uploidy 85-87 but the capacity of the method to rule out temporal bone pathology is very limited 88.

Week 9 Week 10-11 Week 14 Week 21

1

4HCS

32

Figure 5. Schematic representation of the fetal hippocampal development. 1. Gyrus dentatus. 2. Cornu ammonis. 3. Subiculum. 4. Parahippocampal gyrus. HCS. Hippocampal sulcus.

20


The ear is also very difficult to evaluate at brain-targeted fetal MRI studies, as these are prone to suffer from movement, slice tilting and also from partial-volume averaging. Some isolated reports do exist on inner ear malformations depicted by fetal MRI 89, 90 and there is even a published se-ries addressing the development of the middle ear 91, but a comprehensive evaluation of which structures of the ear can be seen on fetal MRI at different stages of pregnancy is still lacking in the literature.

b) Lip and palate Ultrasound has been the gold standard for the pre-natal screening of the lip and palate, but its accu-racy is very heterogeneous, depending heavily on the user and examination protocol 92. However, its ability to identify facial anomalies has been im-proving significantly in recent years, even in non-selected screening examinations 93.

The most common facial malformations are cleft lips, with or without an associated cleft palate, at-taining approximately 1/700 live births 94, 95. The importance of detecting these entities in the pre-natal period is several-fold as it modifies patient

management 96 and parental expectations 97, but also due to the fact that a cleft lip is associated to a genetic syndrome in more than 10% of patients 98.

At least five prospective studies 98-102 and one retrospective study 103 have dealt with the role of MRI in the evaluation of orofacial clefts. All agree that MRI provides a better evaluation of the secondary palate than US and only one study 99 did not find the method to be more helpful when identifying primary palate involvement. MRI has also been shown to be important in the evaluation of the intracranial anomalies that may exist in as-sociation to clefts, these ranging from 6.3% 104 to 23% 103 in previous reports.

In only one of these MRI series 100, regarding fe-tuses in the third trimester, the examinations were read in a blind fashion. In all other studies the reader, the MRI protocol or both were not blinded to previous ultrasound findings. Therefore, the question of how well does MRI independently detect orofacial anomalies remains unanswered, even knowing that in one series with 34 fetuses the method disclosed 5 (14.7%) clefts that were not previously seen by ultrasound 103.

21

Aims of the study

AIMS OF THE STUDY

General aim To evaluate with MRI the normal development of some clinically relevant structures of the fe-tal brain and cranium that previously are insuf-ficiently described. We were particularly focused on the period before 25 GW, as this is a time limit for pregnancy management in a large number of countries 25.

Specific aims 1) To obtain brain measurements for MRI of

normal fetuses before the end of the second trimester of gestation.

2) To evaluate with MRI in vivo the development of the fetal hippocampus.

3) To evaluate with MRI in vivo the development of the fetal ear.

4) To test the reliability of brain-targeted MRI in vivo for the evaluation of the lip and palate across gestation.

22


Table 3, (study IV). Post-natally confirmed facial clefts

Fetus nº GW Cleft lip Cleft primary palate Cleft secondary palate

1 23 Bilateral Bilateral Yes

2 23 Bilateral Bilateral No

3 26 Unilateral Unilateral Yes

4 27 Unilateral No No

5 37 Unilateral Unilateral No

GW = gestational week

23

Materials and Methods

MATERIAL The material for this project came from three sources:A) A database of 71 in vivo fetal MRI examina-

tions performed between 2004 and 2006 at Uppsala University Hospital (Sweden). This source contributed with fetuses for studies I and II.

B) A database of 393 in vivo fetal MRI examina-tions performed between 2002 and 2010 at the Consultório Dr. Campos Costa, Porto (Portu-gal). This source contributed with fetuses for all studies.

C) A database of 21 post mortem fetal MRI per-formed between 2006 and 2010 at Uppsala University Hospital of non-fixed, aborted fe-tuses that were examined before autopsy. This source contributed with fetuses for studies II and III.

Study I Out of all MRI examinations of singleton fetuses aged from 17 to 23 GW from databases A and B, we selected the studies that fulfilled the following criteria: 1) to have been evaluated as normal in what concerned the brain examination, and 2) to have T2-weighted images in orthogonal sagittal, coronal and axial planes, with no obliquities or movement artefacts. A total of 70 examinations was obtained.

The main indications for the MRI studies were suspicion of mild ventriculomegaly (n=35), brain screening in case of non-CNS abnormalities (n=13), abnormal biochemical screening tests (n=9) and US suspicion of structural brain anom-aly (n=9).

In nine cases, the pregnancy was terminated due to non-CNS disorders. All these fetuses had normal CNS findings at autopsy. The remaining 61 fetuses were delivered, and all neonates had a normal neurological status and clinical biometric parameters between the 10th and 90th percentiles at birth.

Study II Twelve fetuses from database C and 306 from databases A and B were assessed. All the fe-tuses that underwent post mortem examinations in which central nervous system pathology was revealed in autopsy or MRI were excluded. For the in vivo group, all the subjects with brain pa-thology, including ventriculomegaly (even mild), and the examinations in which the hippocampal regions were not well assessable bilaterally or the coronal slices were not symmetrically positioned were also excluded. Three post mortem examina-tions and 60 in vivo examinations were accepted for final analysis.

Study III Post mortem group: 16 fetuses from database C, with ages ranging from 16 to 22 GW.

In vivo group: From database B, we retrospec-tively reviewed the brain investigations that ful-filled the following criteria: 1) to have been re-ported as normal or with mild ventriculomegaly, 2) to have T2-weighted images in orthogonal sag-ittal, coronal and axial planes, with no movement artefact and 3) to have a normal clinical evaluation at the perinatal period, including normal audiom-etry values at the newborn hearing-screening test (UNHS). 122 examinations were obtained in to-tal, from 116 fetuses.

Study IV From database B, we selected all the examinations in which a cleft lip or cleft palate had been diag-nosed and confirmed post-natally. A total of five fetuses were found, presenting with seven cleft lips, six cleft primary palates and two cleft sec-ondary palates (table 3, see left). All of them had a previous suspicion of orofacial pathology at ultra-sound that lead to the MRI examination.

A group of 55 gestational age-matched normal brain MRI studies was also selected from the same database, covering fetuses from 20 to 38 GW. All of them had a post-natal confirmation of a normal clinical status.

24


METHODS The in vivo MR studies were performed in Portu-gal and Sweden using 1.5 Tesla MRI units (Philips Intera, Eindhoven, Netherlands). Abdominal phased-array flexible coils were used. No sedation was given to the mothers-to-be.

The imaging protocols in vivo can be seen in ta-ble 2.

The post mortem examinations were performed using a birdcage knee-foot coil. T2-weighted 2D FSE sequences were obtained in the sagittal, coro-nal and transverse planes, with a slice thickness of 2 mm and in-plane resolution of 0.50x0.53 mm (transverse plane) and 0.59x0.62 mm (coronal and sagittal planes).

The gestational age at the moment of the MRI studies was based on ultrasound data and defined as the number of complete gestational weeks.

One reviewer of study II (DB) was a pediatric radiologist with a large experience in MRI. The other reviewers of the studies - NCM, JT and VR -

were senior neuroradiologists with 12, 10 and 6 six years experience in fetal MRI, respectively.

The statistical treatment of the data included Mann-Whitney non-parametric tests to evaluate scoring levels, and kappa evaluation of interob-server disagreement 105.

The ethics committees of the two centres ap-proved the study protocol.

Study I Five measurements of the brain were taken, ac-cording to the criteria published by Garel et al 29, 46 (see fig. 6).

The examinations were reviewed and measure-ments made by two neuroradiologists (NCM and JT). In case of disparity, a measurement consen-sus was obtained. For each GW and parameter, the median, maximum and minimum values were recorded and rounded to the nearest millimetre.

Figure 6.Fetus of 20 GW. T2W-SSFSE. a) Fronto-occipital diameter (FOD). Maximal occipital to frontal distance of the brain. Sagittal plane. b) Cerebral biparietal diameter (CBD). Maximal transversal diameter of the brain above the Sylvian fissures.

Coronal plane. c) Transverse cerebellar diameter (TCD). Largest cerebellar diameter. Coronal plane. d) Vermian height (VH) (dashed line) and antero-posterior diameter (VAP) (full line). Sagittal plane.

25


Study II The hippocampal regions were assessed in one or more coronal slices (fig. 7). A comparison be-tween the right and left side was made in every case. The side of the heart was regarded as the left side of the fetus, but the liver and the gastric chamber were also used as additional visual refer-ences.

The angle between the upper and lower lip of the hippocampal sulcus was measured. When this sulcus was closed, then the shape of the hip-pocampus was classified as non-oval or oval. The presence and orientation of the collateral sulcus

Figure 7. Images of the hippocampal regions in coronal planes.a) Fetus of 25 GW. The arrows indicate the hippocampal sulci. The hippocampal sulcus is larger on the left side.b) Fetus of 29 GW. Closed hippocampal sulcus and a non-oval hippocampus with a vertical long axis bilaterally.

The arrows indicate the hippocampal region and the stars the collateral sulcus.c) Fetus of 29 GW. Closed hippocampal sulcus and an oval hippocampus with a horizontal long axis bilaterally.

The white arrows indicate the hippocampal region, the black arrow closed hippocampal sulcus and the star the collateral sulcus.

was also recorded. If the angle between the col-lateral sulcus and the hippocampus was more than 70 degrees, it was defined as vertical 74.

Two reviewers (DB and NCM) made independ-ent evaluations and the interobserver agreement was calculated. In case of disagreement, the two radiologists re-evaluated the images and a consen-sus was achieved.

Differences in proportions were analyzed by the likehood ratio χ2 method. P values < .05 were considered significant. All tests were 2-tailed.

26


Study III The cochlea, vestibular apparatus, middle ear

and external auditory canals were evaluated by one reader (NCM) that had no information about the gestational age of the fetuses.

Each ear was assessed individually and classi-fied as follows:

Cochlea (fig. 8):

Level 0: Not clearly identifiable. Level 1: Identifiable, but no individual turns are

depicted. Level 2: Turns are seen. Possibility to exclude

cochlear aplasia, cystic cavity or severe hypoplasia.

Level 3: Most turns are identifiable. Possibility to exclude a classic Mondini malformation, but minor partition abnormalities cannot be ruled out.

Level 4: Complete cochlear identification.

Vestibule and semicircular canals (fig. 9):

Level 0: Not clearly identifiable. Level 1: Identifiable and separated from the coch-

lea, but no detail. Level 2: Well-defined vestibule, but LSC not

totally delineated. Level 3: LSC totally delineated. Level 4: Complete delineation of the vestibule

and the semicircular canals.

Middle ear (fig. 10):

Level 0: Middle ear not identified. Level 1: Middle ear depicted but not enough

detail as to identify ossicles. Level 2: Ossicles depicted but not discriminated. Level 3: Possibility to discriminate each ossicle.

External ear (fig. 11):

Level 0: EAC not depicted.Level 1: EAC depicted unilaterally.Level 2: EAC depicted bilaterally.

In the case of different scores arising between the left and right ear, the lower score was used for classification.

In order to test inter-observer variability of the in vivo group, a random sample of 25 fetuses was evaluated independently by another experienced neuroradiologist (JT), this one being also blind to the clinical outcome of the fetuses. For final ana-lysis, only the scoring of the main reader was used.

27


Figure 8. Examples of scoring levels for cochlear evaluation (in vivo series): a) Fetus of 21 GW – level 1, (coronal plane).b) Fetus of 28 GW – level 3, (coronal plane).

Figure 9. Examples of scoring levels for vestibular evaluation (in vivo series): a) Fetus of 22 GW – level 1, (axial plane). b) Fetus of 28 GW – level 3, (axial plane).

Figure 10.Examples of scoring levels for middle ear evaluation (in vivo series): a) Fetus of 23 GW – level 1, (coronal plane). b) Fetus of 34 GW – level 2, (coronal plane).

Figure 11.Examples of scoring levels for external ear evalua-tion (in vivo series): a) Fetus of 34 GW – level 1, (coronal plane). b) Fetus of 32 GW – level 2, (axial plane).

Study IV All the 60 studies were reviewed independently by two senior neuroradiologists (VR and JT). They were blinded to the aims and methodology of the study, including the presence or absence of pathological cases in the cohort. If there were any sequences specially directed for the face - as in the fetuses with a previous suspicion of a cleft at US - these were removed from review to prevent reading bias.

The readers were asked to grade the delineation and normality of the upper lip, primary palate, secondary palate and nasal septum of each fetus,

according to the following scale:

Level 1: Evidently normal. Level 2: Probably normal, even if not completely

delineated. Level 3: Probably abnormal, even if not com-

pletely delineated. Level 4: Evidently abnormal.

If pathology was suspected, the readers were also asked to make an assumptive diagnosis.

28


29

Results

RESULTS

Study I Seventy fetal MRI examinations were included in the study. 22 of them were performed in Sweden and 48 in Portugal. The planned measurements were obtainable at all studies.

Age distribution and the results for each GW are shown in table 4.

Table 4. Age distribution of the fetuses and measurements from 17 to 23 gestation weeks. GW = Gestation week; n = Number of fetuses; FOD = Fronto-occipital diameter; CBD = Cerebral biparietal diameter; TCD = Transverse cerebellar diameter; VH = Vermian cranio-caudal diameter; VAP = Vermian anterio-posterior diameter.

GW Measurement Median Rangemm mm

FOD 49 45-49CBD 32 31-35

17 (n=6) TCD 16 15-17VH 7 6-8

VAP 4 4-5

FOD 49 43-54CBD 33 32-35

18 (n=6) TCD 17 15-17VH 7 6-8

VAP 5 4-6

FOD 49 42-56CBD 35 31-41

19 (n=8) TCD 18 17-19VH 8 7-8

VAP 5 5-6

FOD 54 45-63CBD 39 37-42

20 (n=15) TCD 20 18-21VH 9 7-11

VAP 6 4-9

FOD 59 49-65CBD 42 38-47

21 (n=14) TCD 21 20-23VH 9 7-12

VAP 7 6-9

FOD 62 57-70CBD 45 41-48

22 (n=15) TCD 23 21-24VH 11 9-12

VAP 7 6-9

FOD 64 60-66

CBD 46 44-4723 (n=6) TCD 25 24-27

VH 12 12-13VAP 8 8-9

30


Study II The three post mortem examinations were per-formed at 17-18 GW and the 60 in vivo examina-tions at 19-36 GW. The inter-observer agreement for the hippocampus and collateral sulcus was 90% and 97%, respectively.

The hippocampal sulcus could be seen in all fe-tuses. From figure 12, it can be seen that it was open in 39 of the 51 fetuses that were examined from 17 GW until 32 GW, but not later. In 35 of these, the sulcus was open bilaterally but in 21 the angle between its upper and lower lip was already less than 90 degrees bilaterally. Asymmetry was observed in the closing process in 11 fetuses that had both hippocampal sulci still open; the nar-rower was found on the right side in 8 fetuses and on the left side in 3.

A closed hippocampal sulcus was found in 28 fetuses. In four of these, from 21 to 30 GW, the right sulcus was closed but the left one was still open.

Non-oval hippocampal shapes could be found bilaterally at 24-29 GW, and unilaterally still at 31-36 GW (fig. 12). In the 12 fetuses aged GW 33 or

more, when all the hippocampal sulci were closed, the shape of the hippocampus was still asymmet-ric in 4 (33%), being non-oval on the left side and oval on the right.

The collateral sulcus (fig. 13) was detected at the earliest at 17 GW (post mortem) but in one subject this sulcus could not be seen even at 29 GW. A deep, well-defined, collateral sulcus was seen at the earliest at 26 GW on the right and at 27 GW on the left.

In 14 (58%) of the 24 fetuses that presented a bilateral deep collateral sulcus, there was a uni- or bilateral vertical orientation. The collateral sulcus was vertical in 5/26 fetuses (19%) on the right side and in 13/25 fetuses (52%) on the left side, a differ-ence that was statistically significant (p = 0.014). In the nine fetuses that associated a deep collater-al sulcus with a non-oval hippocampal shape, six had a vertical orientation of the collateral sulcus.

In total, asymmetric hippocampal development was found in 26/63 fetuses (41%). The develop-ment was faster on the right in 23 fetuses and on the left in 3.

31

Results

R LR L

R LR L R L R L

R L

R L R LR L

R LR L R L R L

R L R L L RL R L RL R

17

1

4

32

567

8

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Age of the fetus (GW)

Number of fetuses

Figure 12. Closure of the hippocampal sulcus and the shapes of the hippocampi in fetuses with a closed hippocampal sul-cus. Each square pair describes a fetus. R = right side, L = left side. = open hippocampal sulcus, = non-oval shape of the inverted hippocampus, = oval shape of the inverted hippocampus, the long axis horizontal (complete inversion).

Figure 13. Collateral sulcus at different gestation ages. Each square pair describes a fetus. R = right side, L = left side. Collateral sulcus: = not visible, = shallow, = vertical, = oblique or horizontal.

R LR L

R LR L R L R L

R L

R L R LR L

R LR L R L R L

R L R L L RL R R L R L

17

1

4

32

567

8

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Age of the fetus (GW)

Number of fetuses

32


Study III Post mortem group: All the structures of the in-ner and middle ear were normal and they could be visualised with the maximum detail on our scal-ing score (fig. 14). The outer margin of the lateral semicircular canals could only be delineated on axial slices in 2 of the 16 fetuses (12.5%) but its normality could always be confirmed at coronal slices.

The external auditory canals could be delineated in 12.5% (2 /16) of the fetuses.

In vivo group: Inter-observer agreement was 84% for the cochlea (WK = 0.67), 65% for the ves-tibular apparatus (WK = 0.49), 80% for the middle ear (WK = 0.56) and 84% (WK = 0.83) for the external ear.

Of all 122 examinations, symmetry of the find-ings was observed in 96% of the cases for the cochlea, 93% for the vestibule and 97% for the middle ear.

The cochlea (fig. 15a) could be identified in all examinations after GW 21, although never with sufficient detail to visualise all its components (i.e. detail level 4). In only two examinations, at GW 27 and 38, was it possible to delineate all of

the cochlear turns (level 3). Nevertheless, coch-lear turns could be identified in 75% of all the ex-aminations in our cohort. Before 25 GW cochlear turns were seen in 50% (21/42) of cases, and in 89% later on.

The vestibular apparatus (fig. 15b) could always be depicted after GW 23 but a complete identifica-tion of all its components could never be reached. It was possible to identify the vestibule and the lateral semicircular canals in 72% (88 /122) of all the examinations, and in all of those above 33 GW. Before and after 25 GW, respectively, these structures could be delineated in 31% (13/42) and 93% (74/80) of cases. A complete delineation of the lateral semicircular canals (level 3) could only be achieved in 30 examinations (25%), all but one later than GW 24.

Ossicles (fig. 15c) could be identified in 70% of the examinations (86/122), and in one case as early as GW 21. The ability to identify the ossicles was only 33% (14/42) before 25 GW, but afterwards it reached 90% (72/80). The image detail was never sufficient to discriminate the malleus from the in-cus, or to identify the stapes.

Figure 14. Post mortem series. a) Fetus at 22 GW, coronal plane. All the cochlear turns are identifiable. The head of the malleus and the exter-

nal auditory canal are also seen. Note the low signal of the EAC at this age, occluded by an ectodermal plug.b) Fetus at 17 GW, axial plane. Inside the cochlea (arrow), it is possible to differentiate the scala vestibularis

from the scala tympanica. The modiolus is also seen on the contralateral side. c) Fetus at 18 GW, axial plane. The LSC is almost completely delineated, at the exception of its outer margin

(arrow). d) Fetus at 20 GW, coronal plane. The malleus head, long process of the incus and even the stapes are identified.

33

Results

Figure 15. In vivo series (n=122). Ability to identify components of the a) cochlea, b) vestibule / LSC, c) middle and d) external ear from 20 to 38 GW. The level of detail for each structure (see definitions in methods) is displayed on a gray scale. Level 4 detail was never reached in this group.

Before 25 GW, the external auditory canals (fig. 15d) could only be seen in two examinations, and only on one side. Both EAC were identified at 26

GW at the earliest, and in 59% of cases after 29 GW (29/49 examinations). In one case, the EAC could not be identified as late as 34 GW.

Cochlea

0

2

4

6

8

10

12

14

16

18 level 3

level 2

level 1

level 0

level 3

level 2

level 1

level 0

level 2

level 1

level 2

level 1

level 0

External ear

0

2

4

6

8

10

12

14

16

18

20

Middle ear

0

2

4

6

8

10

12

14

16

18

Vestibule

0

2

4

6

8

10

12

14

16

18

Gestation weeks

d

c

b

a

Num

ber

of fe

tuse

s

21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

34


Study IV The readers could interpret all the 60 examina-tions, and the results are summarized in table 5.

The interobserver agreement (weighted kappa) was 0.79 for the upper lip, 0.70 for the primary palate, 0.86 for the secondary palate and 0.90 for the nasal septum. By clustering scoring levels 1+2 and 3+4, i.e clustering probable and evident nor-mality on one hand, and probable and evident ab-normality on the other, interobserver agreement (linear kappa) was 0.81, 0.78, 0.73 and 1 respec-tively.

There was no statistically significant differences in the scoring levels of the readers across gesta-tional age (p ≥ 0.58).

Of the 55 normal upper lips, 53 (96%) were cor-rectly identified by both readers when the scores of the levels 1 and 2 were combined. However, reader 1 incorrectly suspected an abnormal upper lip in two fetuses of 34 and 36 GW.

Of the 56 normal primary palates, 54 (96%) were correctly identified by both readers (levels 1 and 2 combined), but reader 1 also suspected abnormal primary palates in two fetuses with nor-mal primary palates, these being the same 34 GW fetus referred to above and a 27-GW-old fetus that only had a cleft lip.

The secondary palate was normal in 58 fetuses. Of these, 54 (93%) were correctly identified by both readers when the scores of the levels 1 and 2 were combined. Reader 1 erroneously suspected an isolated cleft of the secondary palate (29 GW), and both readers suspected extensions to the sec-ondary palate in three fetuses with clefts only in-volving the lip and primary palate (23, 27 and 37 GW).

The nasal septum was considered normal (scor-ing levels 1 or 2) by both readers in all cases, at the exception of the two fetuses with confirmed unilateral primary palate clefts.

When summing the evaluations of both read-ers (i.e. 60x2 evaluations), incorrect suspicions of pathology existed in nine instances: four times a normal fetus was mislabeled as pathological and five times the degree of an existing cleft was over-rated. By anatomical structure, pathology was incorrectly presumed in 1.6% of the lips and pri-mary palates, and in 5.8% of secondary palates.

In all fetuses that had confirmed cleft lips or cleft palates, both readers assessed the existence of orofacial pathology to be likely or evident on MRI.

35

Results

Table 5. Scoring of readers 1 and 2 and interobserver agreement. Scoring levels: 1) evidently normal, 2) probably normal, 3) probably abnormal 4) evidently abnormal. True pathological cases in the series are marked with *. For the nasal septum, cases with ** represent unilateral palate clefts.

Upper lipReader 2 Score

1 2 3 4 Total1 43 8 0 0 512 0 2 0 0 23 1 0 0 *1 2

Reader 1

4 1 0 0 *4 5Total 45 10 0 5 60

Primary palateReader 2 Score

1 2 3 4 Total1 29 24 0 0 532 0 1 0 0 13 1 0 0 0 1

Reader 1

4 0 1 *1 *3 5Total 30 26 1 3 60

Secondary palateReader 2 Score

1 2 3 4 Total1 40 7 0 0 472 3 4 0 0 73 1 1 1 0 3

Reader 1

4 0 0 1 *2 3Total 44 12 2 2 60

Nasal SeptumReader 2 Score

1 2 3 4 Total1 50 2 0 0 522 2 4 0 0 63 0 0 0 0 0

Reader 1

4 0 0 0 **2 2Total 52 6 0 2 60

36


Vermian height vs. gestation weeks

017 18 19 20 21 22 23 24

5

10

15

mm

Gestation weeks

fVermis A-P diameter vs. gestation weeks

017 18 19 20 21 22 23 24

5

10

15

mm

Gestation weeks

e

Transverse cerebellar diameter vs. gestation weeks

15

17 18 19 20 21 22 23 24

20

25

30m

m

Gestation weeks

dCerebral biparietal diameter vs. gestation weeks

30

17 18 19 20 21 22 23 24

35

40

45

50

55

mm

Gestation weeks

c

Fronto-occipital diameter vs. gestation weeks

40

45

50

55

60

65

70

75

mm

Gestation weeks17 18 19 20 21 22 23 24

b

6 6 8

15 14

15

6 8

24

15

23

13

0

5

10

15

20

25

30

17 18 19 20 21 22 23 24

num

ber o

f fet

uses

Gestation weeks

Our cohort (n=70) Parazzini et al (n=84)

a

37

Discussion

DISCUSSION In the studies of this thesis we have investigated the normal development of several cranio-cere-bral structures that are not sufficiently covered in the fetal MRI literature, with special focus on the second trimester. The obtained results add to pre-existing knowledge and we believe they will be of benefit for evaluation of normality in these areas.

Our criteria to define “normality” and “normal” development must be clarified, since fetal MRI is a tertiary tool that is only performed when a disease is suspected. For the purpose of our pro-ject we used a normal postnatal clinical status for validation of the MRI examinations reported as “normal”, as it would not be ethical to perform post-natal control studies without any clinical in-dication.

Study I was the first published MRI series on fe-tal brain biometry that included fetuses below 20 GW. By using the same methodology as Parazzini et al 54, Tilea et al 55 and Garel et al 29, 46, 76, our work contributed to a potential larger biometrical database spanning the period from 17 GW to full term (fig. 16 a-f).

CBD and cerebral FOD are both intended for evaluation of the brain growth, using roughly perpendicular axes. The CBD is a standard ultra-sound measurement, but it is taken in a different manner than on MRI (axial plane, bone-to-bone), and therefore larger reference values are obtained 47, 48. The FOD is difficult to assess by ultrasound but obtainable at MRI. Previous studies demon-strated a linear increase in FOD at the second tri-mester 29, 54, 55 and we also found this pattern, even if with some scattering of the data.

The TCD is a standard measurement in fetal ul-trasound. As expected, our data concurred com-pletely with that method, since both US and MRI are able to evaluate this diameter in a precise way.

The vermis measurements showed the highest level of data dispersion in the series, which also was observed in Parazzini’s cohort 54. This sug-gests that MRI might not be accurate enough to evaluate vermian size at the 17-23 GW range.

Study II was the second in vivo fetal MRI study focused on the hippocampus 77 and the first in which gyral development in the hippocampal re-gion and the shape of the hippocampus was as-sessed. Our results show that there are wide individual variations, and a longer time for hip-pocampal closure than described before. We did also find some evidence that a non-oval shape of the hippocampus precedes a more mature oval shape and we demonstrated that there are signifi-cant differences in development between the left and right hippocampus.

The hippocampal region was well assessable in all post mortem examinations, but in vivo the image quality was much lower. In fact, 59% of the examinations available in the database were rejected for technical reasons. The youngest liv-ing fetus with a resolution good enough for hip-pocampal evaluation was examined at 19 GW.

The hippocampal sulcus was identifiable in all of the fetuses in our series, earlier than described by Garel et al in 2003 76, a feature that might be related to the more recent imaging techniques that we used.

Concerning hippocampal shape, previous stud-ies on formalin-fixed brain indicate that the hip-pocampus should assume an adult appearance by 21 GW 56, 67, 68, but we found a closed hippocam-pal sulcus only in one fetus at that age. Open hip-pocampal sulci could still be seen at 32 GW, thus reflecting a larger time variation for closure than previously thought.

When a closed hippocampal sulcus was seen, its

Figure 16 (see left).a) Distribution of the fetuses by GW, compared with the series by Parazzini et al 54. b–f) Median (dots), maximum, and minimum (hori-zontal bars) values for each measurement per GW are represented in full. The corresponding data by Parazzini et al 54 are represented alongside, dashed line.

38


shape was classified as oval or non-oval, using the same criteria as in earlier studies on adults, chil-dren and premature neonates 72, 74, 75. As almost all the fetuses in our series with an oval hippocam-pus were at least 27 GW old and bilateral non-oval hippocampi were not observed after 29 GW, our data agree with the hypothesis that the non-oval shape precedes the oval shape in the hippocampal development.

The development of the hippocampus, includ-ing the hippocampal sulcus and the inverted hip-pocampal formation, was often asymmetric in our series (26/63 = 41%) and in a vast majority of these, the right side (23/26 = 88%) developed faster. This asymmetry was not observed in a previous prena-tal MR study, where hippocampal infolding an-gles were measured at 20-37 GW 77, but another group that used the same methodology in infants and children found a left to right asymmetry and concluded for a slower development of the left hip-pocampus 106. The work by Chi et al, who evaluat-ed photographs of 507 formalin-fixed brains from 10 to 44 GW also suggest a faster gyral develop-ment of the right cerebral hemisphere 107, the same being reported at a recent volumetric MR study that included the hippocampal sulcus 108.

The collateral sulcus also had a tendency to de-velop earlier on the right side and it was identifi-able as early as at GW 17, even if still not visible in one fetus at GW 29. A vertical orientation was seen uni- or bilaterally in 58% of the fetuses in our cohort that presented a well-defined sulcus, a slightly larger number (p = 0.045) than the 36% de-scribed for children or adult populations 75. Also, this vertical orientation occurred more commonly on the left side, thus suggesting the possibility of a transitional phase towards a more mature hori-zontally oriented axis. However, this proposed hy-pothesis could not be confirmed within our own cohort, as we failed to find any significant associa-tion between the gestational age and the percent-age of vertical collateral sulci.

Our study is limited by the small number of the fetuses, even if larger than earlier hippocampal studies on formalin-fixed brains. Repeated exami-nations of healthy fetuses would provide a better

evaluation of the normal development, but this is not ethically acceptable.

Study III was the first publication on the develop-ment of the complete fetal ear in vivo. This study suffered from the expected difficulties when evaluating a very small organ, but it was possible to depict the components of the ear in more than 70% of the fetuses. However, by introducing a cut-off at the end of the second trimester, there was a significant difference in the scoring levels before and after that age, suggesting that MRI findings of the ear should be taken very cautiously before the third trimester.

The development of the ear begins at 4 GW 109

and the labyrinth reaches a final adult shape and size by approximately 11 GW 110. The ossicles ma-ture later, until 15-20 GW 110, and the tympanic cavity continues to enlarge until approximately GW 37 111. The external auditory canal only at-tains its final shape late in childhood 112.

Delineation of the small ear structures with fetal MRI is limited by movement, slice tilting and also by partial-volume averaging. Therefore, in order to better understand these limitations we decided to start our project by retrospectively analyzing a series of post mortem MRI examinations, to serve as a standard of what was possible to see in the ear using thin slices, absence of fetal motion and unlimited scan time.

The results of this post mortem group fully agree with the literature 110, 113, with a near-perfect imag-ing quality being achieved for all organs. Howev-er, we encountered some difficulties upon assess-ing the lateral semicircular canals on axial views, as did Nemzek et al in their post mortem series 110. This might be related to the off-plane orientation of these brain-oriented axial slices, and also to a lack of T2 contrast with the surrounding non-os-sified mesenchyme at these early ages.

In the in vivo series, the symmetry in scoring for the inner and middle ear was always above 93%, attesting for an absence of significant slice tilting at the MRI studies we evaluated, but an identifica-tion of structures at a similar level as in the post mortem studies was never reached.

39

Discussion

The cochlea is known to be more prone to devel-opmental malformations than the phylogenetically older vestibule 114. Severe cochlear malformations are rare, but incomplete cochlear partitions are frequent 115, being the second most common cause of congenital deafness after Sheibe’s membranous dysplasia 116. In our series it was possible to obtain a detail that would be enough to exclude severe cochlear malformations in more than 70% of the fetuses, but a sufficient detail to rule out a simpli-fication of the cochlea to 1.5 turns as described by Mondini (quoted in 114) could only be achieved in two cases (1.6%). To our surprise, the ability to identify the inner structure of the cochlea im-proved with gestational age, a feature we believe is related to a progressive gain of T2 contrast against the surrounding otic capsule as it ossifies.

The semicircular canals and remaining vesti-bule are also formed before 11 GW, with the ex-ception of the vestibular sac and aqueduct that only reach full development in adolescence 114. The lateral semicircular canals are the last ones to develop 117 and they are particularly prone to malformations 112. Indeed, it is very unusual for a malformation of the vestibular complex not to in-clude the LSC 114, and we postulate that their nor-mality might be used as a hallmark for a normal vestibule as a whole. We were able to identify the LSC in more than 70% of the examinations in our cohort, but to see its entire contour was only pos-sible in 24%, probably for the same reasons that disturbed its evaluation at the postmortem series. Again, the delineation of the LSC was only pos-sible in 43% (18/42) of the studies before 25 GW, and we suppose this was related to the same lack of T2 contrast against the surrounding mesen-chyme that affected the evaluation of the cochlea in young fetuses.

The visualization of the middle ear with fetal MRI has been the subject of a previous report 91 concerning fetuses older than 25 GW. In that study the authors reported visualization of ossicles in 100% of the cases, but never with enough detail to discriminate each ossicle. In our study we were not able to identify ossicles in 30% of the fetuses, but most of these were younger than 25 GW. If

we only take into account fetuses above 25 GW, as in the study referred above, ossicular identifi-cation was possible in 90% of the cases. In one fetus, however, the ossicles could not be identified as late as 33 GW.

The external auditory canals are normally only patent after 28 GW, after the progressive involu-tion of an ectodermal plug that occludes them un-til then 118. This absence of fluid inside the EAC surely contributed to the difficulty we had upon identifying this structure in our cohort, more of-ten before 29 GW. Even afterwards, it could only be identified bilaterally with normal characteris-tics in 59% of the fetuses, thus suggesting MRI to be a suboptimal method for EAC evaluation before birth.

Overall, our goal to provide an insight into the development of the ear using in vivo fetal MRI proved to be a difficult task, which is reflected in the degree of interobserver agreement. The concordance of only 78% on average probably reflected our decision to use a rating system for structures that are at the limit of the spatial resolu-tion of MRI, but blinding the second reviewer to the clinical outcome of the fetuses may also have been an important factor influencing the disagree-ment. Indeed, the main reader knew that a normal hearing of the subjects was part of the inclusion criteria, therefore introducing a reading bias due to expectation of normality.

We also realize that partial volume averaging did play a role in our results, as the 3-4 mm thick slices that are used in brain fetal MR studies rep-resent a technical limitation that contributed to a suboptimal delineation of the very small struc-tures we tried to evaluate.

Study IV is, to our knowledge the first MRI study to evaluate the lip and palate through all gesta-tional ages, blinded to the results of previous ul-trasound examinations. Our results support MRI as a reliable method for independent evaluation of these structures. One previous MRI report ex-ists on the evaluation of the normal fetal profile including the lip and palate, but this report only included cases at 16 to 26 GW 119.

40


The routine mid-trimester fetal ultrasound ex-amination that is performed in most developed countries is accepted to have a sensitivity of 45 - 68% for the detection of cleft lips and a very low detection rate for isolated clefts of the palate 93, 120. As it is then possible for facial anomalies to be missed on ultrasound and be later found on a fe-tal MRI that is referred for other purposes 103 it is important to have a notion of how well MRI inde-pendently evaluates the lip and palate on standard brain-targeted scans, out of its usual complemen-tary role towards ultrasound.

In our study, it was possible to correctly delin-eate these structures in almost all of the fetuses with a “substantial” 121 level of interobserver agreement, regardless of gestational age. We used quadratic weighted kappa as opposed to linear kappa to measure interobserver agreement for the four-grade scale, since the different levels in our scale did not carry a linear difference. Both read-ers correctly identified all the fetuses with clefts included in the series, but pathology was overdi-agnosed.

The formation of the upper lip is completed at 7 GW, after the fusion of the paired medial nasal prominences and maxillary prominences. These medial nasal prominences also merge at a deeper level, forming the intermaxillary segment that includes the four incisor teeth and the triangle-shaped primary palate 122. The secondary palate is formed independently from two outgrowths of the maxillary prominences, the palatine shelves. The maxillary prominences also form the remaining of the upper alveolar ridge 122 and the soft palate. Disturbances in the process may produce clefts, but these do not seem however to result merely from a mechanical mismatch in coupling, and lo-cal mesoderm penetration failure or a disturbance in migration /organization of neuroepithelial cells have been proposed as potential contributing fac-tors 123. Isolated clefts of the secondary palate might be distinct embryological entities 122, as these are more common in females and not influ-enced by maternal age, as opposed to cleft lip / primary palate.

Visualizing the upper lip is a part of the check-

list of the standard mid-trimester sonography 26. However, fetal age and motion, covering of the face with the hands or arms, maternal constitu-tion or equipment-related limitations can preclude a correct identification of this area, thus justify-ing the relatively low accuracy 2D ultrasound has for detecting lip malformations. False positives, on the other hand, are extremely rare in low-risk populations when using this technique, without any single case being found by Offerdal et al 93 in a prospective series of 49 341 fetuses.

MRI examinations are likewise limited by a number of fetal and mother-related factors that may disturb the evaluation of the fetal facial pro-file. The lips may be obscured by the contact with the placenta 119 or by superposing the arms and, in fact, it was a contact of the fetal face with the arm the cause of one falsely abnormal lip suspected by Reader 1 in our series (fig. 17). This represents a pitfall of MRI by not being a method as interac-tive as ultrasound, where a delay of just a few sec-onds may help to clarify any posture-related false images.

On the contrary to the facial profile, both the primary and secondary palate of fetuses are very difficult to investigate with ultrasound, especially if there is no suspicion of an associated cleft lip 92. Even in its presence, a 25% error can be expect-ed to exist when evaluating the degree of palatal involvement by this method 124. The detection of isolated palate abnormalities by ultrasound can be as low as 0 to 7% 92, but the accuracy increas-es for high-risk populations seen at tertiary care units 92, or when volume acquisitions for the face are used 125.

MRI allows a clearer visualization of the palate (fig. 18), but fetal motion, young fetal age, or the absence of amniotic fluid between the palate and the tongue 94, 119 may limit its accuracy. This lack of T2 contrast between the tongue and the palate has been held responsible for a 16% failure rate for palate delineation in a previous series 119 and the same probably explained the incorrect suspicions both readers in our work had of posterior exten-sions of primary palate clefts to the secondary pal-ate (fig. 19).

41

Discussion

Figure 17.The arm of this 36 GW fetus is seen touching the lip and hiding its contours. Lip pathology was erroneously suspected (scoring level 3) by one reader.

Figure 18. Normal anatomy of the palate and lips. Fetus of 27 GW. a) Sagittal plane. The posterior

limit of the primary palate (arrow), the secondary palate (short dashed arrow) and upper alveolar ridge (long dashed ar-row) are pointed.

b) Axial plane. The bone of the hard palate (dashed arrow) is seen, with a lower signal than the posteriorly located soft pal-ate (arrowhead).

Figure 19. A cleft primary palate can be seen (arrow). The absence of fluid between the tongue and palate - compare with figure 17 - lead both readers to a false suspicion of a cleft secondary palate in this 37 GW fetus (scoring level 3).

42


It is known 126 that the nasal septum tends to be deviated to the homolateral side when a unilateral cleft palate is present. In fact, in our series devi-ated septa were only found in such instances, thus suggesting this finding to be a valuable indirect sign of an unilateral cleft palate.

Our study was intended to be descriptive, with the aim to assess how well experienced neurora-diologists could delineate the normal anatomy of the lip and palate on fetal MRI, blinded to pre-vious ultrasound findings. The detection of the abnormal fetuses and their assumptive diagnosis was only a secondary aim, because the amount of pathological cases in our database was too small to draw conclusions about this group. Further-

more, a selection bias for this particular material could not be excluded, since all of them had been earlier suspected on US and diagnosed on MRI. Our reason to include these pathological cases at the series was mainly to induce some incertitude in the readers and to avoid a potential bias towards expecting normality, as it is known that only 1/700 fetuses have malformations in this area. However, a reading bias in the reverse sense (i.e. towards over-scoring pathology) might have influenced the results, since the readers were asked to focus on the palate and lips, which is not a special area of attention when performing brain fetal MRI in daily practice.

43

Conclusions and Future Directions

CONCLUSIONS This MRI project focused on the normal development of the fetal brain and skull.

We provided:

a) New MRI reference data for brain measurements in young fetuses.

b) New data on the fetal hippocampus, that seems to develop later and more asymmetrically than previ-ously thought.

c) The first research on the development of the ear in vivo.

d) The notion that brain-targeted MRI is reliable for independent evaluation of the lip and palate across gestation.

FUTURE DIRECTIONS I do not dare to predict the future of fetal MRI (or of any other medical area, for that matter…). However, it seems clear that the method is nowadays waiting for “ethical clearance” to move into higher field strengths that may provide a higher spatial resolution and a better access to more “physiological” approaches to the fetal development, as functional MRI or MR spectroscopy.

At the pace permitted by electronics and software, the quality of imaging will surely improve in the future, including methods to overcome the limitations that result from fetal and maternal motion.

But MRI is just a tool, and it needs skilful operators in order to be useful. Therefore, I hope for the future development of a broader network of knowledge in fetal MRI, spanning as many centres as possible, and also for an international “teaching core” for the medical training in this field.

For myself, I would like to know more about the functional aspects of the fetal development. Although probably difficult in practice, research on “resting-state” functional MRI seems to be worth a try, as has been recently shown by the Vienna group 127.

We shall see….

44


RESUMO EM PORTUGUÊS A ultrassonografia é o método imagiológico pa-drão para estudo fetal, mas a RM tem hoje um pa-pel complementar importante, uma vez que mui-tas vezes revela alterações que alteram a gestão da gravidez de um modo significativo.

A RM tem sido também usada para estudar o desenvolvimento fetal normal, mas há falta de conhecimento em algumas áreas clinicamente relevantes, tais como padrões para as dimensões normais do cérebro ao longo da gestação, o de-senvolvimento normal do hipocampo ou do crânio fetais.

O objetivo desta tese foi colmatar algumas des-sas falhas, focando no período antes do fim do segundo trimestre, prazo limite para interrupção de gravidez por causa fetal em muitos países (in-cluindo Portugal). Procurámos a obtenção de da-dos biométricos para normalidade antes de 24 se-manas de gestação, estudámos o desenvolvimento do hipocampo e do ouvido fetais, e tentámos ain-da avaliar o valor da ressonância magnética para avaliar o lábio e palato fetais.

Para isso, foi analisada retrospectivamente uma base de dados com aproximadamente 450 RM fe-tais in vivo e 21 post mortem.

O estudo I avaliou uma série de 70 exames de RM fetais in vivo normais. Uma tabela de medidas normais do cérebro entre as 17e 23 semanas de gestação foi construída, a primeira na literatura que inclui idades abaixo de 20 semanas de gesta-ção.

O estudo II centrou-se na evolução do hipocam-po a partir das 18 semanas de gestação, avaliando 3 RM post mortem e 60 in vivo. Este estudo sugere que o desenvolvimento do hipocampo fetal se faz mais tarde e de um modo mais assimétrico do que previamente descrito.

No estudo III foi analisada uma série de 122 RM fetais in vivo normais e 16 post mortem. Descre-vemos pela primeira vez na literatura o desenvol-vimento do ouvido fetal in vivo, dando-nos conta que a RM fetal é limitada pelo tamanho das estru-turas avaliadas, sobretudo até ao fim do segundo trimestre de gestação.

No estudo IV, os lábios e palatos de 55 fetos nor-mais e 5 fetos com fendas orofaciais foram anali-sados de modo cego por dois revisores. Os nossos resultados sugerem uma alta precisão da RM na avaliação desta área, independentemente da idade fetal ou de dados ultrassonográficos anterior.

Este trabalho traz portanto novos conhecimen-tos sobre o desenvolvimento normal do cérebro e crânio fetais.

SAMMANFATTNING PÅ SVENSKA Ultraljudsundersökning är första-handsmetod för fosteravbildning men magnetisk resonanstomo-grafi (MRT) kan vara ett värdefullt komplement. MRT kan ofta ge tilläggsinformation som kan på-verka rådgivning och handläggning och har gett ökad förståelse för hjärnans normala utveckling. Inom vissa kliniskt relevanta områden är kunska-pen dock bristfällig, t.ex. avseende normalinter-vall för olika hjärnmått, hippokampus utveckling och bedömningen av öra och ansikte. Syftet med studierna i denna avhandling var därför att med MRT samla normaldata för olika hjärnmått i tidig graviditet, studera utvecklingen av hippokampus, undersöka hur väl olika strukturer i fosterörat kan bedömas med MRT och undersöka hur väl fost-rets överläpp och gom kan bedömas med MRT.

För att rätt kunna bedöma resultatet av en MR-undersökning av fosterhjärnan är det viktigt att känna till normalvariationen. Referensdata för olika mått inom hjärnan finns sedan tidigare för ultraljud. En del av de mått som kan göras med MRT dock kan inte utföras med ultraljud och vissa mått mäts på olika sätt med ultraljud och MRT. Separata referensmaterial för MRT behövs därför. För MRT finns endast ett fåtal studier med normaldata publicerade och endast efter gravidi-tetsvecka 19. I 70 MR-undersökningar av foster vid 17-23 graviditetssveckor uppmättes olika mått i storhjärna och lillhjärna. Alla mätningar kunde

45

Summary

göras. En tabell med normalintervall från gravi-ditetsvecka 17 till 23 skapades. För de av mått-ten som kan mätas med ultraljud överensstämmer resultaten väl med ultraljudsreferensdata. För de graviditetsveckor där det finns tidigare MR-stu-dier överensstämmer resultaten väl med dessa.

Hippokampus är en struktur i tinninglobens inre del som tidigare studier av vuxna visat ofta ha asymmetrisk form. In vivo-studier av hippo-kampus form hos foster med MRT saknas. I 60 normala in vivo och 3 post mortem MR-under-sökningar studerades förekomst och slutning av sulcus hippocampalis, hippokampus form och förekomst och orientering av sulcus collateralis. Sulcus hippocampalis sågs i alla foster. Den var oftast sluten efter 28 gestationsveckor och aldrig öppen efter 32 gestationsveckor. Icke-ovala hip-pokampusformer sågs bilateralt vid 24-29 gesta-tionsveckor och unilateralt fortfarande vid 31-36 graviditetsveckor. Asymmetrisk hippokampusut-veckling sågs i 41% av fostren och snabbare på höger sida i 88% av dem.

Missbildningar inom inneröra, mellanöra eller hörselgång är inte ovanliga och kan ibland vara del i ett missbildningssyndrom. Bedömningen av dessa strukturer med ultraljud är svår. Det har inte systematiskt utvärderats hur väl man kan be-döma fosterörat vid olika graviditetslängder med MRT. Bedömbarhet med MRT in vivo av snäcka, båggångsapparat, hörselben och yttre hörselgång studerades hos 122 foster vid 20-38 graviditets-veckor. Samtliga hade normal postnatal uppfölj-ning. Samma strukturer studerades dessutom hos 16 foster undersökta post mortem vid 15-22 gra-viditetsveckor. Snäckans vindlingar, laterala båg-gången och hörselbenen sågs med MRT in vivo i

50/30/33% och 89/93/90% före respektive efter 25 graviditetsveckor. Skillnaden i bedömning av ör-ats strukturer före resp. efter 25 graviditetsveckor tolkas bero på ökad kontrast mellan olika struktu-rer med ökande grad av förbening. Med undantag av yttre hörselgången kunde alla studerade struk-turer i örat avgränsas i 100% av post mortem-un-dersökningarna, där detaljupplösningen var bättre pga. att undersökningstiden inte var begränsad. Detta talar för att bedömningen av fosterörat med MRT in vivo kan förbättras om tekniken utveck-las.

Läpp-käk-gomspalt (LKG-spalt) ses i 1/700 föds-lar och kan vara svår att diagnosticera intrauterint med ultraljud, särskilt vid isolerad gomspalt. Det har inte tidigare studerats hur väl överläpp och gom kan bedömas med MRT vid olika graviditets-längder. Överläpp och gom studerades av två obe-roende undersökare i 60 MR-undersökningar vid 20-38 gestationsveckor. Bedömare 1 och 2 iden-tifierade 100% av fall med LKG-spalt, 96-100% av normala överläppar och primära gomanlag och 93-100% av normala sekundära gomanlag. Be-dömbarheten skiljde sig inte signifikant med gra-viditetslängden. Resultaten tyder på att överläpp och gom är väl bedömbara med MRT, oberoende av graviditetslängd. Bedömning av dessa struktu-rer bör därför ingå som rutin vid MRT av foster och kan leda till upptäckt av annars odiagnostice-rade LKG-spalter.

Avhandlingen tillför: referensvärden för hjärn-mått i tidig graviditet med MRT, kunskap om ut-vecklingen av hippokampus och kunskap om be-dömbarheten av fosteröra, överläpp och gom med MRT vid olika graviditetslängder.

46


ACKNOWLEDGEMENTS It is common statement that a PhD reflects the work of many people, but in the particular case of this mid-dle-aged man with a large family and several commitments in two countries, this becomes especially true!

There were many “actors” in the project, but an even larger number of people stayed in the backstage, sweeping out the difficulties and providing the environment that made it all possible. My sincere gratitude to you all.

The “actors”:I want to thank first of all my tutors Johan Wikström and Raili Raininko. As my scientific ”parents”

they have been omnipresent, friendly and caring, helping me to find the right pathway along the process.

João Teixeira, co-author in most of the papers, has been my brother-in-arms for almost two decades. His profound knowledge of Neuroradiology, gentle personality and true friendship have been pillars of the whole project.

I want to thank my co-author and friend Dragan Bajic for his leadership in the paper concerning the fetal hippocampus. And also Valentina Ribeiro, Hashem Amini, Ove Axelsson and Raquel Themudo for their precious input as co-authors in papers of this thesis.

All the MR examinations were performed by devoted (and very patient!) technicians: in particular, I want to thank Julio Dias, Pedro Viegas and Andreia Monteiro for their extensive collaboration in hun-dreds of studies at Campos Costa clinic.

I thank Christl Richter-Frohm, our precious link with the University. Christl has helped in many steps of the PhD process, from courses to paperwork, always with a kind attitude and a smile.

The words you are reading at this moment reflect the work by Håkan Pettersson and Nora Velastegui, the gentle “wonder-couple” that patiently edited the text and layout into its final version. Thank you!

Last but not least, I want to thank my friend Per Liss, the former Head of the Department, for his in-vitation to collaborate with the Uppsala University Hospital in 2007, and for challenging me to start this project.

47

Acknowledgements

The “backstage” (summary of a long list…):I thank my family: Catarina, Jaime, Zé, Luzia, Ana Clara, Isilda, André, Clare, my Parents and also Isabel for their continuous love, support and for keeping the family “afloat and on due course” all these years. I love you with all my heart.

I thank all my co-workers at Hospital S. Teotónio (Viseu), University Hospital (Uppsala), Consultório Dr Campos Costa (Porto), UBI Medical School (Covilhã), SIMX (Viseu), CEDIR (Guarda) and Cerdeira Guerra (Lamego). The boards, doctors, technicians, nurses, co-teachers and support staff of these Institu-tions played an important role in this project, by being supportive, patient and coping with all my absences for research. I thank you immensely.

I thank my friends both in Portugal and Sweden, for honouring me with their presence in my life. Every moment spent with each one of them (a glass of wine, a chat, a good laughter or just sharing a bad day…) is – and will be - meaningful and worthwhile. In particular, I thank Helen, Paul, Sam and Eddie for pro-viding me with a “home away from home” at their house for such a long period this last year.

I thank Romeu Cruz and Charles Raybaud, my first medical role models, for teaching me how to be a Doctor and a Neuroradiologist.

A final thought for two important people I never had the pleasure of meeting: Steve Jobs, the creator of my friendly and trustworthy “Mac” and Antoine Rosset, the creator of OSIRIX, a wonderful (free!) com-puter tool for the post-processing of medical images. Thank you, wherever you are.

Obrigado, do fundo da alma!

48


49

References

REFERENCES 1. Clementi M, Stoll C. The Euroscan study. Ultra-

sound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology 2001;18:297-300

2. Grandjean H, Larroque D, Levi S. Sensitivity of routine ultrasound screening of pregnancies in the Eurofetus database. The Eurofetus Team. Ann N Y Acad Sci 1998;847:118-124

3. Sonographic examination of the fetal central ner-vous system: guidelines for performing the ‘ba-sic examination’ and the ‘fetal neurosonogram’. Ultrasound Obstet Gynecol 2007;29:109-116

4. Malinger G, Ben-Sira L, Lev D, et al. Fetal brain imaging: a comparison between magnetic resonance imaging and dedicated neurosonog-raphy. Ultrasound in obstetrics & gynecology 2004;23:333-340

5. Malinger G, Lev D, Lerman-Sagie T. Fetal cen-tral nervous system: MR imaging versus dedicat-ed US:need for prospective, blind, comparative studies. Radiology 2004;232:306; author reply 306-307

6. Pistorius LR, Hellmann PM, Visser GH, et al. Fetal neuroimaging: ultrasound, MRI, or both? Obstet Gynecol Surv 2008;63:733-745

7. Levine D, Barnes PD, Madsen JR, et al. Fetal central nervous system anomalies: MR imag-ing augments sonographic diagnosis. Radiology 1997;204:635-642

8. Levine D, Barnes PD, Robertson RR, et al. Fast MR imaging of fetal central nervous system ab-normalities. Radiology 2003;229:51-61

9. Twickler DM, Magee KP, Caire J, et al. Second-opinion magnetic resonance imaging for suspect-ed fetal central nervous system abnormalities. American journal of obstetrics and gynecology 2003;188:492-496

10. Whitby EH, Paley MN, Sprigg A, et al. Compari-son of ultrasound and magnetic resonance imag-ing in 100 singleton pregnancies with suspected brain abnormalities. BJOG 2004;111:784-792

11. Griffiths PD, Reeves MJ, Morris JE, et al. A pro-spective study of fetuses with isolated ventricu-lomegaly investigated by antenatal sonography and in utero MR imaging. AJNR American jour-nal of neuroradiology 2010;31:106-111

12. Smith FW, Adam AH, Phillips WD. NMR imag-ing in pregnancy. Lancet 1983;1:61-62

13. Smith FW. The potential use of nuclear magnetic resonance imaging in pregnancy. J Perinat Med 1985;13:265-276

14. McCarthy SM, Filly RA, Stark DD, et al. Obstet-rical magnetic resonance imaging: fetal anatomy. Radiology 1985;154:427-432

15. Baker PN, Johnson IR, Gowland PA, et al. Fetal weight estimation by echo-planar magnetic reso-nance imaging. Lancet 1994;343:644-645

16. Baker PN, Johnson IR, Gowland PA, et al. Es-timation of fetal lung volume using echo-planar magnetic resonance imaging. Obstet Gynecol 1994;83:951-954

17. Daffos F, Forestier F, Mac Aleese J, et al. Fetal curarization for prenatal magnetic resonance im-aging. Prenat Diagn 1988;8:312-314

18. Girard N, Raybaud C, Dercole C, et al. In vivo MRI of the fetal brain. Neuroradiology 1993;35:431-436

19. Levine D, Barnes PD, Edelman RR. Obstetric MR imaging. Radiology 1999;211:609-617

20. Simon EM, Goldstein RB, Coakley FV, et al. Fast MR imaging of fetal CNS anomalies in ute-ro. AJNR Am J Neuroradiol 2000;21:1688-1698

21. Malamateniou C, McGuinness AK, Allsop JM, et al. Snapshot inversion recovery: an optimized single-shot T1-weighted inversion-recovery se-quence for improved fetal brain anatomic delin-eation. Radiology 2011;258:229-235

22. Prayer D, Brugger PC, Asenbaum U. Indications for Fetal MRI. In: Prayer D, ed.: Springer Berlin Heidelberg; 2011:1-17

23. Glenn OA. MR imaging of the fetal brain. Pedi-atric Radiology 2010;40:68-81

24. Triulzi F, Manganaro L, Volpe P. Fetal magnetic resonance imaging: indications, study protocols and safety. Radiol Med 2011;116:337-350

25. International Planned Parenthood F. Abortion Legislation In Europe. 2009

26. AIUM practice guideline for the performance of obstetric ultrasound examinations. J Ultrasound Med 2010;29:157-166

27. Malinger G, Lev D, Lerman-Sagie T. Normal and abnormal fetal brain development during the third trimester as demonstrated by neurosonogra-phy. European journal of radiology 2006;57:226-232

50


28. Coakley FV, Glenn OA, Qayyum A, et al. Fetal MRI: a developing technique for the developing patient. AJR Am J Roentgenol 2004;182:243-252

29. Garel C. The role of MRI in the evaluation of the fetal brain with an emphasis on biom-etry, gyration and parenchyma. Pediatr Radiol 2004;34:694-699

30. Ehrstrom C. Circadian rhythm of fetal move-ments. Acta Obstet Gynecol Scand 1984;63:539-541

31. Clements H, Duncan KR, Fielding K, et al. In-fants exposed to MRI in utero have a normal pae-diatric assessment at 9 months of age. The British journal of radiology 2000;73:190-194

32. Kok RD, de Vries MM, Heerschap A, et al. Ab-sence of harmful effects of magnetic resonance exposure at 1.5 T in utero during the third trimes-ter of pregnancy: a follow-up study. Magnetic resonance imaging 2004;22:851-854

33. Heinrichs WL, Fong P, Flannery M, et al. Midg-estational exposure of pregnant BALB/c mice to magnetic resonance imaging conditions. Magn Reson Imaging 1988;6:305-313

34. Chen MM, Coakley FV, Kaimal A, et al. Guide-lines for computed tomography and magnetic resonance imaging use during pregnancy and lactation. Obstet Gynecol 2008;112:333-340

35. Prayer D, Kasprian G, Krampl E, et al. MRI of normal fetal brain development. Eur J Radiol 2006;57:199-216

36. Baker PN, Johnson IR, Harvey PR, et al. A three-year follow-up of children imaged in utero with echo-planar magnetic resonance. Am J Obstet Gynecol 1994;170:32-33

37. Reeves MJ, Brandreth M, Whitby EH, et al. Neo-natal cochlear function: measurement after expo-sure to acoustic noise during in utero MR imag-ing. Radiology 2010;257:802-809

38. Edwards MJ. Review: Hyperthermia and fever during pregnancy. Birth Defects Res A Clin Mol Teratol 2006;76:507-516

39. Dimbylow PJ, Nagaoka T, Xu XG. A compari-son of foetal SAR in three sets of pregnant fe-male models. Physics in medicine and biology 2009;54:2755-2767

40. Hand JW, Li Y, Hajnal JV. Numerical study of RF exposure and the resulting temperature rise in the foetus during a magnetic resonance proce-dure. Phys Med Biol 2010;55:913-930

41. Gowland PA, De Wilde J. Temperature increase in the fetus due to radio frequency exposure dur-ing magnetic resonance scanning. Phys Med Biol 2008;53:L15-18

42. Brugger PC, Prayer D. Actual imaging time in fetal MRI. Eur J Radiol 2011

43. Merkle EM, Dale BM, Paulson EK. Abdominal MR Imaging at 3T. Magnetic resonance imaging clinics of North America 2006;14:17-26

44. Feess-Higgins A, Larroche J-C. Le développe-ment du cerveau foetal humain : atlas anatomique = Development of the human foetal brain : an anatomical atlas. Paris: INSERM CNRS ; Mas-son; 1987

45. Toi A, Lister WS, Fong KW. How early are fetal cerebral sulci visible at prenatal ultrasound and what is the normal pattern of early fetal sulcal development? Ultrasound in Obstetrics & Gyne-cology 2004;24:706-715

46. Garel C. Le développement du cerveau foetal : atlas IRM et biométrie. Montpellier: Sauramps médical; 2000

47. Pilu G, Nicolaides KH. Diagnosis of fetal abnor-malities : the 18-23-week scan. Diploma in fetal medicine series, 1999:viii, 149 p.

48. Timor-Tritsch IE, Monteagudo A, Cohen HL. Ul-trasonography of the prenatal and neonatal brain. 2001:xix, 550 p.

49. Grossman R, Hoffman C, Mardor Y, et al. Quan-titative MRI measurements of human fetal brain development in utero. Neuroimage 2006;33:463-470

50. Gholipour A, Estroff JA, Barnewolt CE, et al. Fetal brain volumetry through MRI volumetric reconstruction and segmentation. Int J Comput Assist Radiol Surg 2011;6:329-339

51. Scott JA, Habas PA, Kim K, et al. Growth trajec-tories of the human fetal brain tissues estimated from 3D reconstructed in utero MRI. Int J Dev Neurosci 2011;29:529-536

52. Clouchoux C, Guizard N, Evans AC, et al. Nor-mative fetal brain growth by quantitative in vivo magnetic resonance imaging. Am J Obstet Gyne-col 2011

53. Reichel TF, Ramus RM, Caire JT, et al. Fetal central nervous system biometry on MR imag-ing. AJR Am J Roentgenol 2003;180:1155-1158

54. Parazzini C, Righini A, Rustico M, et al. Pre-natal magnetic resonance imaging: brain nor-mal linear biometric values below 24 gestational weeks. Neuroradiology 2008;50:877-883

51

References

55. Tilea B, Alberti C, Adamsbaum C, et al. Cerebral biometry in fetal magnetic resonance imaging: new reference data. Ultrasound Obstet Gynecol 2009;33:173-181

56. Rados M, Judas M, Kostovic I. In vitro MRI of brain development. European journal of radiol-ogy 2006;57:187-198

57. Kostovic I, Judas M, Rados M, et al. Laminar or-ganization of the human fetal cerebrum revealed by histochemical markers and magnetic reso-nance imaging. Cereb Cortex 2002;12:536-544

58. Levitt P. Structural and functional matura-tion of the developing primate brain. J Pediatr 2003;143:S35-45

59. Brazel CY, Romanko MJ, Rothstein RP, et al. Roles of the mammalian subventricular zone in brain development. Prog Neurobiol 2003;69:49-69

60. Fogliarini Cl, Chaumoitre K, Chapon Fdr, et al. Assessment of cortical maturation with prenatal MRI. Part I: normal cortical maturation. Euro-pean Radiology 2005;15:1671-1685

61. Kostovic I, Rakic P. Developmental history of the transient subplate zone in the visual and so-matosensory cortex of the macaque monkey and human brain. The Journal of comparative neu-rology 1990;297:441-470

62. Judas M, Rados M, Jovanov-Milosevic N, et al. Structural, immunocytochemical, and mr imag-ing properties of periventricular crossroads of growing cortical pathways in preterm infants. AJNR American journal of neuroradiology 2005;26:2671-2684

63. Barkovich AJ, Kjos BO, Jackson DE, Jr., et al. Normal maturation of the neonatal and in-fant brain: MR imaging at 1.5 T. Radiology 1988;166:173-180

64. Counsell SJ, Maalouf EF, Fletcher AM, et al. MR imaging assessment of myelination in the very preterm brain. AJNR American journal of neuro-radiology 2002;23:872-881

65. Limperopoulos C, Robertson RL, Khwaja OS, et al. How Accurately Does Current Fetal Imaging Identify Posterior Fossa Anomalies? American Journal of Roentgenology 2008;190:1637-1643

66. Limperopoulos C, Robertson RL, Estroff JA, et al. Diagnosis of inferior vermian hypoplasia by fetal magnetic resonance imaging: poten-tial pitfalls and neurodevelopmental outcome. American journal of obstetrics and gynecology 2006;194:1070-1076

67. Kier EL, Fulbright RK, Bronen RA. Limbic lobe embryology and anatomy: dissection and MR of the medial surface of the fetal cerebral hemi-sphere. AJNR American journal of neuroradiol-ogy 1995;16:1847-1853

68. Kier EL, Kim JH, Fulbright RK, et al. Embry-ology of the human fetal hippocampus: MR im-aging, anatomy, and histology. AJNR American journal of neuroradiology 1997;18:525-532

69. Humphrey T. The development of the hu-man hippocampal fissure. Journal of anatomy 1967;101:655-676

70. Arnold SE, Trojanowski JQ. Human fetal hip-pocampal development: I. Cytoarchitecture, myeloarchitecture, and neuronal morphologic features. The Journal of comparative neurology 1996;367:274-292

71. Arnold SE, Trojanowski JQ. Human fetal hip-pocampal development: II. The neuronal cyto-skeleton. The Journal of comparative neurology 1996;367:293-307

72. Bajic D, Ewald U, Raininko R. Hippocampal de-velopment at gestation weeks 23 to 36. An ultra-sound study on preterm neonates. Neuroradiol-ogy 2010;52:489-494

73. Baker LL, Barkovich AJ. The large temporal horn: MR analysis in developmental brain anom-alies versus hydrocephalus. AJNR American journal of neuroradiology 1992;13:115-122

74. Bajic D, Wang C, Kumlien E, et al. Incom-plete inversion of the hippocampus:a common developmental anomaly. European Radiology 2008;18:138-142

75. Bajic D, Kumlien E, Mattsson P, et al. Incom-plete hippocampal inversion-is there a relation to epilepsy? European Radiology 2009;19:2544-2550

76. Garel C, Chantrel E, Elmaleh M, et al. Fetal MRI: normal gestational landmarks for cerebral biometry, gyration and myelination. Child’s ner-vous system 2003;19:422-425

77. Righini A, Zirpoli S, Parazzini C, et al. Hip-pocampal infolding angle changes during brain development assessed by prenatal MR imag-ing. AJNR American journal of neuroradiology 2006;27:2093-2097

78. Nemzek WR, Brodie HA, Hecht ST, et al. MR, CT, and plain film imaging of the developing skull base in fetal specimens. AJNR Am J Neuro-radiol 2000;21:1699-1706

52


79. Jeffery N. A high-resolution MRI study of linear growth of the human fetal skull base. Neuroradi-ology 2002;44:358-366

80. Robinson AJ, Blaser S, Toi A, et al. MRI of the fetal eyes: morphologic and biometric assessment for abnormal development with ultrasonographic and clinicopathologic correlation. Pediatric Ra-diology 2008;38:971-981

81. Paquette LB, Jackson HA, Tavare CJ, et al. In utero eye development documented by fetal MR imaging. AJNR American journal of neuroradi-ology 2009;30:1787-1791

82. Brugger PC, Weber M, Prayer D. Magnetic reso-nance imaging of the fetal efferent lacrimal path-ways. European Radiology 2010;20:1965-1973

83. Robson CD. Congenital hearing impairment. Pe-diatr Radiol 2006;36:309-324

84. Fryer A. Syndromes associated with hereditary deafness. In: King SJ BA, ed. Pediatric ENT ra-diology. Berlin: Springer; 2003:35–54

85. Shih JC, Shyu MK, Lee CN, et al. Antenatal de-piction of the fetal ear with three-dimensional ul-trasonography. Obstet Gynecol 1998;91:500-505

86. Chitkara U, Lee L, Oehlert JW, et al. Fetal ear length measurement: a useful predictor of aneu-ploidy? Ultrasound Obstet Gynecol 2002;19:131-135

87. Shimizu T, Salvador L, Hughes-Benzie R, et al. The role of reduced ear size in the prenatal detec-tion of chromosomal abnormalities. Prenat Di-agn 1997;17:545-549

88. Isaacson G, Mintz MC. Prenatal sonographic vi-sualization of the inner ear. J Ultrasound Med 1986;5:409-410

89. Tilea B, Garel C, Menez F, et al. Contribution of fetal MRI to the diagnosis of inner ear ab-normalities: report of two cases. Pediatr Radiol 2006;36:149-154

90. Milic A, Blaser S, Robinson A, et al. Prenatal detection of microtia by MRI in a fetus with tri-somy 22. Pediatr Radiol 2006;36:706-710

91. Katorza E, Nahama-Allouche C, Castaigne V, et al. Prenatal evaluation of the middle ear and diagnosis of middle ear hypoplasia using MRI. Pediatric Radiology 2011;41:652-657

92. Maarse W, Berge SJ, Pistorius L, et al. Diagnos-tic accuracy of transabdominal ultrasound in de-tecting prenatal cleft lip and palate: a systematic review. Ultrasound Obstet Gynecol 2010;35:495-502

93. Offerdal K, Jebens N, Syvertsen T, et al. Prenatal ultrasound detection of facial clefts: a prospec-tive study of 49,314 deliveries in a non-selected population in Norway. Ultrasound Obstet Gyne-col 2008;31:639-646

94. Stroustrup Smith A, Estroff JA, Barnewolt CE, et al. Prenatal diagnosis of cleft lip and cleft palate using MRI. AJR Am J Roentgenol 2004;183:229-235

95. Abramowicz S, Cooper ME, Bardi K, et al. De-mographic and prenatal factors of patients with cleft lip and cleft palate. A pilot study. J Am Dent Assoc 2003;134:1371-1376

96. Costello BJ, Edwards SP, Clemens M. Fetal diag-nosis and treatment of craniomaxillofacial anom-alies. J Oral Maxillofac Surg 2008;66:1985-1995

97. Nusbaum R, Grubs RE, Losee JE, et al. A quali-tative description of receiving a diagnosis of clefting in the prenatal or postnatal period. J Genet Couns 2008;17:336-350

98. Manganaro L, Tomei A, Fierro F, et al. Fetal MRI as a complement to US in the evaluation of cleft lip and palate. Radiol Med 2011

99. Ghi T, Tani G, Savelli L, et al. Prenatal imaging of facial clefts by magnetic resonance imaging with emphasis on the posterior palate. Prenat Di-agn 2003;23:970-975

100. Descamps MJ, Golding SJ, Sibley J, et al. MRI for definitive in utero diagnosis of cleft palate: a useful adjunct to antenatal care? Cleft Palate Craniofac J 2010;47:578-585. Epub 2010 Feb 2028.

101. Kazan-Tannus JF, Levine D, McKenzie C, et al. Real-time magnetic resonance imaging aids pre-natal diagnosis of isolated cleft palate. J Ultra-sound Med 2005;24:1533-1540

102. Wang G, Shan R, Zhao L, et al. Fetal cleft lip with and without cleft palate: Comparison be-tween MR imaging and US for prenatal diagno-sis. European Journal of Radiology 2011;79:437-442

103. Mailath-Pokorny M, Worda C, Krampl-Bettel-heim E, et al. What does magnetic resonance imaging add to the prenatal ultrasound diagno-sis of facial clefts? Ultrasound Obstet Gynecol 2010;36:445-451

104. Rosen H, Chiou GJ, Stoler JM, et al. Magnetic Resonance Imaging for Detection of Brain Ab-normalities in Fetuses With Cleft Lip and/or Cleft Palate. Cleft Palate Craniofac J 2011;48:619-622. Epub 2010 Aug 2017.

53

References

105. http://faculty.vassar.edu/lowry/VassarStats.html. 106. Okada Y, Kato T, Iwai K, et al. Evaluation of hip-

pocampal infolding using magnetic resonance imaging. Neuroreport 2003;14:1405-1409.

107. Chi JG, Dooling EC, Gilles FH. Gyral develop-ment of the human brain. Annals of neurology 1977;1:86-93

108. Habas PA, Scott JA, Roosta A, et al. Early Fold-ing Patterns and Asymmetries of the Normal Hu-man Brain Detected from in Utero MRI. Cereb Cortex 2011

109. Kenna M. Embryology and developmental anato-my of the ear. In: Bluestone CD SS, Scheetz MD, ed. Pediatric Otolaryngology. Philadelphia: WB Saunders; 1990:77–87

110. Nemzek WR, Brodie HA, Chong BW, et al. Im-aging findings of the developing temporal bone in fetal specimens. AJNR Am J Neuroradiol 1996;17:1467-1477

111. Sperber G. Craniofacial Embryology. London: Wright, Butterworth Scientific; 1989

112. Tortori-Donati P, Rossi A, Castillo M, et al. Dis-orders of the Temporal Bone. Pediatric Neurora-diology: Springer Berlin Heidelberg; 2005:1361-1389

113. Jeffery N, Spoor F. Prenatal growth and devel-opment of the modern human labyrinth. J Anat 2004;204:71-92

114. Som PM, Curtin HD. Head and neck imaging. St. Louis, Mo.: Mosby; 2003

115. Krombach GA, Honnef D, Westhofen M, et al. Imaging of congenital anomalies and acquired lesions of the inner ear. Eur Radiol 2008;18:319-330

116. Paparella MM. Mondini’s deafness. A review of histopathology. The Annals of otology, rhinology & laryngology Supplement 1980;89:1-10

117. Jackler RK, Luxford WM, House WF. Congeni-tal malformations of the inner ear: a classifica-tion based on embryogenesis. Laryngoscope 1987;97:2-14

118. Marple B, Roland P. External Auditory canal. In: P Roland BM, W Meyerhoff, ed. Hearing Loss. New York: Thieme; 1997:133

119. Schmook MT, Brugger PC, Weber M, et al. Forebrain development in fetal MRI: evalua-tion of anatomical landmarks before gestational week 27. Neuroradiology 2010;52:495-504

120. Sommerlad M, Patel N, Vijayalakshmi B, et al. Detection of lip, alveolar ridge and hard pal-ate abnormalities using two-dimensional ul-trasound enhanced with the three-dimensional reverse-face view. Ultrasound Obstet Gynecol 2010;36:596-600

121. Landis JR, Koch GG. The measurement of ob-server agreement for categorical data. Biometrics 1977;33:159-174

122. Sadler TW, Langman J. Langman’s essential medical embryology. Philadelphia: Lippincott Williams & Wilkins; 2005

123. Urata MM, Kawamoto Jr HK. Median clefts of the upper lip: a review and surgical manage-ment of a minor manifestation. J Craniofac Surg 2003;14:749-755

124. Demircioglu M, Kangesu L, Ismail A, et al. In-creasing accuracy of antenatal ultrasound diag-nosis of cleft lip with or without cleft palate, in cases referred to the North Thames London Re-gion. Ultrasound Obstet Gynecol 2008;31:647-651

125. Wang LM, Leung KY, Tang M. Prenatal evalua-tion of facial clefts by three-dimensional extend-ed imaging. Prenat Diagn 2007;27:722-729

126. Suzuki H, Yamaguchi T, Furukawa M. Rhinolog-ic computed tomographic evaluation in patients with cleft lip and palate. Archives of otolaryngol-ogy--head & neck surgery 1999;125:1000-1004

127. Schopf V, Kasprian G, Brugger PC, et al. Watch-ing the fetal brain at ‘rest’. Int J Dev Neurosci 2011

mri studies of the fetal brain and craniumuu.diva-portal.org/smash/get/diva2:469024/fulltext02.pdf14...

Documents