- Email: [email protected]
Advancing Fetal Brain MRI: Targets for the Future
Advancing Fetal Brain MRI: Targets for the Future Catherine Limperopoulos, PhD,*,†,‡,§ and Cedric Clouchoux, PhD*,§ Fetal MRI is becoming an increasingly powerful imaging tool for studying brain development in vivo. Until recently, the application of advanced magnetic resonance imaging techniques was limited by motion in the nonsedated fetus. Extensive research efforts currently underway are focusing on the development of dedicated magnetic resonance imaging sequences and sophisticated postprocessing techniques that are revolutionizing our ability to study the healthy and compromised fetus. The ongoing refinement of these magnetic resonance imaging techniques will undoubtedly lead to the development of cornerstone biomarkers that will provide healthcare caregivers with vital, and currently lacking, information upon which to counsel parents effectively, and base rational decisions regarding the timing and type of novel medical and surgical interventions currently on the horizon. Semin Perinatol 33:289-298 © 2009 Elsevier Inc. All rights reserved. KEYWORDS fetal, brain, MRI, development
I
n vivo fetal magnetic resonance imaging (MRI) is revolutionizing our ability to study human brain development. To date, fetal MRI has been used primarily for the qualitative morphologic evaluation of the fetal brain. However, the recent application of advanced MRI techniques to the fetus has provided an unprecedented opportunity to investigate the developing brain in vivo. The ability to begin reliable investigation of brain growth and development in the healthy and compromised fetus promises a range of new quantitative biomarkers that can be applied clinically, and that will help to formulate a better understanding of brain development and lead to improved management of high-risk pregnancies. This article explores the advancing role of fetal brain MRI, provides an overview of the current obstacles in MRI of the living fetus, and highlights future targets in this rapidly evolving field.
*Department of Neurology and Neurosurgery, McGill University, Montreal, Quebec, Canada. †School of Physical and Occupational Therapy, McGill University, Montreal, Quebec, Canada. ‡Department of Pediatrics, McGill University, Montreal, Quebec, Canada. §McConnell Brain Imaging Centre, Neurological Institute, McGill University, Montreal, Quebec, Canada. This work was partially supported by the Canadian Research Chairs Program (to C.L.) and the Canadian Institutes of Health Research. Address reprint requests to Catherine Limperopoulos, PhD, Canada Research Chair in Brain and Development, Montreal Children’s Hospital, Pediatric Neurology, 2300 Tupper St A-334, Montreal, Quebec, Canada, H3H 1P3. E-mail: [email protected]
0146-0005/09/$-see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1053/j.semperi.2009.04.002
Challenges of Fetal MRI The in vivo study of the developing cerebral parenchyma is very complex and continues to be challenged by several factors. One of the biggest problems encountered in performing nonsedated fetal MRI is fetal motion.1,2 Despite the recent development in ultrafast MRI that facilitates image acquisition in milliseconds, advanced fetal MRI acquisitions, such as three-dimensional (3-D)-acquired T2-weighted sequences for 3-D brain reconstruction, diffusion tensor imaging (DTI), or magnetic resonance spectroscopic imaging, typically take longer to acquire and are therefore more susceptible to motion artifact. Although adult brain motion can be controlled and easily corrected,3 fetal motion is uncontrollable without sedation, is unpredictable, and occurs in all planes.4,5 In normal conditions, motion implies that the resulting slices are not perfectly parallel to each other, and that motion occurs with overlapping signal tissues resulting in image artifact.3 Briefly, the problem can be summarized as a compromise between short acquisition time and good image resolution. A short acquisition time is needed to reduce both maternal and fetal motion artifact as well as subject discomfort; however, improved image resolution requires an increased scanning time. The high water content of the immature, largely unmyelinated fetal brain also results in poor contrast in the cerebral parenchyma (ie, low resolution), making it inherently difficult to obtain high-resolution MR images. Consequently, poor tissue contrast can affect the diagnostic accuracy of fetal MRI. Decreased tissue contrast, for example, can hinder the 289
C. Limperopoulos and C. Clouchoux
290 reliable identification of diffuse white matter abnormalities, which is of critical importance for detecting conditions that may affect the fetal brain, such as hypoxia-ischemia and viral infections.6,7 Another important challenge is the clinical interpretation of more subtle anatomical anomalies of unclear long-term significance, which are increasingly detected by fetal MRI. Their consideration often requires taxing decisionmaking by both care providers and families, and may lead to termination of pregnancy.8,9 Moreover, it has been shown that in cases in which a viable fetus is delivered, a misdiagnosis and resulting errors in prognostication have serious long-term consequences for parents.8,10 It is noteworthy that very few MRI series of normal fetal brain development are currently available, resulting in a fundamentally poor understanding of the limits of normality and variability.1,8 These ongoing fetal MRI challenges are subject to extensive research, including the development of dedicated fetal MRI sequences, as well as sophisticated postprocessing algorithms described later in the text.
Overcoming Current Challenges of Fetal MRI MR Image Acquisition Several strategies have been explored to decrease motionrelated artifact. The development of specific fetal MRI sequences is one way to attempt to overcome or minimize image degradation associated with motion artifact; however, this remains a multifaceted task because of the inherent na-
ture of the immature brain and surrounding uterine environment. Jiang et al4 have proposed the use of repeated dynamic single-shot MRI sequences to sample the region of interest, leading to the acquisition of multiple overlapping slices. On the basis of the assumption that every part of the fetal brain is sampled, a registration postprocessing technique allows for the recovery of the original information in its entirety. This preliminary work has shown promising results, although ongoing research in fetal MRI acquisition is still needed.
MRI Post-Processing Techniques Another approach to overcome image degradation secondary to fetal motion relies on the application of advanced postprocessing techniques. Development in dedicated motion artifact detection and removal tools is currently under research. Algorithms developed for adults often fail, given that fetal and maternal motion are often mixed, as well as the intrinsic immaturity and limited contrast of the fetal brain. Typically, 3 fetal scans are acquired, 1 in each direction, providing a high-resolution view for each direction. The challenge then is to re-create a single, isotropic high-resolution volume.4,5 These techniques rely on the assumption that the entire brain has been sampled in multiple shots, and that the resulting images contain the entire information set. Various steps are then mandatory, and specific algorithms must be developed to recover this information and create high-resolution images (Fig. 1). A complex set of registration and segmentation algorithms are involved; however, a detailed description of these methodologies is beyond the scope of this article.
Figure 1 High resolution image reconstruction of a healthy 35 week gestational age fetus. Left column: original T2-weighted images. Each row represents low-resolution scans in 1 direction (axial, sagittal, and coronal). Right column: resulting high-resolution isotropic volume (1 ⫻ 1 ⫻ 1 mm3) after preprocessing and coregistration.
Advancing fetal brain MRI
291
Current Status of Advanced Fetal MR Imaging Techniques Volumetric Brain Growth of the Fetal Brain
Figure 2 MR image reconstruction using 3-D inpainting in a healthy 34 week gestational age fetus. Axial T2-weighted image on the left represents the original corrupted MRI slice (red arrow), and the image on the right demonstrates the same reconstructed slice using 3-D inpainting. (Color version of figure is available online.)
The presence of corrupted slices during the acquisition phase is another motion-related challenge in fetal MRI studies. Although conventional interpolation methods may be used to overcome the problem, an elegant and simple solution has recently been proposed,11 in which corrupted slices are replaced using a novel in-painting process, taking advantage of existing data from surrounding slices (Fig. 2). The ongoing development and refinement of dedicated postprocessing algorithms is mandatory for image contrast enhancement and the creation of high-resolution images is necessary for quantitative fetal brain measurements.
Quantitative 3-D Volumetric MRI Quantitative 3-D volumetric MRI has provided major insights into the developmental changes in specific brain structures and tissue subtypes of the immature brain infant.12,13 The ability to make quantitative measurements of gray and white matter volumes in the preterm infant over the third postconceptional trimester has advanced our understanding of the rate and progression of brain development,13 as well as normal and abnormal cerebral cortical development and myelination. More recent studies have begun to explore the feasibility of acquiring 3-D volumetric data in utero, allowing for 3-D reconstruction and volume rendering of fetal brain parenchyma and extra-axial cerebrospinal fluid.14,15 However, 3-D fetal MRI remains difficult because of motion during image acquisition. Motion corrupts the 3-D position and orientation of individual slices during the acquisition. Studies are now beginning to examine fetal brain growth in healthy and high-risk fetuses. To date, determination of fetal brain volume has been performed manually (Fig. 3) to quantify 3-D brain growth.16,17 A recent study using quantitative 3-D volumetric MRI demonstrated a strong linear relationship between total brain volume and gestational age in healthy second and third trimester fetuses. It also described the first in vivo evidence of abnormal brain growth in fetuses with con-
Figure 3 Three-dimensional manual segmentation of total brain volume (blue) and intracranial cavity volume (red), using multiplanar T2-weighted images. (Color version of figure is available online.)
C. Limperopoulos and C. Clouchoux
292
Figure 4 Automatic volumetric parcellation of a healthy 33-week fetal brain. First row represents mulitplanar original T2-weighted images. Bottom row represents automatic parcellation of the cerebrum (white), cerebellum (yellow), and brainstem (blue). (Color version of figure is available online.)
genital heart disease as compared with healthy control fetuses, characterized by progressively smaller third trimester total brain volume.18 These preliminary data promise exciting opportunities to begin to quantify the timing and type of aberrant brain growth in compromised fetuses in utero. The various postprocessing steps currently in use in fetal MRI are cumbersome and time-consuming, and their automation would represent a major advancement for several reasons. The first advantage is the obvious gain of time, particularly when processing large MRI datasets. The second advantage of automated processing pipelines is minimizing human bias. Understandably, extensive validation studies would be needed to ensure the reliability of these postprocessing automated techniques. Innovative automated modelbased techniques for fetal brain parcellation (Fig. 4) are being developed and validated.19 These techniques are already facilitating the quantitative study of third trimester regional brain growth in healthy fetuses (Fig. 4). Application of these automated approaches to characterize temporal and regional brain growth in fetuses at risk for impaired brain development is currently underway.
Microstructural Development of the Fetal Brain Diffusion-Weighted Imaging and DTI Diffusion-weighted imaging (DWI) and DTI are techniques used to evaluate maturation-dependent microstructural changes associated with brain growth and development of cerebral white matter.13,20-23 From a technical perspective, diffusion refers to the migration of water molecules over intracapillary distances.23 On the basis of this principle, DWI is based on the preferential diffusion of water molecules in a magnetic field. Increased diffusion reduces MR signal in a specific direction, whereas lower diffusion results in less signal loss and a brighter display in the image. The degree of diffusion weighting corresponds to the strength of diffusion
gradients, characterized by their b-value. Specific acquisition sequences are then used to capture diffusion-weighted data and to generate images that give precise information on molecular displacements over short distances, comparable to cell dimensions.24 To date, these advanced MRI techniques have been limited mainly to postnatal MRI studies in high-risk pre-term and full-term infants.1 Technical challenges encountered with fetal diffusion imaging relate primarily to echo-planar imaging with inherently noisy sequences very sensitive to motion artifacts. Additionally, the high-water content of the immature brain amplifies this phenomenon of diffusion.24 Spatial resolution is poor, which may in turn lead to a partial volume effect. Despite these challenges, several studies have explored the role of DWI in the fetus and have demonstrated that DWI can be used to investigate normal and abnormal brain development.24-30 A progressive decrease of the apparent diffusion coefficient after 30 weeks gestation age in the supratentorial regions has been demonstrated, using fetal MRI,86 similar to that previously reported in premature infants. Normal values of fetal brain apparent diffusion coefficient measures have been established, which allow detection of pathologic changes, such as hemorrhage and acute ischemia.26,27,31,32 DTI is a technique derived from DWI, determining the direction and magnitude of the water molecules diffusion.33 In particular, DTI allows the visualization and quantification of white matter fiber direction. This is made possible by measuring fractional anisotropy, which gives information on the shape of the diffusion tensor for each voxel (based on normalized variance of eigenvalues). The difference between isotropic and anisotropic diffusion is then obtained, giving information on white matter structure and state. Even though there is an intrinsic low anisotropy in the developing brain, white matter fiber tract maturational changes have been described from birth.34 Using fractional anisotropy in this manner, 3 broad phases in postnatal white matter development
Advancing fetal brain MRI
293 is an important landmark for future studies of abnormal white matter development in the living fetus.
Metabolic Development of the Fetal Brain
Figure 5 In vivo diffusion tensor tractography (courtesy of Drs Prayer and Kasprian). The fibers of the bilateral corticospinal tracts (green and blue fibers), the genu (pink fibers), and splenium (yellow fibers) of the corpus callosum have been superimposed to an anatomical T2-weighted image. Fiber tracking was performed based on a diffusion-tensor sequence with 36 directions. (Color version of figure is available online.)
have been described: that is, rapid changes in the size and shape of white matter tracts over the first 12 months, slower changes over the second year, and a relatively stable picture thereafter. These findings have been used to generate a normative database34 that is available online and provides a useful framework for studies of postnatal brain development. Despite its potential importance for the study of fetal brain development, the clinical application of fetal DTI remains limited. To date, only 1 study has successfully applied DT tractography to the apparently normal fetus35 to describe the in vivo microstructural development of white matter (Fig. 5) from as early as 18 weeks of gestation. This pioneering work
Proton Magnetic Resonance Spectroscopy Proton magnetic resonance spectroscopy (1H-MRS) is becoming a powerful noninvasive tool for examining cerebral metabolism in the fetus in vivo. This technique is based on hydrogen resonance frequency that is altered by the immediate chemical environment,36 and the concentration of the metabolite is directly linked to the resulting signal intensity. 1H-MRS allows measurement of specific brain metabolites, such as the N-acetyl aspartate (NAA), a neuroaxonal marker reflecting development of dendrites and synapses, as well as mitochondrial metabolism, creatine (Cr) reflecting cellular energy metabolism, choline (Cho) a marker of myelination, myoinositol, a glial marker, and lactate that accumulates during anaerobic metabolism (Fig. 6). The application of 1H-MRS to the fetus continues to face inherent scanning challenges, such as the lack of availability of dedicated coils, the distance from the fetal brain, and requirements of a long acquisition time.1 Nonetheless, several studies have described the in vivo metabolic maturation of the fetal brain between 22 and 39 weeks’ gestational age.36-43 In the normal fetus, Cr is visible as early as 22 weeks, along with a small NAA peak. Both NAA and Cr values increase with increasing gestational age, presumably reflecting synapse and dendrite development, whereas Cho gradually decreases likely representing variations in substrate required for membrane synthesis and myelination.36-39,43 Although cerebral lactate has been identified by neonatal 1H-MRS in stable premature infants, it has not been reported in normal fetuses.40,44 The availability of normative 1H-MRS data for fetal brain metabolites provides a valuable reference for measurement of cerebral metabolites in pathologic conditions of the fetal brain. Several studies have begun to explore metabolic alterations associated with MRI-detected structural changes, as well as in conditions of potential fetal compromise.45-47 In a
Figure 6 Proton MRS (TE 144 ms) acquired in a healthy, 33-week gestational age fetus with voxel positioned in the left cerebral hemisphere. Identifiable peaks on magnetic resonance (MR) spectra include N-acetyl-aspartate (NAA), creatine (Cr), and choline (Cho). (Color version of figure is available online.)
294 recent study, Limperopoulos et al48 described progressively lower NAA and/or Cho ratios during the third trimester in fetuses with congenital heart disease as compared with healthy fetuses. In addition, the presence of cerebral lactate was detected in 20% of fetuses with congenital heart disease, suggesting impaired third trimester brain metabolism.48 Long-term follow-up studies are underway to examine the predictive validity of impaired brain metabolism in utero.
Novel Fetal MR Imaging Techniques on the Horizon Fetal Brain Activity by MRI Functional MRI has been used extensively in the adult and an older child to study the regional activation of the brain to specific stimuli and activities.49-51 This approach is based on the normal regional neuronal perfusion coupling in the brain and investigates regional changes in blood oxygenation using the blood oxygen level-dependent response, which can be summarized as follows. Specifically, regional neuronal activation responses to specific stimuli trigger a local increase in blood flow, blood volume, and venous blood oxygenation. This in turn changes the oxy- and deoxyhemoglobin difference, thereby increasing the local MRI blood oxygen leveldependent contrast signal.52,53 With regard to applying functional MRI to the fetus, important technical challenges persist related to fetal motion, applying stimuli to the fetus, and the analysis and interpretation of data.53,54 Therefore, functional MRI in the fetus remains limited to research protocols at this time. More details on this technique and its clinical application are reviewed by Gowland and Fulford52 elsewhere in this issue.
Fetal Behavior by MRI Preliminary studies are beginning to explore the application of MRI for the evaluation of fetal movement to provide a snapshot of the functional development of the fetal nervous system at different gestational ages. This approach has the potential to expand the assessment of normal brain function in the fetus through assessment of spontaneous movement patterns.2 Fetal behavior comprises spontaneous and reflexive movement55 and increases in complexity with increasing maturity. Fetal movements rely on an intact neuromuscular system and a normal metabolic state of the central nervous system.56 A range of known movements characterize normal fetal behavior. As described by Prayer,2 the different growth stages have specific and characteristic movements. For instance, the first fetal movements occur at around the 8th week of gestation and are characterized by flexion and extension of the vertebral column. It is noteworthy that until the 19th week of gestation, these movements do not originate from cerebral activity, and likely result from spontaneous discharges at the spinal and brainstem levels. Coordinated complex movements, involving the arm, leg, neck and trunk, appear from the 9th week of gestation.57 More organized movements are observable from the 14th week, while independent movements of the extremities appear between 26
C. Limperopoulos and C. Clouchoux and 32 weeks’ gestational age.58 This approach could become a useful adjunct to other techniques for studying fetal neurologic development. Deviation of gestational age-appropriate movement repertoires may be used in future to identify disturbed fetal development.
Placental Perfusion by Fetal MRI Normal function of the uteroplacental unit is critical for fetal growth and development. Therefore, reliable techniques for evaluating dynamic placental function are of major importance for the assessment and management of high-risk pregnancies, such as those complicated by pre-eclampsia and intrauterine growth restriction.59-61 To date, the mainstay technique in the field has been Doppler ultrasound. However, MRI has several potential advantages over Doppler ultrasound in that it is not dependent on adequate amniotic fluid volume or affected by maternal obesity or a posteriorly positioned placenta.62 Several features of the placenta, including its relative immobility, high blood volume (about 50% of placental volume), and high rate of perfusion, make it particularly amenable in theory to MRI perfusion studies. Despite the theoretical ease of placenta imaging, to date, few MR perfusion imaging studies have been carried out. Perfusion imaging techniques depend on the availability of a contrast agent to track microscopic perfusion exchanges63 and a detection system with high temporal resolution, as the tracer may be visible only briefly. In animal models, it has been shown that MRI perfusion techniques provide a robust measure of placental blood flow between the fetus and the mother.64 In these animal studies, microcirculatory MRI perfusion studies have become possible because of the development of new contrast agents with specific biodistribution and faster MRI acquisition sequences.64-68 Data from these experimental studies have enabled the development of new models of placental perfusion and permeability in vivo.64,69,70 Despite the promising results provided by the use of MRI contrast agents in experimental studies, the extent to which contrast agents cross the placenta and their rate of clearance remains unclear, thus limiting their application for clinical evaluation of placental perfusion. Consequently, other methods for studying placental perfusion by MRI techniques have been explored. Arterial spin labeling is an MRI technique widely used to measure perfusion. Specifically, the flow-sensitive alternating inversion recovery technique compares the rate of recovery of the magnetization following nonselective and selective inversion pulses. In studies testing this approach in both healthy and high-risk pregnant women,71 it was found that although average global placental perfusion was similar in the 2 groups, regions of decreased perfusion were described in the placentas of growth-restricted fetuses. Another MRIbased approach that has been applied to measurement of placental perfusion is the intravoxel incoherent motion (IVIM) technique. The IVIM technique was first developed to measure the perfusion (or blood volume) in capillary networks by measuring signal attenuation caused by phase dis-
Advancing fetal brain MRI persion of randomly moving water protons when a magnetic field gradient is applied.72 Water moving in a capillary network can be modeled as a fraction of the voxel that is diffused very rapidly. Therefore, the volume of moving blood in the placenta can be measured as that fraction of water that is moving fast enough to be dephased during acquisition of IVIM sequence.72 Preliminary studies applying the IVIM technique to measure placental perfusion have provided promising results.73 Specifically, this technique has demonstrated a band of high moving-blood fraction on the side of the placenta adjacent to the uterine wall.73 In cases of preeclampsia or fetal growth retardation, IVIM studies in late gestation have shown reduced perfusion of the placental basal plate. Conversely, the fraction of moving blood on the fetal side of the placenta is increased in fetuses with intrauterine growth restriction, changes ascribed by the authors to an alteration of the villous structure and a regional decrease in vascular resistance. Although the application of MR imaging techniques for studying placental function are promising, substantially more longitudinal data with these techniques are needed to characterize normal placental perfusion, and to determine whether early changes in placental perfusion are predictive of subsequent fetal compromise.
Development of an In Vivo 3-D Fetal MRI Brain Atlas Although brain atlases are widely used for image processing in adult MRI, no such atlas is currently available for the fetal brain. The development of a 3-D MRI fetal brain atlas would expedite our understanding of normal and aberrant in vivo development of the fetal brain. The fundamental steps in development of an MRI brain atlas involve building a template from the statistical analysis of multiple brain studies in the population of interest.74,75 The basic principles would be
295 similar to those used for the creation of adult brain atlases, including brain segmentation, registration, and averaging techniques of a set of fetal brain volumes to a common reference. However, there would be important differences in approach extending beyond the technical considerations described earlier. Perhaps the most fundamental difference is the dynamic anatomic changes that characterize normal brain development. In the early second trimester, for example, the surface of the cerebral cortex is normally smooth (lissencephalic) with primary sulci emerging between 18 and 20 weeks’ gestation and secondary sulci around 30 to 32 weeks. Consequently, a meaningful fetal brain MRI atlas would actually consist of multiple gestational age-dependent atlases that reflect these and multiple other evolving features of the developing brain. This is no trivial challenge given the need for a large-scale normative fetal MRI database that spans across all gestational ages. Recent innovations in fetal image reconstruction using advanced registration methods to correct for fetal motion5,76,77 have allowed acquisition of high-resolution 3D-MRI images of the fetal brain, and will facilitate the creation of a fetal brain atlas. Development of such a 3D-MRI fetal brain atlas extending from 25 to 37 weeks of gestation is in preparation,78 based on a large normative database of nonsedated fetal MRI studies (Fig. 7). The availability of a 3-D fetal MRI brain atlas will assist in the development of advanced post-processing techniques for fetal volumetric brain reconstruction and automatic volume rendering of fetal brain structures and tissue types (e.g., cortical grey matter, unmyelinated and myelinated white matter, germinal matrix, cerebrospinal fluid) for the future study of regional tissue-specific in utero brain development. In addition to volumetric studies of the major brain structures, this approach may ultimately allow volume measurements of tissue subtypes (eg, cortical gray matter, unmyelinated and myelinated white matter, cerebrospinal
Figure 7 The development of a high-resolution in vivo fetal MRI brain atlas. The top row represents the first atlas (mean gestational age of 27 weeks). The bottom row illustrates the second fetal atlas (mean gestational age of 33 weeks).
C. Limperopoulos and C. Clouchoux
296
Figure 8 Extracted cortical surface from a 28-week gestational age fetus. Top row shows right and left cerebral hemispheres, viewed from 3 different angles. Bottom row shows the left hemisphere (external and internal surfaces).
fluid) in specific parcellated brain regions, facilitating the study of regional tissue-specific brain development. The developmental schedule for gyral formation in the developing fetus has been well defined through both pathology and qualitative MRI studies.78-80 Quantitative techniques for measurement of cortical folding previously described in adults, have recently been applied to preterm infants ex utero.81 An extension of this approach to the study of fetal sulcal development is currently underway, using a novel methodological framework for fetal cortical surface rendering (Fig. 8) (Clouchoux et al, unpublished data). This approach is based on a technique previously validated in adults using a 2-D mesh of the cortical surface brains.82-85 Other potential future developments might include cortical thickness measurement and volume-based morphometry.
Summary and Future Directions Fetal MRI is rapidly becoming a powerful tool for investigating in vivo fetal brain development in both clinical and research studies. To date, our understanding of the dynamic and highly intricate evolution of the fetal brain has been impeded by a lack of standardized and reliable MRI benchmarks. The recent development and ongoing refinement of quantitative 3-D volumetric MRI, DTI, MR spectroscopy, and functional and perfusion imaging now provide us with the opportunity to evaluate the fetal brain from a complementary, integrated perspective to include volume, microstructure, metabolism, and function, respectively. These timely methodological developments now provide us with an unlimited potential to study in unprecedented detail normal fetal brain development as well as the mechanisms and consequences that underlie impaired brain development in the compromised fetus. Advancing the role and accuracy of MRI in the fetus will require ongoing methodological developments in the areas of image acquisition and advanced postprocessing techniques to overcome current challenges of fetal motion and image
degradation. Moreover, the creation of a large-scale highresolution 3-D in vivo fetal brain atlas will permit greater insight into neuronal and axonal pathway development, as well as sulcal and gyral formation, and provide a unique window on the timing of insults that might disrupt normal brain development. Advances expected to result from this work will greatly improve our ability to reliably detect and monitor the high-risk fetus, improve parental counseling, and evaluate the potential benefits of prenatal interventions.
References 1. Garel C: Fetal MRI: What is the future? Ultrasound Obstet Gynecol 31:123-128, 2008 2. Prayer D: Investigation of the normal organ development with fetal MRI. Eur Radiol 17:2458-2471, 2006 3. Jenkinson M, Smith S: Improved optimization for the robust an accurate linear registration and motion correction of brain images. Neuroimage 17:825-841, 2002 4. Jiang S, Xue H, Counsell S, et al: In-utero three dimension high resolution fetal brain diffusion tensor imaging. MICCAI 10:18-26, 2007 5. Rousseau F, Glenn OA, Iordanova B, et al: Registration-based approach for reconstruction of high-resolution in utero fetal MR brain images. Acad Radiol 13:1072-1081, 2006 6. Garel C, Delezoide: MRI of the Fetal Brain: Normal Development and Cerebral Pathologies. New York, NY, Springer, 2004 7. Girard N, Gire C, Sigaudy S, et al: MR imaging of acquired fetal brain disorders. Childs Nerv Syst 19:490-500, 2003 8. Limperopoulos C, Robertson RL, Estroff JA, et al: Diagnosis of inferior vermian hypoplasia by fetal magnetic resonance imaging: Potential pitfalls and neurodevelopmental outcome. Am J Obstet Gynecol 194: 1070-1076, 2006 9. Limperopoulos C, Robertson RL Jr, Khwaja OS, et al: How accurately does current fetal imaging identify posterior fossa anomalies? Am J Roentgenol 190:1637-1643, 2008 10. Sjogren B: Psychological indications for prenatal diagnosis. Prenat Diagn 16:449-454, 1996 11. Manjon JV, Guizard N, Robles M, et al: Reconstruction using 3D inpainting. Hum Brain Mapp (in press) 12. Limperopoulos C, Soul JS, Gauvreau K, et al: Late gestation cerebellar growth is rapid and impeded by premature birth. Pediatrics 115:68895, 2005 13. Hüppi PS, Warfield S, Kikinis R: Quantitative magnetic resonance im-
Advancing fetal brain MRI
14.
15.
16.
17.
18.
19.
20.
21.
22.
23. 24. 25. 26.
27.
28.
29.
30. 31. 32.
33. 34.
35. 36.
37.
38.
aging of brain development in premature and mature newborns. Ann Neurol 43:224-235, 1998 Kubik-Huch RA, Wildermuth S, Cettuzzi L, et al: Fetus and Uteroplacental unit: Fast MR imaging with three-dimensional reconstruction and volumetry—Feasibility study. Radiology 219:567-73, 2001 Schierlitz L, Dumanli H, Robinson JN, et al: In-utero three-dimensional magnetic resonance imaging of fetal brains. Lancet 357:1177-1178, 2001 Grossman R, Hoffman C, Mardor Y, et al: Quantitative MRI measurements of human fetal brain development in utero. Neuroimage 33:463470, 2006 Kazan-Tannus JF, Dialani V, Kataoka ML, et al: MR volumetry of brain and CSF in fetuses referred for ventriculomegaly. Am J Roentgenol 189:145-151, 2007 Limperopoulos C, Tworetzky W, Newburger JW, et al: Third-trimester volumetric brain growth is impaired in fetuses with congenital heart disease. Circulation 118:S908, 2008 Guizard N, Evans AC, Lepage C, et al: Automatic model-based fetal brain parcellation to quantify in vivo fetal brain development. Hum Brain Mapp (in press) Mukherjee P, Miller JH, Shimony JS, et al: Diffusion-tensor MR imaging of gray matter and white matter development during normal human maturation. Am J Neuroradiol 23:1445-1456, 2002 Anjari M, Srinivasan L, Allsop JM, et al: Diffusion tensor imaging with tract-based spatial statistics reveals local white matter abnormalities in preterm infants. Neuroimage 35:1021-1027, 2007 Dubois J, Dehaene-Lambertz G, Soarès C, et al: Microstructural correlates of infant functional development: Example of the visual pathways. J Neurosci 28:1943-1948, 2008 Stadnik T, Luypaert R, Jager T, et al: Diffusion imaging: From basic physics to practical imaging. RSNA EJ Radiog 3, 1999 Prayer D, Prayer L: Diffusion-weighted magnetic resonance imaging of cerebral white matter development. Eur J Radiol 45:235-243, 2003 Garel C: New advances in fetal MR neuroimaging. Pediatr Radiol 36: 621-625, 2006 Baldoli C, Righini A, Parazzini C, et al: Demonstration of acute ischemic lesions in the fetal brain by diffusion magnetic resonance imaging. Ann Neurol 52:243-246, 2002 Righini A, Bianchini E, Parazzini C, et al: Apparent diffusion coefficient determination in normal fetal brain: A prenatal MR imaging study. Am J Neuroradiol 24:799-804, 2003 Agid R, Lieberman S, Nadjari M, et al: Prenatal MR diffusion-weighted imaging in a fetus with hemimegalencephaly. Pediatr Radiol 36:138140, 2006 Bui T, Daire JL, Chalard F, et al: Microstructural development of human brain assessed in utero by diffusion tensor imaging. Pediatri Radiol 36:1133-1140, 2006 Kim DH, Chung S, Vigneron DB, et al: Diffusion-weighted imaging of the fetal brain in vivo. Magn Reson Med 59:216-220, 2008 Brunel H, Girard N, Confort-Gouny S, et al: Fetal brain injury. J Neuroradiol 31:123-137, 2004 Erdem G, Celik O, Hascalik S, et al: Diffusion-weighted imaging evaluation of subtle cerebral microstructural changes in intrauterine fetal hydrocephalus. Magn Reson Imaging 25:1417-1422, 2007 Le Bihan D: Molecular diffusion nuclear magnetic resonance imaging. Magn Reson Q 7:1-30, 1991 Hermoye L, Saint Martin C, Cosnard G, et al: Pediatric diffusion tensor imaging: Normal database and observation of the white matter maturation in early childhood. Neuroimage 29:493-504, 2005 Kasprian G, Brugger PC, Weber M, et al: In utero tractography of fetal white matter development. Neuroimage 43 213-24, 2008 Kok RD, van der Bergh AJ, Heerschap A, et al: Metabolic information from the human fetal brain obtained with proton magnetic resonance spectroscopy. Am J Obstet Gynecol 185:1011-1015, 2001 Kok RD, van der Berg PP, van der Bergh AJ, et al: Maturation of the human fetal brain as observed by 1H MR spectroscopy. Magn Reson Med 48:611-616, 2002 Heerschap A, Kok RD, van der berg PP: Antenatal proton MR spectroscopy of the human brain in vivo. Childs Nerv Syst 19:418-421
297 39. Kreis R: Brain metabolite composition during early human brain development as measured by quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med 48:949-958, 2002 40. Girard N, Gouny SC, Viola A, et al: Assessment of normal fetal brain maturation in utero by proton magnetic resonance spectroscopy. Magn Reson Med 56:768-775, 2006 41. Roelant-Van Rijn AM, Groenendaal F, Stoutenbeek P, et al: Lactate in the fetal brain: Detection and implications. Acta Pædiatr 93:937-940, 2004 42. Van der Voorn JP, Pouwels PJ, Hart AA, et al: Childhood white matter disorders: Quantitative MR imaging and spectroscopy. Radiology 241: 510-517, 2006 43. Fenton BW, Lin CS, Macedonia C, et al: The fetus at term: In utero volume-selected proton MR spectroscopy with breath-hold technique—A feasibility study. Radiology 219:563-566, 2001 44. Girard N, Fogliarini C, Viola A, et al: MRS of normal and impaired fetal brain development. Eur J Radiol 57:217-225, 2006 45. Borowska-Matwiejczuk K, Lemancewicz A, Tarasow E, et al: Assessment of fetal distress based on magnetic resonance examinations: Preliminary report. Acad Radiol 10:1274-1282, 2003 46. Wolfberg AJ, Robinson JN, Mulkern R, et al: Identification of fetal cerebral lactate using magnetic resonance spectroscopy. Am J Obstet Gynecol 196:e9-e11, 2007 47. Azpurua H, Alvarado A, Mayobre F, et al: Metabolic assessment of the brain using proton magnetic resonance spectroscopy in a growth-restricted human fetus: Case report. Am J Perinatol 25:305-309, 2008 48. Limperopoulos C, Tworetzky W, Robertson RL, et al: Impaired brain metabolism in fetuses with congenital heart disease. Circulation 118: S651-S-652, 2008 49. Hirsch J, Kim KHS, Souweidane MM, et al: fMRI reveals a developing language system in a 15-month old sedated infant (Abstract). Soc Neurosci 23: 222, 1997 50. Huettel SA, Song AW, McCarthy G: Functional Magnetic Resonance Imaging. Sunderland, MA, Sinauer Associates, 2004 51. Iannetti GD, Wise RG: BOLD functional MRI in disease and pharmacological studies: Room for improvement? Magn Reson Imaging 25: 978-988, 2007 52. Gowland P, Fulford J: Initial experiences of performing fetal fMRI. Exp Neurol 190 (suppl 1):S22-S27, 2004 53. Fulford J, Vadeyar SH, Dodampahala SH, et al: Fetal brain activity and hemodynamic response to a vibroacoustic stimulus. Hum Brain Mapp 22:116-121, 2004 54. Fulford J, Vadeyar SH, Dodampahala SH, et al: Fetal brain activity in response to a visual stimulus. Hum Brain Mapp 20:239-245, 2003 55. Sparling JW: Concepts in Fetal Movement Research. New York, Haworth Press, 1993 56. Olesen AG, Svare JA: Decreased fetal movements: Background, assessment, and clinical management. Acta Obstet Gynecol Scand 83:818826, 2004 57. Prechtl HF, Einspieler C: Is neurological assessment of the fetus possible? Eur J Obstet Gynecol Reprod Biol 75:81-84, 1997 58. Sparling JW, Van Tol J, Chescheir NC: Fetal and neonatal hand movement. Phys Ther 79:24-39, 1999 59. Gagnon R: Placental insufficiency and its consequences. Eur J Obstet Gynecol Reprod Biol 110 (suppl 1):99-107, 2003 60. Mihaly GW, Morgan DJ: Placental drug transfer: Effects of gestational age and species. Pharmacol Ther 23:253-266, 1983 61. Polliotti BM, Holmes R, Cornish JD, et al: Long-term dual perfusion of isolated human placental lobules with improved oxygenation for infectious diseases research. Placenta 17:57-68, 1996 62. Duncan KR: Fetal and placental volumetric and functional analysis using echo-planar imaging. Top in Magn Res Imaging 12:52-66, 2001 63. Le Bars E, Gondry-Jouet C, Deramond H, et al: MR diffusion and perfusion imaging in clinical practice. J Neurorad 57:56-76, 2000 64. Salomon LJ, Siauve N, Balvay D, et al: Placental perfusion MR imaging with contrast agents in a mouse model. Radiology 235:73-80, 2005 65. Zhu XP, Li KL, Kamaly-Asl ID, et al: Quantification of endothelial permeability, leakage space, and blood volume in brain tumors using
C. Limperopoulos and C. Clouchoux
298
66.
67.
68.
69. 70.
71.
72. 73.
74.
combined T1 and T2* contrast-enhanced dynamic MR imaging. J Magn Reson Imaging 11:575-585, 2000 Brasch R, Turetschek K: MRI characterization of tumors and grading angiogenesis using macromolecular contrast media: Status report. Eur J Radiol 34:148-155, 2000 Vallee JP, Sostman HD, MacFall JR, et al: Quantification of myocardial perfusion with MRI and exogenous contrast agents. Cardiology 88:90105, 1997 Pradel C, Siauve N, Bruneteau G, et al: Reduced capillary perfusion and permeability in human tumour xenografts treated with the VEGF signaling inhibitor ZD4190: An in vivo assessment using dynamic MR imaging and macromolecular contrast media. Magn Reson Imaging 21:845-851, 2003 Roberts JM, Lain KY: Recent insights into the pathogenesis of preeclampsia. Placenta 23:359-372, 2002 Zygmunt M, Herr F, Munstedt K, et al: Angiogenesis and vasculogenesis in pregnancy. Eur J Obstet Gynecol Reprod Biol 110 (suppl 1):1018, 2003 Gowland PA, Francis ST, Duncan KR, et al: In vivo perfusion measurements in the human placenta using echo planar imaging at 0.5T. Magn Reson Med 40:467-473, 1998 Gowland PA: Placental MRI. Seminars in fetal and neonatal. Medicine 10:485-490, 2005 Moore RJ, Ong SS, Tyler DJ, et al: Spiral artery blood volume in normal pregnancies and those compromised by pre-eclampsia. NMR Biomed 21:376-380, 2008 Evans A, Collins D, Mills S, et al: 3rd statistical neuranatomical models from 305 MRI volumes. Proc. IEEE Nuclear Science Symposium and Medical Imaging Conference 1813-1817, 1993
75. Kochunov P, Lancaster J, Thompson P, et al: An optimised individual target brain in the Talairach coordinates system. Neuroimage 17:922927, 2002 76. Rousseau F, Glenn O, Iordanova B, et al: A novel approach to high resolution fetal brain MR imaging. MICCAI 8:548-555, 2005 77. Guizrad N, Lepage C, Fonov V, et al: Development of fetus brain atlas from multi-axial MR acquisitions. ISMRM. (in press) 78. Rados M, Judas M, Kostovic I: In vitro MRI of brain development. Eur J Radiol 57:187-198, 2006 79. Perkins L, Hughes E, Srinivasan L, et al: Exploring cortical subplate evolution using magnetic resonance imaging of the fetal brain. Dev Neurosci 30:211-220, 2008 80. Salomon LJ, Garel C: Magnetic resonance imaging examination of the fetal brain. Ultrasound Obstet Gynecol 30:1019-1032, 2007 81. Dubois J, Benders M, Cachia A, et al: Mapping the early cortical folding process in the preterm newborn brain. Cereb Cor Advanc 18:14441454, 2007 82. Van Essen D, Drury H, Joshi S, et al: Functional and structural mapping of human cerebral cortex: Solution are in the surfaces. Proc Natl Acad Sci USA 95:788-795, 1998 83. Fischl B, Sereno M, Tootell R, et al: Cortical surface-based analysis. II. Segmentation and surface reconstruction. Neuroimage 9:179-194, 1999 84. Toro R, Burnod Y: Geometric atlas: Modeling the cortex as an organized surface. Neuroimage 20:1468-1484, 2003 85. Clouchoux C, Coulon O, Rivière D, et al, LNCS: Anatomically constrained surface parameterization for cortical localization. MICCAI 3750:344-35, 2005 86. Schneider JF, Confort-Gouny S, Le Fur Y, et al: Diffusion-weighted imaging in normal fetal maturation. Eur Radiol 17:2422-2429, 2007
I
n vivo fetal magnetic resonance imaging (MRI) is revolutionizing our ability to study human brain development. To date, fetal MRI has been used primarily for the qualitative morphologic evaluation of the fetal brain. However, the recent application of advanced MRI techniques to the fetus has provided an unprecedented opportunity to investigate the developing brain in vivo. The ability to begin reliable investigation of brain growth and development in the healthy and compromised fetus promises a range of new quantitative biomarkers that can be applied clinically, and that will help to formulate a better understanding of brain development and lead to improved management of high-risk pregnancies. This article explores the advancing role of fetal brain MRI, provides an overview of the current obstacles in MRI of the living fetus, and highlights future targets in this rapidly evolving field.
*Department of Neurology and Neurosurgery, McGill University, Montreal, Quebec, Canada. †School of Physical and Occupational Therapy, McGill University, Montreal, Quebec, Canada. ‡Department of Pediatrics, McGill University, Montreal, Quebec, Canada. §McConnell Brain Imaging Centre, Neurological Institute, McGill University, Montreal, Quebec, Canada. This work was partially supported by the Canadian Research Chairs Program (to C.L.) and the Canadian Institutes of Health Research. Address reprint requests to Catherine Limperopoulos, PhD, Canada Research Chair in Brain and Development, Montreal Children’s Hospital, Pediatric Neurology, 2300 Tupper St A-334, Montreal, Quebec, Canada, H3H 1P3. E-mail: [email protected]
0146-0005/09/$-see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1053/j.semperi.2009.04.002
Challenges of Fetal MRI The in vivo study of the developing cerebral parenchyma is very complex and continues to be challenged by several factors. One of the biggest problems encountered in performing nonsedated fetal MRI is fetal motion.1,2 Despite the recent development in ultrafast MRI that facilitates image acquisition in milliseconds, advanced fetal MRI acquisitions, such as three-dimensional (3-D)-acquired T2-weighted sequences for 3-D brain reconstruction, diffusion tensor imaging (DTI), or magnetic resonance spectroscopic imaging, typically take longer to acquire and are therefore more susceptible to motion artifact. Although adult brain motion can be controlled and easily corrected,3 fetal motion is uncontrollable without sedation, is unpredictable, and occurs in all planes.4,5 In normal conditions, motion implies that the resulting slices are not perfectly parallel to each other, and that motion occurs with overlapping signal tissues resulting in image artifact.3 Briefly, the problem can be summarized as a compromise between short acquisition time and good image resolution. A short acquisition time is needed to reduce both maternal and fetal motion artifact as well as subject discomfort; however, improved image resolution requires an increased scanning time. The high water content of the immature, largely unmyelinated fetal brain also results in poor contrast in the cerebral parenchyma (ie, low resolution), making it inherently difficult to obtain high-resolution MR images. Consequently, poor tissue contrast can affect the diagnostic accuracy of fetal MRI. Decreased tissue contrast, for example, can hinder the 289
C. Limperopoulos and C. Clouchoux
290 reliable identification of diffuse white matter abnormalities, which is of critical importance for detecting conditions that may affect the fetal brain, such as hypoxia-ischemia and viral infections.6,7 Another important challenge is the clinical interpretation of more subtle anatomical anomalies of unclear long-term significance, which are increasingly detected by fetal MRI. Their consideration often requires taxing decisionmaking by both care providers and families, and may lead to termination of pregnancy.8,9 Moreover, it has been shown that in cases in which a viable fetus is delivered, a misdiagnosis and resulting errors in prognostication have serious long-term consequences for parents.8,10 It is noteworthy that very few MRI series of normal fetal brain development are currently available, resulting in a fundamentally poor understanding of the limits of normality and variability.1,8 These ongoing fetal MRI challenges are subject to extensive research, including the development of dedicated fetal MRI sequences, as well as sophisticated postprocessing algorithms described later in the text.
Overcoming Current Challenges of Fetal MRI MR Image Acquisition Several strategies have been explored to decrease motionrelated artifact. The development of specific fetal MRI sequences is one way to attempt to overcome or minimize image degradation associated with motion artifact; however, this remains a multifaceted task because of the inherent na-
ture of the immature brain and surrounding uterine environment. Jiang et al4 have proposed the use of repeated dynamic single-shot MRI sequences to sample the region of interest, leading to the acquisition of multiple overlapping slices. On the basis of the assumption that every part of the fetal brain is sampled, a registration postprocessing technique allows for the recovery of the original information in its entirety. This preliminary work has shown promising results, although ongoing research in fetal MRI acquisition is still needed.
MRI Post-Processing Techniques Another approach to overcome image degradation secondary to fetal motion relies on the application of advanced postprocessing techniques. Development in dedicated motion artifact detection and removal tools is currently under research. Algorithms developed for adults often fail, given that fetal and maternal motion are often mixed, as well as the intrinsic immaturity and limited contrast of the fetal brain. Typically, 3 fetal scans are acquired, 1 in each direction, providing a high-resolution view for each direction. The challenge then is to re-create a single, isotropic high-resolution volume.4,5 These techniques rely on the assumption that the entire brain has been sampled in multiple shots, and that the resulting images contain the entire information set. Various steps are then mandatory, and specific algorithms must be developed to recover this information and create high-resolution images (Fig. 1). A complex set of registration and segmentation algorithms are involved; however, a detailed description of these methodologies is beyond the scope of this article.
Figure 1 High resolution image reconstruction of a healthy 35 week gestational age fetus. Left column: original T2-weighted images. Each row represents low-resolution scans in 1 direction (axial, sagittal, and coronal). Right column: resulting high-resolution isotropic volume (1 ⫻ 1 ⫻ 1 mm3) after preprocessing and coregistration.
Advancing fetal brain MRI
291
Current Status of Advanced Fetal MR Imaging Techniques Volumetric Brain Growth of the Fetal Brain
Figure 2 MR image reconstruction using 3-D inpainting in a healthy 34 week gestational age fetus. Axial T2-weighted image on the left represents the original corrupted MRI slice (red arrow), and the image on the right demonstrates the same reconstructed slice using 3-D inpainting. (Color version of figure is available online.)
The presence of corrupted slices during the acquisition phase is another motion-related challenge in fetal MRI studies. Although conventional interpolation methods may be used to overcome the problem, an elegant and simple solution has recently been proposed,11 in which corrupted slices are replaced using a novel in-painting process, taking advantage of existing data from surrounding slices (Fig. 2). The ongoing development and refinement of dedicated postprocessing algorithms is mandatory for image contrast enhancement and the creation of high-resolution images is necessary for quantitative fetal brain measurements.
Quantitative 3-D Volumetric MRI Quantitative 3-D volumetric MRI has provided major insights into the developmental changes in specific brain structures and tissue subtypes of the immature brain infant.12,13 The ability to make quantitative measurements of gray and white matter volumes in the preterm infant over the third postconceptional trimester has advanced our understanding of the rate and progression of brain development,13 as well as normal and abnormal cerebral cortical development and myelination. More recent studies have begun to explore the feasibility of acquiring 3-D volumetric data in utero, allowing for 3-D reconstruction and volume rendering of fetal brain parenchyma and extra-axial cerebrospinal fluid.14,15 However, 3-D fetal MRI remains difficult because of motion during image acquisition. Motion corrupts the 3-D position and orientation of individual slices during the acquisition. Studies are now beginning to examine fetal brain growth in healthy and high-risk fetuses. To date, determination of fetal brain volume has been performed manually (Fig. 3) to quantify 3-D brain growth.16,17 A recent study using quantitative 3-D volumetric MRI demonstrated a strong linear relationship between total brain volume and gestational age in healthy second and third trimester fetuses. It also described the first in vivo evidence of abnormal brain growth in fetuses with con-
Figure 3 Three-dimensional manual segmentation of total brain volume (blue) and intracranial cavity volume (red), using multiplanar T2-weighted images. (Color version of figure is available online.)
C. Limperopoulos and C. Clouchoux
292
Figure 4 Automatic volumetric parcellation of a healthy 33-week fetal brain. First row represents mulitplanar original T2-weighted images. Bottom row represents automatic parcellation of the cerebrum (white), cerebellum (yellow), and brainstem (blue). (Color version of figure is available online.)
genital heart disease as compared with healthy control fetuses, characterized by progressively smaller third trimester total brain volume.18 These preliminary data promise exciting opportunities to begin to quantify the timing and type of aberrant brain growth in compromised fetuses in utero. The various postprocessing steps currently in use in fetal MRI are cumbersome and time-consuming, and their automation would represent a major advancement for several reasons. The first advantage is the obvious gain of time, particularly when processing large MRI datasets. The second advantage of automated processing pipelines is minimizing human bias. Understandably, extensive validation studies would be needed to ensure the reliability of these postprocessing automated techniques. Innovative automated modelbased techniques for fetal brain parcellation (Fig. 4) are being developed and validated.19 These techniques are already facilitating the quantitative study of third trimester regional brain growth in healthy fetuses (Fig. 4). Application of these automated approaches to characterize temporal and regional brain growth in fetuses at risk for impaired brain development is currently underway.
Microstructural Development of the Fetal Brain Diffusion-Weighted Imaging and DTI Diffusion-weighted imaging (DWI) and DTI are techniques used to evaluate maturation-dependent microstructural changes associated with brain growth and development of cerebral white matter.13,20-23 From a technical perspective, diffusion refers to the migration of water molecules over intracapillary distances.23 On the basis of this principle, DWI is based on the preferential diffusion of water molecules in a magnetic field. Increased diffusion reduces MR signal in a specific direction, whereas lower diffusion results in less signal loss and a brighter display in the image. The degree of diffusion weighting corresponds to the strength of diffusion
gradients, characterized by their b-value. Specific acquisition sequences are then used to capture diffusion-weighted data and to generate images that give precise information on molecular displacements over short distances, comparable to cell dimensions.24 To date, these advanced MRI techniques have been limited mainly to postnatal MRI studies in high-risk pre-term and full-term infants.1 Technical challenges encountered with fetal diffusion imaging relate primarily to echo-planar imaging with inherently noisy sequences very sensitive to motion artifacts. Additionally, the high-water content of the immature brain amplifies this phenomenon of diffusion.24 Spatial resolution is poor, which may in turn lead to a partial volume effect. Despite these challenges, several studies have explored the role of DWI in the fetus and have demonstrated that DWI can be used to investigate normal and abnormal brain development.24-30 A progressive decrease of the apparent diffusion coefficient after 30 weeks gestation age in the supratentorial regions has been demonstrated, using fetal MRI,86 similar to that previously reported in premature infants. Normal values of fetal brain apparent diffusion coefficient measures have been established, which allow detection of pathologic changes, such as hemorrhage and acute ischemia.26,27,31,32 DTI is a technique derived from DWI, determining the direction and magnitude of the water molecules diffusion.33 In particular, DTI allows the visualization and quantification of white matter fiber direction. This is made possible by measuring fractional anisotropy, which gives information on the shape of the diffusion tensor for each voxel (based on normalized variance of eigenvalues). The difference between isotropic and anisotropic diffusion is then obtained, giving information on white matter structure and state. Even though there is an intrinsic low anisotropy in the developing brain, white matter fiber tract maturational changes have been described from birth.34 Using fractional anisotropy in this manner, 3 broad phases in postnatal white matter development
Advancing fetal brain MRI
293 is an important landmark for future studies of abnormal white matter development in the living fetus.
Metabolic Development of the Fetal Brain
Figure 5 In vivo diffusion tensor tractography (courtesy of Drs Prayer and Kasprian). The fibers of the bilateral corticospinal tracts (green and blue fibers), the genu (pink fibers), and splenium (yellow fibers) of the corpus callosum have been superimposed to an anatomical T2-weighted image. Fiber tracking was performed based on a diffusion-tensor sequence with 36 directions. (Color version of figure is available online.)
have been described: that is, rapid changes in the size and shape of white matter tracts over the first 12 months, slower changes over the second year, and a relatively stable picture thereafter. These findings have been used to generate a normative database34 that is available online and provides a useful framework for studies of postnatal brain development. Despite its potential importance for the study of fetal brain development, the clinical application of fetal DTI remains limited. To date, only 1 study has successfully applied DT tractography to the apparently normal fetus35 to describe the in vivo microstructural development of white matter (Fig. 5) from as early as 18 weeks of gestation. This pioneering work
Proton Magnetic Resonance Spectroscopy Proton magnetic resonance spectroscopy (1H-MRS) is becoming a powerful noninvasive tool for examining cerebral metabolism in the fetus in vivo. This technique is based on hydrogen resonance frequency that is altered by the immediate chemical environment,36 and the concentration of the metabolite is directly linked to the resulting signal intensity. 1H-MRS allows measurement of specific brain metabolites, such as the N-acetyl aspartate (NAA), a neuroaxonal marker reflecting development of dendrites and synapses, as well as mitochondrial metabolism, creatine (Cr) reflecting cellular energy metabolism, choline (Cho) a marker of myelination, myoinositol, a glial marker, and lactate that accumulates during anaerobic metabolism (Fig. 6). The application of 1H-MRS to the fetus continues to face inherent scanning challenges, such as the lack of availability of dedicated coils, the distance from the fetal brain, and requirements of a long acquisition time.1 Nonetheless, several studies have described the in vivo metabolic maturation of the fetal brain between 22 and 39 weeks’ gestational age.36-43 In the normal fetus, Cr is visible as early as 22 weeks, along with a small NAA peak. Both NAA and Cr values increase with increasing gestational age, presumably reflecting synapse and dendrite development, whereas Cho gradually decreases likely representing variations in substrate required for membrane synthesis and myelination.36-39,43 Although cerebral lactate has been identified by neonatal 1H-MRS in stable premature infants, it has not been reported in normal fetuses.40,44 The availability of normative 1H-MRS data for fetal brain metabolites provides a valuable reference for measurement of cerebral metabolites in pathologic conditions of the fetal brain. Several studies have begun to explore metabolic alterations associated with MRI-detected structural changes, as well as in conditions of potential fetal compromise.45-47 In a
Figure 6 Proton MRS (TE 144 ms) acquired in a healthy, 33-week gestational age fetus with voxel positioned in the left cerebral hemisphere. Identifiable peaks on magnetic resonance (MR) spectra include N-acetyl-aspartate (NAA), creatine (Cr), and choline (Cho). (Color version of figure is available online.)
294 recent study, Limperopoulos et al48 described progressively lower NAA and/or Cho ratios during the third trimester in fetuses with congenital heart disease as compared with healthy fetuses. In addition, the presence of cerebral lactate was detected in 20% of fetuses with congenital heart disease, suggesting impaired third trimester brain metabolism.48 Long-term follow-up studies are underway to examine the predictive validity of impaired brain metabolism in utero.
Novel Fetal MR Imaging Techniques on the Horizon Fetal Brain Activity by MRI Functional MRI has been used extensively in the adult and an older child to study the regional activation of the brain to specific stimuli and activities.49-51 This approach is based on the normal regional neuronal perfusion coupling in the brain and investigates regional changes in blood oxygenation using the blood oxygen level-dependent response, which can be summarized as follows. Specifically, regional neuronal activation responses to specific stimuli trigger a local increase in blood flow, blood volume, and venous blood oxygenation. This in turn changes the oxy- and deoxyhemoglobin difference, thereby increasing the local MRI blood oxygen leveldependent contrast signal.52,53 With regard to applying functional MRI to the fetus, important technical challenges persist related to fetal motion, applying stimuli to the fetus, and the analysis and interpretation of data.53,54 Therefore, functional MRI in the fetus remains limited to research protocols at this time. More details on this technique and its clinical application are reviewed by Gowland and Fulford52 elsewhere in this issue.
Fetal Behavior by MRI Preliminary studies are beginning to explore the application of MRI for the evaluation of fetal movement to provide a snapshot of the functional development of the fetal nervous system at different gestational ages. This approach has the potential to expand the assessment of normal brain function in the fetus through assessment of spontaneous movement patterns.2 Fetal behavior comprises spontaneous and reflexive movement55 and increases in complexity with increasing maturity. Fetal movements rely on an intact neuromuscular system and a normal metabolic state of the central nervous system.56 A range of known movements characterize normal fetal behavior. As described by Prayer,2 the different growth stages have specific and characteristic movements. For instance, the first fetal movements occur at around the 8th week of gestation and are characterized by flexion and extension of the vertebral column. It is noteworthy that until the 19th week of gestation, these movements do not originate from cerebral activity, and likely result from spontaneous discharges at the spinal and brainstem levels. Coordinated complex movements, involving the arm, leg, neck and trunk, appear from the 9th week of gestation.57 More organized movements are observable from the 14th week, while independent movements of the extremities appear between 26
C. Limperopoulos and C. Clouchoux and 32 weeks’ gestational age.58 This approach could become a useful adjunct to other techniques for studying fetal neurologic development. Deviation of gestational age-appropriate movement repertoires may be used in future to identify disturbed fetal development.
Placental Perfusion by Fetal MRI Normal function of the uteroplacental unit is critical for fetal growth and development. Therefore, reliable techniques for evaluating dynamic placental function are of major importance for the assessment and management of high-risk pregnancies, such as those complicated by pre-eclampsia and intrauterine growth restriction.59-61 To date, the mainstay technique in the field has been Doppler ultrasound. However, MRI has several potential advantages over Doppler ultrasound in that it is not dependent on adequate amniotic fluid volume or affected by maternal obesity or a posteriorly positioned placenta.62 Several features of the placenta, including its relative immobility, high blood volume (about 50% of placental volume), and high rate of perfusion, make it particularly amenable in theory to MRI perfusion studies. Despite the theoretical ease of placenta imaging, to date, few MR perfusion imaging studies have been carried out. Perfusion imaging techniques depend on the availability of a contrast agent to track microscopic perfusion exchanges63 and a detection system with high temporal resolution, as the tracer may be visible only briefly. In animal models, it has been shown that MRI perfusion techniques provide a robust measure of placental blood flow between the fetus and the mother.64 In these animal studies, microcirculatory MRI perfusion studies have become possible because of the development of new contrast agents with specific biodistribution and faster MRI acquisition sequences.64-68 Data from these experimental studies have enabled the development of new models of placental perfusion and permeability in vivo.64,69,70 Despite the promising results provided by the use of MRI contrast agents in experimental studies, the extent to which contrast agents cross the placenta and their rate of clearance remains unclear, thus limiting their application for clinical evaluation of placental perfusion. Consequently, other methods for studying placental perfusion by MRI techniques have been explored. Arterial spin labeling is an MRI technique widely used to measure perfusion. Specifically, the flow-sensitive alternating inversion recovery technique compares the rate of recovery of the magnetization following nonselective and selective inversion pulses. In studies testing this approach in both healthy and high-risk pregnant women,71 it was found that although average global placental perfusion was similar in the 2 groups, regions of decreased perfusion were described in the placentas of growth-restricted fetuses. Another MRIbased approach that has been applied to measurement of placental perfusion is the intravoxel incoherent motion (IVIM) technique. The IVIM technique was first developed to measure the perfusion (or blood volume) in capillary networks by measuring signal attenuation caused by phase dis-
Advancing fetal brain MRI persion of randomly moving water protons when a magnetic field gradient is applied.72 Water moving in a capillary network can be modeled as a fraction of the voxel that is diffused very rapidly. Therefore, the volume of moving blood in the placenta can be measured as that fraction of water that is moving fast enough to be dephased during acquisition of IVIM sequence.72 Preliminary studies applying the IVIM technique to measure placental perfusion have provided promising results.73 Specifically, this technique has demonstrated a band of high moving-blood fraction on the side of the placenta adjacent to the uterine wall.73 In cases of preeclampsia or fetal growth retardation, IVIM studies in late gestation have shown reduced perfusion of the placental basal plate. Conversely, the fraction of moving blood on the fetal side of the placenta is increased in fetuses with intrauterine growth restriction, changes ascribed by the authors to an alteration of the villous structure and a regional decrease in vascular resistance. Although the application of MR imaging techniques for studying placental function are promising, substantially more longitudinal data with these techniques are needed to characterize normal placental perfusion, and to determine whether early changes in placental perfusion are predictive of subsequent fetal compromise.
Development of an In Vivo 3-D Fetal MRI Brain Atlas Although brain atlases are widely used for image processing in adult MRI, no such atlas is currently available for the fetal brain. The development of a 3-D MRI fetal brain atlas would expedite our understanding of normal and aberrant in vivo development of the fetal brain. The fundamental steps in development of an MRI brain atlas involve building a template from the statistical analysis of multiple brain studies in the population of interest.74,75 The basic principles would be
295 similar to those used for the creation of adult brain atlases, including brain segmentation, registration, and averaging techniques of a set of fetal brain volumes to a common reference. However, there would be important differences in approach extending beyond the technical considerations described earlier. Perhaps the most fundamental difference is the dynamic anatomic changes that characterize normal brain development. In the early second trimester, for example, the surface of the cerebral cortex is normally smooth (lissencephalic) with primary sulci emerging between 18 and 20 weeks’ gestation and secondary sulci around 30 to 32 weeks. Consequently, a meaningful fetal brain MRI atlas would actually consist of multiple gestational age-dependent atlases that reflect these and multiple other evolving features of the developing brain. This is no trivial challenge given the need for a large-scale normative fetal MRI database that spans across all gestational ages. Recent innovations in fetal image reconstruction using advanced registration methods to correct for fetal motion5,76,77 have allowed acquisition of high-resolution 3D-MRI images of the fetal brain, and will facilitate the creation of a fetal brain atlas. Development of such a 3D-MRI fetal brain atlas extending from 25 to 37 weeks of gestation is in preparation,78 based on a large normative database of nonsedated fetal MRI studies (Fig. 7). The availability of a 3-D fetal MRI brain atlas will assist in the development of advanced post-processing techniques for fetal volumetric brain reconstruction and automatic volume rendering of fetal brain structures and tissue types (e.g., cortical grey matter, unmyelinated and myelinated white matter, germinal matrix, cerebrospinal fluid) for the future study of regional tissue-specific in utero brain development. In addition to volumetric studies of the major brain structures, this approach may ultimately allow volume measurements of tissue subtypes (eg, cortical gray matter, unmyelinated and myelinated white matter, cerebrospinal
Figure 7 The development of a high-resolution in vivo fetal MRI brain atlas. The top row represents the first atlas (mean gestational age of 27 weeks). The bottom row illustrates the second fetal atlas (mean gestational age of 33 weeks).
C. Limperopoulos and C. Clouchoux
296
Figure 8 Extracted cortical surface from a 28-week gestational age fetus. Top row shows right and left cerebral hemispheres, viewed from 3 different angles. Bottom row shows the left hemisphere (external and internal surfaces).
fluid) in specific parcellated brain regions, facilitating the study of regional tissue-specific brain development. The developmental schedule for gyral formation in the developing fetus has been well defined through both pathology and qualitative MRI studies.78-80 Quantitative techniques for measurement of cortical folding previously described in adults, have recently been applied to preterm infants ex utero.81 An extension of this approach to the study of fetal sulcal development is currently underway, using a novel methodological framework for fetal cortical surface rendering (Fig. 8) (Clouchoux et al, unpublished data). This approach is based on a technique previously validated in adults using a 2-D mesh of the cortical surface brains.82-85 Other potential future developments might include cortical thickness measurement and volume-based morphometry.
Summary and Future Directions Fetal MRI is rapidly becoming a powerful tool for investigating in vivo fetal brain development in both clinical and research studies. To date, our understanding of the dynamic and highly intricate evolution of the fetal brain has been impeded by a lack of standardized and reliable MRI benchmarks. The recent development and ongoing refinement of quantitative 3-D volumetric MRI, DTI, MR spectroscopy, and functional and perfusion imaging now provide us with the opportunity to evaluate the fetal brain from a complementary, integrated perspective to include volume, microstructure, metabolism, and function, respectively. These timely methodological developments now provide us with an unlimited potential to study in unprecedented detail normal fetal brain development as well as the mechanisms and consequences that underlie impaired brain development in the compromised fetus. Advancing the role and accuracy of MRI in the fetus will require ongoing methodological developments in the areas of image acquisition and advanced postprocessing techniques to overcome current challenges of fetal motion and image
degradation. Moreover, the creation of a large-scale highresolution 3-D in vivo fetal brain atlas will permit greater insight into neuronal and axonal pathway development, as well as sulcal and gyral formation, and provide a unique window on the timing of insults that might disrupt normal brain development. Advances expected to result from this work will greatly improve our ability to reliably detect and monitor the high-risk fetus, improve parental counseling, and evaluate the potential benefits of prenatal interventions.
References 1. Garel C: Fetal MRI: What is the future? Ultrasound Obstet Gynecol 31:123-128, 2008 2. Prayer D: Investigation of the normal organ development with fetal MRI. Eur Radiol 17:2458-2471, 2006 3. Jenkinson M, Smith S: Improved optimization for the robust an accurate linear registration and motion correction of brain images. Neuroimage 17:825-841, 2002 4. Jiang S, Xue H, Counsell S, et al: In-utero three dimension high resolution fetal brain diffusion tensor imaging. MICCAI 10:18-26, 2007 5. Rousseau F, Glenn OA, Iordanova B, et al: Registration-based approach for reconstruction of high-resolution in utero fetal MR brain images. Acad Radiol 13:1072-1081, 2006 6. Garel C, Delezoide: MRI of the Fetal Brain: Normal Development and Cerebral Pathologies. New York, NY, Springer, 2004 7. Girard N, Gire C, Sigaudy S, et al: MR imaging of acquired fetal brain disorders. Childs Nerv Syst 19:490-500, 2003 8. Limperopoulos C, Robertson RL, Estroff JA, et al: Diagnosis of inferior vermian hypoplasia by fetal magnetic resonance imaging: Potential pitfalls and neurodevelopmental outcome. Am J Obstet Gynecol 194: 1070-1076, 2006 9. Limperopoulos C, Robertson RL Jr, Khwaja OS, et al: How accurately does current fetal imaging identify posterior fossa anomalies? Am J Roentgenol 190:1637-1643, 2008 10. Sjogren B: Psychological indications for prenatal diagnosis. Prenat Diagn 16:449-454, 1996 11. Manjon JV, Guizard N, Robles M, et al: Reconstruction using 3D inpainting. Hum Brain Mapp (in press) 12. Limperopoulos C, Soul JS, Gauvreau K, et al: Late gestation cerebellar growth is rapid and impeded by premature birth. Pediatrics 115:68895, 2005 13. Hüppi PS, Warfield S, Kikinis R: Quantitative magnetic resonance im-
Advancing fetal brain MRI
14.
15.
16.
17.
18.
19.
20.
21.
22.
23. 24. 25. 26.
27.
28.
29.
30. 31. 32.
33. 34.
35. 36.
37.
38.
aging of brain development in premature and mature newborns. Ann Neurol 43:224-235, 1998 Kubik-Huch RA, Wildermuth S, Cettuzzi L, et al: Fetus and Uteroplacental unit: Fast MR imaging with three-dimensional reconstruction and volumetry—Feasibility study. Radiology 219:567-73, 2001 Schierlitz L, Dumanli H, Robinson JN, et al: In-utero three-dimensional magnetic resonance imaging of fetal brains. Lancet 357:1177-1178, 2001 Grossman R, Hoffman C, Mardor Y, et al: Quantitative MRI measurements of human fetal brain development in utero. Neuroimage 33:463470, 2006 Kazan-Tannus JF, Dialani V, Kataoka ML, et al: MR volumetry of brain and CSF in fetuses referred for ventriculomegaly. Am J Roentgenol 189:145-151, 2007 Limperopoulos C, Tworetzky W, Newburger JW, et al: Third-trimester volumetric brain growth is impaired in fetuses with congenital heart disease. Circulation 118:S908, 2008 Guizard N, Evans AC, Lepage C, et al: Automatic model-based fetal brain parcellation to quantify in vivo fetal brain development. Hum Brain Mapp (in press) Mukherjee P, Miller JH, Shimony JS, et al: Diffusion-tensor MR imaging of gray matter and white matter development during normal human maturation. Am J Neuroradiol 23:1445-1456, 2002 Anjari M, Srinivasan L, Allsop JM, et al: Diffusion tensor imaging with tract-based spatial statistics reveals local white matter abnormalities in preterm infants. Neuroimage 35:1021-1027, 2007 Dubois J, Dehaene-Lambertz G, Soarès C, et al: Microstructural correlates of infant functional development: Example of the visual pathways. J Neurosci 28:1943-1948, 2008 Stadnik T, Luypaert R, Jager T, et al: Diffusion imaging: From basic physics to practical imaging. RSNA EJ Radiog 3, 1999 Prayer D, Prayer L: Diffusion-weighted magnetic resonance imaging of cerebral white matter development. Eur J Radiol 45:235-243, 2003 Garel C: New advances in fetal MR neuroimaging. Pediatr Radiol 36: 621-625, 2006 Baldoli C, Righini A, Parazzini C, et al: Demonstration of acute ischemic lesions in the fetal brain by diffusion magnetic resonance imaging. Ann Neurol 52:243-246, 2002 Righini A, Bianchini E, Parazzini C, et al: Apparent diffusion coefficient determination in normal fetal brain: A prenatal MR imaging study. Am J Neuroradiol 24:799-804, 2003 Agid R, Lieberman S, Nadjari M, et al: Prenatal MR diffusion-weighted imaging in a fetus with hemimegalencephaly. Pediatr Radiol 36:138140, 2006 Bui T, Daire JL, Chalard F, et al: Microstructural development of human brain assessed in utero by diffusion tensor imaging. Pediatri Radiol 36:1133-1140, 2006 Kim DH, Chung S, Vigneron DB, et al: Diffusion-weighted imaging of the fetal brain in vivo. Magn Reson Med 59:216-220, 2008 Brunel H, Girard N, Confort-Gouny S, et al: Fetal brain injury. J Neuroradiol 31:123-137, 2004 Erdem G, Celik O, Hascalik S, et al: Diffusion-weighted imaging evaluation of subtle cerebral microstructural changes in intrauterine fetal hydrocephalus. Magn Reson Imaging 25:1417-1422, 2007 Le Bihan D: Molecular diffusion nuclear magnetic resonance imaging. Magn Reson Q 7:1-30, 1991 Hermoye L, Saint Martin C, Cosnard G, et al: Pediatric diffusion tensor imaging: Normal database and observation of the white matter maturation in early childhood. Neuroimage 29:493-504, 2005 Kasprian G, Brugger PC, Weber M, et al: In utero tractography of fetal white matter development. Neuroimage 43 213-24, 2008 Kok RD, van der Bergh AJ, Heerschap A, et al: Metabolic information from the human fetal brain obtained with proton magnetic resonance spectroscopy. Am J Obstet Gynecol 185:1011-1015, 2001 Kok RD, van der Berg PP, van der Bergh AJ, et al: Maturation of the human fetal brain as observed by 1H MR spectroscopy. Magn Reson Med 48:611-616, 2002 Heerschap A, Kok RD, van der berg PP: Antenatal proton MR spectroscopy of the human brain in vivo. Childs Nerv Syst 19:418-421
297 39. Kreis R: Brain metabolite composition during early human brain development as measured by quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med 48:949-958, 2002 40. Girard N, Gouny SC, Viola A, et al: Assessment of normal fetal brain maturation in utero by proton magnetic resonance spectroscopy. Magn Reson Med 56:768-775, 2006 41. Roelant-Van Rijn AM, Groenendaal F, Stoutenbeek P, et al: Lactate in the fetal brain: Detection and implications. Acta Pædiatr 93:937-940, 2004 42. Van der Voorn JP, Pouwels PJ, Hart AA, et al: Childhood white matter disorders: Quantitative MR imaging and spectroscopy. Radiology 241: 510-517, 2006 43. Fenton BW, Lin CS, Macedonia C, et al: The fetus at term: In utero volume-selected proton MR spectroscopy with breath-hold technique—A feasibility study. Radiology 219:563-566, 2001 44. Girard N, Fogliarini C, Viola A, et al: MRS of normal and impaired fetal brain development. Eur J Radiol 57:217-225, 2006 45. Borowska-Matwiejczuk K, Lemancewicz A, Tarasow E, et al: Assessment of fetal distress based on magnetic resonance examinations: Preliminary report. Acad Radiol 10:1274-1282, 2003 46. Wolfberg AJ, Robinson JN, Mulkern R, et al: Identification of fetal cerebral lactate using magnetic resonance spectroscopy. Am J Obstet Gynecol 196:e9-e11, 2007 47. Azpurua H, Alvarado A, Mayobre F, et al: Metabolic assessment of the brain using proton magnetic resonance spectroscopy in a growth-restricted human fetus: Case report. Am J Perinatol 25:305-309, 2008 48. Limperopoulos C, Tworetzky W, Robertson RL, et al: Impaired brain metabolism in fetuses with congenital heart disease. Circulation 118: S651-S-652, 2008 49. Hirsch J, Kim KHS, Souweidane MM, et al: fMRI reveals a developing language system in a 15-month old sedated infant (Abstract). Soc Neurosci 23: 222, 1997 50. Huettel SA, Song AW, McCarthy G: Functional Magnetic Resonance Imaging. Sunderland, MA, Sinauer Associates, 2004 51. Iannetti GD, Wise RG: BOLD functional MRI in disease and pharmacological studies: Room for improvement? Magn Reson Imaging 25: 978-988, 2007 52. Gowland P, Fulford J: Initial experiences of performing fetal fMRI. Exp Neurol 190 (suppl 1):S22-S27, 2004 53. Fulford J, Vadeyar SH, Dodampahala SH, et al: Fetal brain activity and hemodynamic response to a vibroacoustic stimulus. Hum Brain Mapp 22:116-121, 2004 54. Fulford J, Vadeyar SH, Dodampahala SH, et al: Fetal brain activity in response to a visual stimulus. Hum Brain Mapp 20:239-245, 2003 55. Sparling JW: Concepts in Fetal Movement Research. New York, Haworth Press, 1993 56. Olesen AG, Svare JA: Decreased fetal movements: Background, assessment, and clinical management. Acta Obstet Gynecol Scand 83:818826, 2004 57. Prechtl HF, Einspieler C: Is neurological assessment of the fetus possible? Eur J Obstet Gynecol Reprod Biol 75:81-84, 1997 58. Sparling JW, Van Tol J, Chescheir NC: Fetal and neonatal hand movement. Phys Ther 79:24-39, 1999 59. Gagnon R: Placental insufficiency and its consequences. Eur J Obstet Gynecol Reprod Biol 110 (suppl 1):99-107, 2003 60. Mihaly GW, Morgan DJ: Placental drug transfer: Effects of gestational age and species. Pharmacol Ther 23:253-266, 1983 61. Polliotti BM, Holmes R, Cornish JD, et al: Long-term dual perfusion of isolated human placental lobules with improved oxygenation for infectious diseases research. Placenta 17:57-68, 1996 62. Duncan KR: Fetal and placental volumetric and functional analysis using echo-planar imaging. Top in Magn Res Imaging 12:52-66, 2001 63. Le Bars E, Gondry-Jouet C, Deramond H, et al: MR diffusion and perfusion imaging in clinical practice. J Neurorad 57:56-76, 2000 64. Salomon LJ, Siauve N, Balvay D, et al: Placental perfusion MR imaging with contrast agents in a mouse model. Radiology 235:73-80, 2005 65. Zhu XP, Li KL, Kamaly-Asl ID, et al: Quantification of endothelial permeability, leakage space, and blood volume in brain tumors using
C. Limperopoulos and C. Clouchoux
298
66.
67.
68.
69. 70.
71.
72. 73.
74.
combined T1 and T2* contrast-enhanced dynamic MR imaging. J Magn Reson Imaging 11:575-585, 2000 Brasch R, Turetschek K: MRI characterization of tumors and grading angiogenesis using macromolecular contrast media: Status report. Eur J Radiol 34:148-155, 2000 Vallee JP, Sostman HD, MacFall JR, et al: Quantification of myocardial perfusion with MRI and exogenous contrast agents. Cardiology 88:90105, 1997 Pradel C, Siauve N, Bruneteau G, et al: Reduced capillary perfusion and permeability in human tumour xenografts treated with the VEGF signaling inhibitor ZD4190: An in vivo assessment using dynamic MR imaging and macromolecular contrast media. Magn Reson Imaging 21:845-851, 2003 Roberts JM, Lain KY: Recent insights into the pathogenesis of preeclampsia. Placenta 23:359-372, 2002 Zygmunt M, Herr F, Munstedt K, et al: Angiogenesis and vasculogenesis in pregnancy. Eur J Obstet Gynecol Reprod Biol 110 (suppl 1):1018, 2003 Gowland PA, Francis ST, Duncan KR, et al: In vivo perfusion measurements in the human placenta using echo planar imaging at 0.5T. Magn Reson Med 40:467-473, 1998 Gowland PA: Placental MRI. Seminars in fetal and neonatal. Medicine 10:485-490, 2005 Moore RJ, Ong SS, Tyler DJ, et al: Spiral artery blood volume in normal pregnancies and those compromised by pre-eclampsia. NMR Biomed 21:376-380, 2008 Evans A, Collins D, Mills S, et al: 3rd statistical neuranatomical models from 305 MRI volumes. Proc. IEEE Nuclear Science Symposium and Medical Imaging Conference 1813-1817, 1993
75. Kochunov P, Lancaster J, Thompson P, et al: An optimised individual target brain in the Talairach coordinates system. Neuroimage 17:922927, 2002 76. Rousseau F, Glenn O, Iordanova B, et al: A novel approach to high resolution fetal brain MR imaging. MICCAI 8:548-555, 2005 77. Guizrad N, Lepage C, Fonov V, et al: Development of fetus brain atlas from multi-axial MR acquisitions. ISMRM. (in press) 78. Rados M, Judas M, Kostovic I: In vitro MRI of brain development. Eur J Radiol 57:187-198, 2006 79. Perkins L, Hughes E, Srinivasan L, et al: Exploring cortical subplate evolution using magnetic resonance imaging of the fetal brain. Dev Neurosci 30:211-220, 2008 80. Salomon LJ, Garel C: Magnetic resonance imaging examination of the fetal brain. Ultrasound Obstet Gynecol 30:1019-1032, 2007 81. Dubois J, Benders M, Cachia A, et al: Mapping the early cortical folding process in the preterm newborn brain. Cereb Cor Advanc 18:14441454, 2007 82. Van Essen D, Drury H, Joshi S, et al: Functional and structural mapping of human cerebral cortex: Solution are in the surfaces. Proc Natl Acad Sci USA 95:788-795, 1998 83. Fischl B, Sereno M, Tootell R, et al: Cortical surface-based analysis. II. Segmentation and surface reconstruction. Neuroimage 9:179-194, 1999 84. Toro R, Burnod Y: Geometric atlas: Modeling the cortex as an organized surface. Neuroimage 20:1468-1484, 2003 85. Clouchoux C, Coulon O, Rivière D, et al, LNCS: Anatomically constrained surface parameterization for cortical localization. MICCAI 3750:344-35, 2005 86. Schneider JF, Confort-Gouny S, Le Fur Y, et al: Diffusion-weighted imaging in normal fetal maturation. Eur Radiol 17:2422-2429, 2007