The Influence of Various Cerebral and Extracerebral Pathologies on Apparent Diffusion Coefficient Values in the Fetal Brain.

BACKGROUND AND
PURPOSE: The changing MRI signal accompanying brain maturation in fetal brains can be quantified on apparent diffusion coefficient (ADC) maps. Deviations from the natural course of ADC values may reflect structural pathology. The purpose of this study was to determine the influence of fetal pathologies on the ADC values in different regions of the fetal brain and their evolution with increasing gestational age.
METHODS: This was a retrospective study of 291 fetuses evaluated between the 14th and the 40th week of gestation using diffusion-weighted imaging (DWI). Fetuses with normal MRI findings but sonographically suspected pathology or fetuses with abnormalities not affecting the brain were analyzed in the control group and compared to fetuses suffering from different pathologies like hydrocephalus/ventriculomegaly, brain malformations, infections, ischemia/hemorrhage, diaphragmatic hernias, and congenital heart disease. Pairwise ADC measurements in each side of the white matter (WM) of the frontal, parietal, and occipital lobes, in the basal ganglia and the cerebellum, as well as a single measurement in the pons were performed and were plotted against gestational age.
RESULTS: In the control group, brain maturation followed a defined gradient, resulting in lower ADC values in the most mature regions. Each disorder group experienced abnormal patterns of evolution of the ADC values over time deviating from the expected course.
CONCLUSIONS: The ADC values in different regions of the fetal brain and their evolution with increasing gestational age are influenced by pathologies compromising the cerebral maturation.

Introduction

MRI has proved to be a valuable method in the assessment of sonographically suspected fetal pathology. For example, ischemia, cortical malformations, infections, as well as normal brain development can be demonstrated by diffusion‐weighted imaging (DWI). , , , Various studies have shown that the changing MR signal accompanying brain maturation in fetal and neonatal brains can be quantified on apparent diffusion coefficient (ADC) maps, , , , , , whereas water diffusivity decreases with ongoing maturation as a result of progressing myelination with increasing lipid content and tissue density, formation of new synapses, fiber organization into bundles, and decreasing water content. Thus, deep ADC values are expected in cell‐dense, highly organized tissues with little interstitial fluid. , , , , , Therefore, deviations from the natural course of these values may reflect structural pathology. , , Recent studies have found that ADC calculation is reproducible method for radiologists at different gestational ages. , ADC is equivalent to mean diffusivity (MD) and is a rotationally invariant quantity obtained from 3 orthogonal measurements for time‐considerations. , MD is affected by the cellular density of a voxel, or many disease processes like edema, necrosis, or inflammation. ,

Diverse pathological conditions have been studied using fetal brain ADC measurement so far. For example, compared with controls, ADC values in the brain fetuses with severe intrauterine growth restriction have been shown to be different. Changes of ADC values in different fetal brain regions in CMV infection have also been described. ,

The knowledge of normal and pathological MR signal due to changing tissue composition at a given time of the fetal maturation is essential to adequately assess the imaging results, identify pathologies, and, in selected cases, initiate appropriate management.

The aim of this study was to investigate the development of ADC values in different brain areas in singleton fetuses affected by different intra‐ or extracranial pathologies in comparison with fetuses with normal brains.

Methods

Patient Data and Ethical Standards

This retrospective study was performed in accordance with the ethical standards and approved by the Institutional Review Board (Kantonale Ethikkommission für die Forschung Bern, Bern, Switzerland: 2016‐00727) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. An informed consent of the patients was waived due to the retrospective study design. The patients were referred to our department for fetal MR imaging after prenatal sonography suggested intra‐ or extracranial anomalies (eg, suspected brain or spine malformations, hydrocephalus, heart defect, diaphragmatic hernia, urogenital anomalies, etc; see Table 1), to rule out anomalies in cases of suspected prenatal infection or in a few cases – after previous pathological pregnancy.

Table 1

Group Assignment and Characteristics

Group	Number of cases	Mean gestational age (weeks)
1. Control Group	143	28.15
2. Hydrocephalus/Ventriculomegaly	27	28.11
3. Ischemia/Hemorrhage	16	25.63
4. Infections	8	28.38
5. Congenital Diaphragmatic Hernia	14	30.36
6. Congenital Heart Disease	8	32.50

Data regarding maternal health and when available, umbilical artery (UA) Doppler as well as cranial ultrasound findings of the fetuses with pathologies were retrospectively reviewed.

MR Protocol and Post Processing

We retrospectively reviewed all studies including DWI as part of the fetal MR imaging protocol from March 2005 to June 2016. DWI was performed on 1.5 T Magnetom Sonata or Magnetom Aera Siemens scanners (Siemens, Erlangen, Germany) with a single‐shot spin‐echo echo‐planar axial DWI sequence acquired by use of a b‐value of 0, 500, and 700 s/mm2 in 3 orthogonal directions without sedation of the mother or fetus.

The application of gradients oriented in 3 planes in space using b‐values between 0 and 700 s/mm2 was chosen, since it is a part of many fetal imaging protocols, , , , where b‐values between 400 and 700 s/mm2 are considered to demonstrate optimal contrast in the fetal brain.

The ADC maps were automatically calculated from the Siemens software. The following parameters were used: TR 4500 ms; TE 84 ms; FOV 320 mm; matrix 128 × 128 mm; section thickness 5 mm; skip 30% (1.5 mm); and bandwidth 1562 Hz.

The scan time of the diffusion sequence was 68 seconds, with overall examination time between 20 and 30 minutes. Sequences that could not be assessed due to motion or other types of artifacts and of poor image quality were excluded (n = 27 from an initial population of 381 fetuses, with a dropout rate of 8,5%). A total of 75 fetuses with cerebral malformations were identified. Due to the extreme heterogeneity of the pathologies (eg, Dandy‐Walker malformation, agenesis of the corpus callosum, polymicrogyria, tuberous sclerosis, cystic malformations, holoprosencephaly, etc), this group was excluded for further analysis, since no general statement regarding the ADC values based on such inhomogeneous population can be made, resulting in a final population of 216 fetuses.

A total of 11 circular regions of interest (ROIs) were placed on the ADC maps by a fellow radiologist (NSc), always under the supervision of senior pediatric neuroradiologists (CW, NSl) using Olea Sphere® software (Version 2.3, Olea Medical®) as follows: bilaterally in the WM of the frontal, parietal and occipital lobes, in the basal ganglia (BG), in the WM of the cerebellum and 1 ROI in the middle of the pons. ROIs placed in the supratentorial WM included the intermediate zone and the subplate zone (Fig 1). The size of the ROIs was adjusted for gestational age (GA).

Fig 1

Placement of the ROIs on the ADC map. Images from the control group, 33 weeks gestational age. A) paired ROIs in the frontal and parietal white matter (WM); B) paired ROIs in the occipital WM; C) paired ROIs n the basal ganglia; D) single ROI in the middle of the pons; paired ROIs in the cerebellum.

Statistical Analysis

For all statistical analysis, a p‐value ≤ .05 two‐sided was considered statistically significant. The graphics were prepared using Microsoft Excel Software (2013) and SPSS (Version 21; IBM, Armonk, New York). ADC values of the groups with disease were compared to the control group using the nonparametric Mann‐Whitney U‐test due to group size differences. ADC values were correlated with GA using Spearman correlation.

Results

Group Assignment and Characteristics

All fetuses were assigned to 6 groups

A Student's t‐test was performed for each side to compare the ADC values for each area and did not show any significant differences between the values of the right and left hemisphere (all p > .05). Therefore, the mean values for each anatomical ROI location of the 2 hemispheres were included for further analysis, except for the ROI in the middle of the pons, where a single central measurement was performed (see Table 2).

Table 2

Group Assignment and ADC Values

	Mean ADC value [×10−6 mm2 / s] ± SD
	Frontal WM	Parietal WM	Occipital WM	Basal Ganglia	Cerebellum	Pons
Control n = 143	1763.42 ± 221.71	1773.15 ± 205.32	1723.34 ± 213.05	1359.71 ± 190.67	1497.17 ± 183.58	1309.43 ± 175.61
Hydrocephalus/Ventriculomegaly n = 27	1705.09 ± 236.74	1759.69 ± 205.32	1781.81 ± 404.11	1389.57 ± 205.56	1501.25 ± 164.00	1323.56 ± 249.17
Ischemia/Hemorrhage n = 16	1506.51* ± 368.46	1592.07 ± 369.44	1597.72 ± 345.72	1195.18* ± 260.52	1475.26 ± 295.61	1294.75 ± 513.88
Infection n = 8	1869.51 ± 172.34	1986.88* ± 212.47	2021.44* ± 318.60	1446.64 ± 251.85	1589.33 ± 216.76	1362.13 ± 227.51
Congenital diaphragmatic hernia n = 14	1765.99 ± 218.87	1877.21 ± 256.38	1796.07 ± 252.71	1459.31 ± 314.44	1468.85 ± 164.60	1251.07 ± 263.56
Congenital heart disease n = 8	1692.38± 221.71	1765.87 ± 183.88	1713.44 ± 218.99	1319.05 ± 139.37	1412.91 ± 194.76	1229 ± 137.58

Data represent mean ± standard deviation unless otherwise indicated. WM: white matter, n = number of cases, *p < .05 compared to controls (Mann‐Whitney U‐test)

The control group contained 143 fetuses with GA post menstruation between 14 and 40 weeks (mean 28.05 weeks). Of them, 57 had normal MRI findings and 86 had extracerebral pathologies, which were considered not interfering with brain development (for details and main indications for the studies please see Table 3).

Table 3

Main Indications of the Control Fetuses

Indicaton	Number of cases
Evaluation of placenta and uterus	n = 14	– Rule out placenta increta/percreta	n = 13
		– Condition after laser ablation of giant placental chorangioma	n = 1
Evaluation of the brain/ Questionable sonographic abnormalities	n = 43	– Arachnoid cyst, megacisterna magna, promiment cavum septi pellucidi	n = 8
		– Twin pregnancy with single fetal death	n = 3
		– Sonographically suspected malformation, subsequent normal fetal MR imaging results	n = 31
		– Eye globe abnormality	n = 1
Previous pregnancy with brain abnormalities	n = 3
Evaluation of the spine	n = 3	– Spinal malformation	n = 2
		– Myelomeningocele/lipomyelomeningocele without evidence of Chiari malformation	n = 1
Sacrococcygeal teratoma	n = 3
Suspected thoracic pathology	n = 23	– Congenital pulmonary airway malformation /Lung sequester	n = 16
		– Suspected congenital diaphragmatic hernia	n = 5
		– Intrathoracic space‐occupying lesion	n = 2
Abdominal pathology	n = 48	– Kidney abnormalities	n = 31
		– Intestinal pathology	n = 9
		– Abdominal mass	n = 5
		– Abdominal wall defect	n = 3
Others	n = 6	– Sibling with pancreatic hypoplasia	n = 1
		– Situs inversus	n = 1
		– Neck cyst	n = 2
		– Mild hydrothorax	n = 1
		– Clubfoot	n = 1

n = number of cases

The second group contained 27 cases of hydrocephalus (n = 17) or ventriculomegaly (n = 10) at a GA between 22 and 35 weeks (mean, 28.11 weeks).

The group of ischemia and/or hemorrhage contained 16 cases aged between 14 and 35 weeks GA (mean 25.63 weeks). Six fetuses suffered from cerebral ischemia, 9 from hemorrhage, and 1 fetus had both pathologies.

In total, 8 fetuses with congenital infections were examined between the 23rd and 34th gestational week (mean 28.38 weeks GA), of them 3 cases of cytomegalovirus (CMV) seroconversion, 2 cases of toxoplasmosis, and 3 cases with suspected TORCH infection based on radiological signs.

Our study included 12 cases of congenital diaphragmatic hernia (CDH), a case of massive pleural effusion, and a case of a chylothorax, all evaluated in the same group (n = 14) due to the identical effect of the pathology on the heart with mediastinal shift. The fetuses were examined between 24 and 37 weeks GA (mean 30.36 weeks).

In total, 8 fetuses with congenital heart disease (CHD) with a GA between 21 and 37 weeks (mean 32.5 weeks) were included. They suffered from an atrioventricular septal defect (n = 2), hypoplastic left heart syndrome (n = 2), cardiomegaly (n = 1), cardiac rhabdomyoma (n = 1), univentricular heart (n = 1), and a complex cardiac malformation with valve atresias and ventricular septal defect (n = 1).

Our population of fetuses with different intra‐ and extracranial pathologies (n = 73) included fetuses with normal and abnormal UA Doppler. In our series, 17% fetuses presented an abnormal UA Doppler (for details regarding pathological UA Doppler and fetal cranial ultrasound findings as well as maternal risk factors, please see Table 4).

Table 4

Placenta Abnormalities and Maternal Health Factors

	Number of fetuses	Records available	Pathological umbilical Doppler	Abnormal fetal cranial ultrasound	Placental disorder (retroplacental hematoma, infarction)	Maternal health factors
Group	n	n (%)	n (%)	n (%)	n (%)	Diabetes n (%)	Hyper‐tension n (%)	Adipositas n (%)
Hydrocephalus/Ventriculomegaly	27	15 (55.6)	0	0	0	3 (20)	0	2 (13.3)
Ischemia/Hemorrhage	16	15 (93.8)	7 (46.7)	11 (73.3)	4 (26.7)	0	0	0
Infections	8	7 (87.5)	0	4 (50)	0	0	0	0
Congenital diaphragmatic hernia	14	11 (78.6)	2 (18.2)	0	0	0	1 (7.1)	2 (14.2)
Congenital heart disease	8	5 (62.5)	0	0	0	0	0	0

n, number of cases.

Control Group

In the control group, the highest (mean) ADC values were measured in the frontal WM (1763.42 × 10−6 mm2/s), followed by the parietal (1773.15 × 10−6 mm2/s) and occipital WM (1723.34 × 10−6 mm2/s); intermediate ADC values were measured in the BG and the cerebellum, and the lowest ADC values – in the pons (1309.43 × 10−6 mm2/s; Table 2).

The evolution of the ADC values showed a significant increase with progressing GA in the frontal WM (r = .208, p = .01). A decrease was observed in the pons (r = –.254, p = .002) and in the cerebellum (r = –.371, p < .001; Fig 2).

Fig 2

Distribution of the ADC values in different areas of the fetal brain with increasing gestational age in the control group. ADC value: ADC [×10‐6 mm2/s]; GA: gestational age in weeks.

Hydrocephalus/Ventriculomegaly Group

The occipital WM showed the highest ADC values (1781.81 × 10−6 mm2/s), the pons – the lowest (1323.56 × 10−6 mm2/s; Table 2).

The values in the pons remained almost unchanged. All other regions but the cerebellum (with a non‐significant decline) showed a non‐significant increase in ADC values with increasing GA (for all tests: r < .349, p > .07). In the occipital WM the highest increase was found (Fig 3).

Fig 3

Distribution of the ADC values in different areas of the fetal brain with increasing gestational age. GA: gestational age in weeks, MD: mean diffusivity; ADC = MD [×10–6 mm2/s].

Ischemia/Hemorrhage Group

The highest ADC values were found in the occipital WM (1597.72 × 10−6 mm2/s), the lowest – in the BG (1195.18 × 10−6 mm2/s; Table 2). The frontal and parietal WM and the BG showed a non‐significant increase with progressing GA (all r > .325, p > .22); the occipital WM, the cerebellum and the pons – a non‐significant decrease (all r < .48, p > .057).

Infections Group

The highest ADC values were measured in the occipital WM (2021.44 × 10−6 mm2/s), the lowest – in the pons (1362.13 × 10−6 mm2/s; Table 2). The development of ADC values with progressing GA did not show any significant changes in any of the examined regions. A slight increase was observed in the frontal and occipital WM, a slight decrease – in the remaining areas (Fig 3).

Congenital Diaphragmatic Hernia Group

The parietal WM showed the highest ADC values (1877.21 × 10−6 mm2/s), the pons – the lowest (1251.07 × 10−6 mm2/s; Table 2). Except for the parietal WM (r = –.547, p = .04), the values in all other regions showed a non‐significant decrease with progressing GA (Fig 3).

Congenital Heart Disease Group

The fetuses with CHD showed the highest ADC values in the parietal WM (1765.87 × 10−6 mm2/s), the lowest – in the pons (1229.0 × 10−6 mm2/s; Table 2). The course of the values with progressing maturation showed neither significant increase nor decrease (Fig 3).

Discussion

Considering the physiological maturation gradient with progressing myelination, the ADC values in the supratentorial WM are expected to be higher than in the cerebellum and the pons, and higher frontally than posteriorly. The development of the ADC values with progressing GA showed the expected decrease in the cerebellum and the pons, but not in the frontal and occipital WM and in the BG, each with a slight increase with advancing GA. Similar results have also been found in other studies and could be explained by transient laminar compartments of the developing human brain, for example, if the ROIs contain the subplate (a transient layer of the fetal brain, visible between the 22nd and 30th gestational week, and characterized by low cellular density). The radial expansion of the compartments at a specific time of the fetal development may explain the different course of the ADC values in healthy fetuses across studies.

Hydrocephalus is one of the most common congenital anomalies. This disorder, which is characterized with a disturbance of cerebrospinal fluid (CSF) circulation resulting from an obstruction to the ventricular drainage system and CSF flow or a reduced CSF resorption, must be distinguished from ventriculomegaly, caused by loss of tissue, for example, in the case of agenesis of the corpus callosum, Arnold‐Chiari malformations or holoprosencephaly, or as a result of periventricular leukomalacia. , , Furthermore, the pathology can be divided into congenital cases, for example, in cases of Dandy‐Walker malformations or Chiari Type II malformation, and in acquired pathologies, for example, as a result of infections or bleeding. In addition to etiology, early formation, cortical thickness of less than 1.5 mm, and loss of stratification on T2‐ and diffusion‐weighted imaging are expected to be prognostically important. The parenchymal compression caused by increased CSF pressure leads to reduced perfusion and ischemia with reduced ADC values of the cerebral parenchyma. , , ,

In our study, lower ADC values were only found in the frontal WM, while all other values were higher compared to the control group.

The higher values as well as their increase with advancing GA may be explained by the time of imaging: in early stages of hydrocephalus, the ADC values rise due to CSF diapedesis into the parenchyma and decrease again after longer exposure to ischemia. , , Another factor may be a delayed myelination.

Ischemia is a common cause of fetal death. The 20.4% of children who died in utero or during the neonatal period had evidence of prenatal ischemic brain damage in neuropathological studies. The assessment of acute hypoxic lesions in adults is 1 of the main application of diffusion imaging. The resulting cytotoxic edema restricts the extracellular diffusion and the ADC values decrease.

The pathophysiological changes following hypoperfusion/hypoxia explain the lower ADC values in all brain regions but pons and cerebellum in this groupcompared to controls. The transient laminar compartments of the developing WM explain their specific spatial and temporal vulnerability since those structures that are in a phase of intense growth at the time of a hypoxic‐ischemic or hemorrhagic event are particularly sensitive. The lowest ADC values in the BG can be explained by their selective vulnerability to ischemia‐hypoxia. The developing WM also experiences specific spatial and temporal vulnerability. , , , , The expected decrease of the ADC values did not appear in the frontal and parietal WM and in the BG. A possible explanation is that it could be due to a more advanced tissue damage to these regions.

TORCH infections can lead to microcephaly, cerebellar hypoplasia, ventriculomegaly/hydrocephalus, calcifications, cysts, hemorrhage, and disorders of cortical development. , , ,

The ADC values of the fetal brains in this group were higher in all regions than in the control group, especially in the parietal and occipital WM. Three fetuses were examined for CMV seroconversion. This virus is transmitted intrauterinely in 2.5% of all newborns and may lead to microcephaly, cerebellar hypoplasia, ventriculomegaly, and disorders of cortical development during the first and second trimesters due to affection of the subventricular and ventricular zones, the origin of the neuronal and glial cells. Subependymal cysts, calcifications, and signal changes in the temporal lobe have typically been described in CMV infections.

Other fetuses in this group of our study suffered from hepatomegaly with or without ascites. In contrast to the study by Yaniv et al. who showed reduced ADC values in all brain regions of CMV infected fetuses, our population showed higher values compared to the control group. We hypothesize that this phenomenon could be explained by the time of intrauterine infection (early or late) or may represent chronic damage and WM gliosis, since a decrease in ADC values due to hypercellularity, inclusion bodies, and cytotoxic edema due to damage of the umbilical cord by the virus is expected only in the acute infection setting. Similar results as in the aforementioned study were obtained in the analysis of the development of ADC values with progressing gestational age, demonstrating decreasing ADC values in all regions except for the frontal and occipital WM.

Another fetus of our study suffered from a toxoplasmosis infection. On imaging, agenesis of the corpus callosum, microcephaly, and a delay of the cerebral maturation was found. In another fetus, an infection with the same pathogen was suspected due to the presence cerebral lesions (germinal matrix bleeding, hydrocephalus, focal gyration disorder/polymicrogyria). Characteristic cerebral lesions of the fetal toxoplasmosis infection are reduction of the parenchymal volume, calcifications of the BG, thalami, in the periventricular parenchyma, the cortex, the meninges, and the eyes. , As a consequence of obliteration of the mesencephalic aqueduct, hydrocephalus may occur.

CDH has an incidence of 1 in 2,500 live births, with predominant left‐sided defects (85%). The intrathoracic herniation of abdominal organs can lead to a shift of the mediastinum, the lungs, and the heart, with heart compression and compromise of the cerebral blood flow. Depending on the size of the defect, the degree of lung hypoplasia and pulmonary hypertension, it is associated with high morbidity and mortality. , In previous studies, brain pathologies such as cortical malformations, intracranial bleeding, dilated external CSF spaces, and WM damage were detected in up to 60% of children with CDH. , , , , The increased incidence of PVL in children with congenital diaphragmatic hernia is indicative of the high vulnerability of the WM in this population. The neurological damage could be attributed to the significantly reduced cerebral blood flow.

In none of the fetuses with CDH examined in our study obvious morphological abnormalities were seen. This may be due to the timing of imaging before birth, as newborns who are postnatally treated with extracorporeal membrane oxygenation (ECMO), inotropic drugs, mechanical ventilation or a gastric tube, or are not orally fed, have an increased incidence of sequelae or neurolodevelopmental disorders, , , “…likely as a result of treatments and therapies.”

CHD are accompanied by high morbidity and global developmental disorders. Newborns with CHD show changes in cerebral microstructure as a result of the deficient intrauterine oxygen and substrate supply. , Many of these children show brain damage before corrective or palliative cardiac surgery, such as ischemia, intracranial hemorrhage, WM injury, delayed brain maturation, and decreased brain volume. , , , ,

The similarity of the pathologies to that seen in premature infants suggests comparable vulnerability due to immaturity of the brain tissue with concomitant increased vulnerability of the oligodendrocyte progenitor cells. , , Fetuses with CHD show a significant increase of water diffusivity in the WM tracts. We observed similar changes, with an increase in the ADC values in all regions of the WM as well as in the BG. None of the fetuses in this group showed morphological brain abnormalities, but the development of ADC values deviating from the control group is suggestive of abnormal brain maturation.

Some limitations have to be mentioned. The direct comparison of studies on the development of ADC values during the brain maturation process is limited due to the use of different MR systems and the different study populations with regard to their size and GA. The statistical power of our study is limited by the slab thickness and the resultant low spatial resolution of the ADC maps. As known, in some cases the quality of the DWI is also influenced by fetal or maternal motion. In our study, in the group of fetuses with CHD, only a small number of pathologies were present, so no general statement about the influence of congenital heart defects on the cerebral development can be made. Infants with CDH can suffer from brain damage resulting from postnatal therapies, which influence is not shown by prenatal imaging. The prognostic value of the observed ADC changes in the different groups of our study is still unknown, further studies and correlation with perinatal outcome are needed.

Regarding ADC estimation, the b‐values for diffusion weighting have to be carefully chosen. Using a baseline b‐value of 50 s/mm² in diffusion‐weighted imaging instead of b = 0 in future studies may be more appropriate in order to eliminate contamination of the images from microperfusion. Besides, the selection of b‐values could be optimized and should be made depending on the expected MD of the parenchyma.

To summarize, fetal diffusion MR imaging is a feasible method of detecting abnormal brain diffusivity associated with early microstructural impact of different pathologies on the developing brain.

Both the ADC values and their evolution curve in diseased fetuses are influenced and deviate from the expected course. Further in utero studies by using advanced MR imaging techniques with larger number of fetal populations in the particular disease groups are needed.