Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2373041625
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wedegärtner, U.
Right arrow Articles by Schröder, H. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wedegärtner, U.
Right arrow Articles by Schröder, H. J.
(Radiology 2005;237:919-926.)
© RSNA, 2005

Experimental Studies

Fetal Sheep Brains: Findings at Functional Blood Oxygen Level–Dependent 3-T MR Imaging—Relationship to Maternal Oxygen Saturation during Hypoxia1

Ulrike Wedegärtner, MD, Mikhail Tchirikov, MD, Sebastian Schäfer, BS, Andrew N. Priest, PhD, Michael Walther, MD, Gerhard Adam, MD and Hobe J. Schröder, MD

1 From the Department of Diagnostic and Interventional Radiology (U.W., S.S., A.N.P., M.W., G.A.), Department of Obstetrics and Prenatal Medicine (M.T.), and Division of Experimental Gynecology, Department of Obstetrics and Prenatal Medicine (H.J.S.), University Medical Center Hamburg-Eppendorf, Martinistrasse 52, 20251 Hamburg, Germany. Received September 21, 2004; revision requested November 24; revision received December 21; accepted January 12, 2005. Supported by Deutsche Forschungsgemeinschaft We 2826/1-1. Address correspondence to U.W. (e-mail: wedegaer{at}uke.uni-hamburg.de).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
PURPOSE: To quantify the dependence of the signal intensity (SI) at blood oxygen level–dependent (BOLD) magnetic resonance (MR) imaging of fetal sheep brains on maternal oxygen saturation and to investigate the influence of positions of regions of interest (ROIs).

MATERIALS AND METHODS: All experimental protocols were reviewed and approved by the local authorities on animal protection. The brains of singleton fetuses of five anesthetized sheep were subjected to rapid sequences (single-shot echo-planar imaging) of BOLD measurements with a 3-T MR imaging unit. Maternal oxygen saturation and heart rate were recorded continuously. After a normoxic phase, hypoxia was induced by reducing the oxygen in a ventilated gas mixture. ROIs were placed in the cerebrum at a reference level and in the cerebellum. Normalized BOLD SI values were calculated from the mean values of steady-state BOLD SIs at the control (SIc) and hypoxic (SIh) plateaus as follows: normalized BOLD SI = (SIh/ SIc) · 100. Normalized BOLD SI values were correlated with maternal oxygen saturation, and linear regression (slope) analysis was performed. Additionally, ROIs were varied in section level and position. Differences in normalized BOLD SI values for ROI placements were calculated by using analysis of variance. A t test was performed to evaluate differences.

RESULTS: Mean maternal oxygen saturation (as the percentage of oxygen in the blood) was 88% (95% confidence interval [CI]: 80%, 96%) in the control period. During hypoxia, it was reduced to 62% (95% CI: 50%, 75%), while fetal normalized BOLD SI decreased to 64% (95% CI: 44%, 85%) in the cerebrum and 56% (95% CI: 32%, 80%) in the cerebellum. Correlations between normalized BOLD SI values and maternal oxygen saturation were as follows: r2 = 0.84 and slope = 1.27 (95% CI: 1.17, 1.36) in the cerebrum and r2 = 0.83 and slope = 1.54 (95% CI: 1.44, 1.63) in the cerebellum. Normalized BOLD SI was 4% lower in the section above the reference level. Variations in normalized BOLD SI for different ROI positions ranged between 0% and 12%.

CONCLUSION: The depletion of oxygen supply is reflected by decreases in fetal brain BOLD SIs that are more distinct in the cerebellum than in the cerebrum. Normalized BOLD SI is influenced only slightly by ROI position.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Intrauterine growth restriction is associated with an increased risk of perinatal mortality, morbidity, and impaired neurodevelopment (1–3). Growth-restricted newborns remain smaller than their counterparts of equal age throughout childhood (4) and have an increased risk of developing hypertension, coronary heart disease, and stroke in adult life (5). Uteroplacental insufficiency, resulting in a poor supply of nutrients and oxygen to the fetus, is thought to be a major cause of intrauterine growth restriction (6–8). Hence, methods of assessing fetal tissue oxygenation to enable detection of the fetus at risk and the option of initiating precautionary or therapeutic measures (eg, induction of preterm delivery) would be desirable. Doppler ultrasonography and cardiotocography, the main diagnostic tools for monitoring fetal well-being, yield indirect data on fetal tissue oxygenation and are of limited value (9,10). There is currently no noninvasive method that provides direct information on fetal oxygenation.

The blood oxygen level–dependent (BOLD) effect of magnetic resonance (MR) imaging is based on the principle that deoxyhemoglobin molecules are paramagnetic particles, whereas oxyhemoglobin is diamagnetic (11). Changes in the concentration of deoxyhemoglobin within blood alters the signal intensity (SI) on T2*-weighted MR images (11). In principle, decreasing pixel brightness on MR images corresponds to increasing deoxyhemoglobin concentrations (11,12). At present, functional BOLD MR imaging is used mainly to depict brain activation due to, for example, acoustic stimulation in the adult brain (11,13–15). It has been shown recently that functional MR imaging is capable of depicting changes in fetal tissue oxygenation during hypoxia (16) and while the mother breathes oxygen-enriched air (17). Thus, BOLD MR imaging may be a tool for noninvasively measuring the state of tissue oxygenation in the fetus. Before fetal BOLD MR imaging can be used as a test modality for fetal surveillance, basic research is necessary to develop methodologic standards for image evaluation and to compare fetal BOLD SI changes with defined changes in maternal or fetal tissue oxygenation. Thus, the purpose of our study was to quantify the dependence of the fetal brain BOLD SI on maternal oxygen saturation and to investigate the influence of position of regions of interest (ROIs).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
All experimental protocols were reviewed and approved by the local authorities on animal protection.

Animal Preparation
MR imaging measurements were performed in five ewes that were carrying singleton fetuses at gestational ages that ranged from 112 to 133 days (full term, 145 days). The ewes were sedated with 20 mg of diazepam (Ratiopharm, Ulm, Germany) administered intravenously through the external jugular vein and 250 mg of esketamine (Ketanest; Parke-Davis, Karlsruhe, Germany); the dose of the latter drug was repeated as required. The ewes were then intubated and transferred from the animal care facility into the MR imaging unit. During MR imaging, the ewes were artificially ventilated (1%–2% isoflurane [Baxter, München-Unterschleissheim, Germany] in O2/N2O; tidal volume, 600–1000 mL; 12–14 cycles per minute).

Maternal oxygen saturation (as the percentage of oxygen in the blood) and heart rate (in beats per minute) were monitored with a pulse oximeter system (Nonin Medical, Plymouth, Minn). The sensor was attached to the maternal tongue and connected to the system by means of a 9-m glass fiber cable. The pulse oximetry data were stored (sampling rate, 0.39 datum per second) along with time information on a personal computer with a data acquisition system (WP 100; Biopac Systems, Santa Barbara, Calif). At the beginning of each experiment, the MR imaging system and the personal computer were synchronized. The respirator, the data acquisition system, and the pulse oximeter were situated outside the imaging room. The animal was connected to the respirator by using rigid pipes (total length, 8 m) and standard anesthesia hoses that were assembled for the experiments. All animals were examined in a lateral position.

MR Imaging
MR imaging was performed once in each animal by using a 3.0-T MR imaging unit (Intera; Philips Medical Systems, Best, the Netherlands) and a six-element phased-array cardiac coil for signal reception. Initially, T2-weighted turbo-spin-echo sequences (repetition time msec/echo time msec, 4454/120; 294 x 256 matrix) were acquired in the fetus for general orientation. Stacks of sections were manipulated with the aim of examining the fetal head in the transverse orientation. For BOLD imaging, T2*-weighted single-shot gradient-echo echo-planar sequences (echo time, 45 msec; field of view, 180 mm; acquired spatial resolution, 2.25 x 2.25 mm; reconstructed spatial resolution, 1.4 x 1.4 mm; acquired matrix, 80 x 80; reconstructed matrix, 128 x 128; section thickness, 5 mm; intersection gap, 1 mm; flip angle, 90°) were used. Saturation bands were used to minimize fold-over artifacts in the phase-encoding direction. To reduce the likelihood of motion artifacts caused by maternal ventilation, measurements were triggered by a pressure sensor attached to the ewe's thorax and were performed during expiration.

For each BOLD examination, a total of 250–300 consecutive BOLD measurements with a stack of seven to 13 sections were acquired at intervals of 8–16 seconds, depending on the maternal respiration rate, for a total sampling time of 30–40 minutes. During BOLD imaging, the MR imaging software enabled observations of the time course of the spatial mean gray-scale value within a selected ROI. It was therefore possible to determine during an experiment when stable values of fetal BOLD SI during hypoxia had been reached. MR imaging measurements were performed once in each fetus.

For each BOLD measurement, the MR imaging system generated a 5-msec spike signal that was also stored by the data acquisition system. This allowed the precise temporal alignment of maternal and MR imaging data (see below).

Hypoxia Protocol
Each protocol consisted of a control period and one level of hypoxia; BOLD images were acquired continuously. During the control period, maternal oxygen saturation was held at 100% with a mixture of 0.64 L of O2 (95% confidence interval [CI]: 0.37, 0.91) and 0.90 L of N2O (95% CI: 0.38, 1.42) in the ventilated gas. Hypoxia was induced by reducing the oxygen in the ventilated gas mixture (to 0.24 L of O2 [95% CI: 0.19, 0.29] with 1.10 L of N2O [95% CI: 0.82, 1.38]). During hypoxia, the gas mixture of O2 and N2O was adjusted as necessary (to 0.44 L of O2 [95% CI: 0.37, 0.51] with 1.00 L of N2O [95% CI: 0.56, 1.44]) to keep maternal oxygen saturation at an approximately constant level of 60%–80%. In all experiments, a considerable time delay (up to 10 minutes) between the beginning of oxygen reduction and a detectable decrease in maternal oxygen saturation was observed. When stable values of fetal BOLD SI during hypoxia were reached (on average, 3–5 minutes after the beginning of the maternal hypoxic plateau), imaging continued for another 5 minutes.

Data Analysis
MR imaging data were exported from the MR imaging system in Digital Imaging and Communications in Medicine format for further analysis. A program was developed in the programming language Delphi (Borland, Scotts Valley, Calif) that allowed the user to measure the time course of the BOLD SIs within a circular or polygonal ROI of arbitrary size (measured in pixel numbers). ROIs were placed in the cerebrum and cerebellum. With visual inspection, the position of the ROI inside the brain during the time series was monitored. ROI data (ie, the mean and standard deviation of the SI and the time of measurement) for the time series were exported as text files (MR imaging data) and saved.

With the ACQ program, version 3.2 (Biopac Systems), the maternal oxygen saturation and heart rate recordings corresponding to the MR imaging data were selected. On average, the number of data points in the control or hypoxic phase was 24 (range, 14–48), which is equivalent to time periods of 4–6 minutes. Data were exported with time information as text files (ACQ data). Another program was developed by using Delphi to automatically merge MR imaging and ACQ data files. Because BOLD images and maternal data were not collected at exactly the same time points, maternal data were linearly interpolated from the mean of two values immediately preceding or following acquisition of the BOLD image and were assigned the time of the BOLD image. The merged and adjusted (time-matched) data sets were saved as text files, and statistical analyses (Statistica; Statsoft, Tulsa, Okla) were based on these data.

Normalization of BOLD SI
To enable comparison between the results of different examinations, BOLD SI during hypoxia was expressed as normalized BOLD SI—a percentage of the mean control values—and calculated with the following equation: normalized BOLD SI = (SIh/SIc) · 100, where SIh is the SI during hypoxia and SIc is the SI during the control period. The following procedure was used: The initial control phase and the subsequent steady-state hypoxic phases were identified on a scatterplot. Because BOLD measurements were performed every 10–15 seconds, temporal mean values (for BOLD SIs in the ROI, maternal oxygen saturation, and maternal heart rate) were derived for the selected phases.

ROI Placement
ROIs were placed in the cerebrum (in both hemispheres and in the left and right hemisphere separately) and in the cerebellum (in both cerebellar hemispheres). Cerebral normalized BOLD SI corresponds to ROI measurements in both cerebral hemispheres. ROI placement was performed by one author (U.W.), who had 5 years of experience in radiology.

For ROI placement in the cerebrum, a reference section was defined according to the following criteria: There had to be at least one adjacent section that depicted the fetal brain above and below the reference section and it had to permit a reasonable ROI size. To reduce data variation, we set the number of pixels within a cerebral ROI to greater than 100 because we had observed that standard deviations of the ROI mean values increased distinctly at lower ROI sizes (results not shown). The average ROI size was 405 pixels (95% CI: 120, 691) in both hemispheres of the cerebrum, 194 pixels (95% CI: 85, 303) in the left hemisphere, and 202 pixels (95% CI: 75, 329) in the right hemisphere. To avoid the influence of data in the skull base, the second or third cranial section through the cerebrum in a transverse orientation at the level of the basal ganglia and the lateral ventricles was typically chosen.

For measurements in the cerebellum, the section on which the largest ROI possible could be drawn was selected. ROI measurements in the cerebellum were mostly caudal of the reference section and cranial to the pons in a transverse orientation. The average ROI size was 60 pixels (95% CI: 17, 103) for the cerebellum.

Additionally, ROI placements were varied so that we could assess the influence of ROI position on the results. ROIs were also placed in the sections immediately above and below the reference sections and at defined positions within the reference section (a circular ROI of 37 pixels was placed within the reference section at three different positions—frontal, temporal, and occipital—in each of the hemispheres, for a total of six positions). This process was important in enabling us to examine how the normalized BOLD SI changed when ROIs were not positioned in the same anatomic levels of the brain—for example, if an ROI was positioned one or two sections cranial or caudal to the reference section or if the position of an ROI changed within a section because of fetal motion.

Statistical Analysis
Statistica was used for statistical analysis. Mean values and 95% CIs of the normalized BOLD SI and the maternal oxygen saturation and heart rate in the control and hypoxic phases were calculated. Linear regression analysis and correlation of normalized BOLD SI and maternal oxygen saturation were performed by using the control and hypoxic values in each experiment. For analysis of the effect of variation of ROI placement, differences in normalized BOLD SI between the ROI settings were calculated by using analysis of variance and expressed as percentage differences. A t test was used to evaluate differences. Differences were considered to be significant when P was less than .05.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Table 1 shows results of the experiment in each animal.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Normalized BOLD SI and Maternal Oxygen Saturation and Heart Rate Values at Each Experiment

 
Changes in Fetal BOLD SI, Maternal Oxygen Saturation, and Maternal Heart Rate during Hypoxia
Figure 1 shows the position of an ROI in the reference section, and Figure 2 shows an example of a typical experiment (in animal 3). After oxygen reduction, maternal blood oxygen saturation started to decline from its control value of about 92%, and BOLD SIs decreased concomitantly (Fig 2a, 2c). Maternal blood oxygen saturation leveled off at nearly 50%, and the BOLD SI reached plateau values 3–4 minutes later. As is typical during an acute hypoxic insult, maternal heart rate rose to 125 beats per minute.



View larger version (170K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. (a, b) Transverse T2*-weighted single-shot gradient echo echo-planar BOLD MR images (echo time, 45 msec; field of view, 180 mm; acquired resolution, 2.25 x 2.25; reconstructed resolution, 1.4 x 1.4 mm; acquired matrix, 80 x 80; reconstructed matrix, 128 x 128; section thickness, 5 mm; intersection gap, 1 mm; and flip angle, 90°) of the fetal cerebrum in animal 3 acquired (a) during the control period with a maternal oxygen saturation of 92% and (b) during the hypoxic plateau a maternal oxygen saturation of 53%. The SI decrease during hypoxia is clearly visible. An ROI that includes both hemispheres is positioned in the cerebrum (dotted line).

 


View larger version (168K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. (a, b) Transverse T2*-weighted single-shot gradient echo echo-planar BOLD MR images (echo time, 45 msec; field of view, 180 mm; acquired resolution, 2.25 x 2.25; reconstructed resolution, 1.4 x 1.4 mm; acquired matrix, 80 x 80; reconstructed matrix, 128 x 128; section thickness, 5 mm; intersection gap, 1 mm; and flip angle, 90°) of the fetal cerebrum in animal 3 acquired (a) during the control period with a maternal oxygen saturation of 92% and (b) during the hypoxic plateau a maternal oxygen saturation of 53%. The SI decrease during hypoxia is clearly visible. An ROI that includes both hemispheres is positioned in the cerebrum (dotted line).

 


View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a. (a–c) Scatterplots of data acquired in animal 3 during an initial control period and hypoxia induced by a reduction in maternal oxygen saturation (MatSO2) show the time course of (a) absolute and (c) relative BOLD SI in both cerebral hemispheres ({bullet}), the left cerebral hemisphere ({square}), the right cerebral hemisphere ({diamond}), and the cerebellum ({blacktriangleup}) and (b) maternal data—that is, maternal oxygen saturation, indicated by the straight line, and maternal heart rate (MatHR), indicated by the dashed line.

 


View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b. (a–c) Scatterplots of data acquired in animal 3 during an initial control period and hypoxia induced by a reduction in maternal oxygen saturation (MatSO2) show the time course of (a) absolute and (c) relative BOLD SI in both cerebral hemispheres ({bullet}), the left cerebral hemisphere ({square}), the right cerebral hemisphere ({diamond}), and the cerebellum ({blacktriangleup}) and (b) maternal data—that is, maternal oxygen saturation, indicated by the straight line, and maternal heart rate (MatHR), indicated by the dashed line.

 


View larger version (35K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2c. (a–c) Scatterplots of data acquired in animal 3 during an initial control period and hypoxia induced by a reduction in maternal oxygen saturation (MatSO2) show the time course of (a) absolute and (c) relative BOLD SI in both cerebral hemispheres ({bullet}), the left cerebral hemisphere ({square}), the right cerebral hemisphere ({diamond}), and the cerebellum ({blacktriangleup}) and (b) maternal data—that is, maternal oxygen saturation, indicated by the straight line, and maternal heart rate (MatHR), indicated by the dashed line.

 
Tables 1 and 2 summarize the major results of all five experiments. Average maternal oxygen saturation and heart rate values over all experiments during the control period were 88% (95% CI: 80%, 96%) and 94 beats per minute (95% CI: 72, 117). During hypoxia, maternal oxygen saturation dropped to 62% (95% CI: 50%, 75%) and maternal heart rate increased to 109 beats per minute (95% CI: 81, 137). Reduction in maternal oxygen saturation during hypoxia resulted in a distinct decrease in normalized BOLD SI, from 100% (control plateau) to 56% (95% CI: 32%, 80%) in the fetal cerebellum and 64% (95% CI: 44%, 85%) in the cerebrum. The decrease in normalized BOLD SI was significantly more pronounced in the cerebellum than in the cerebrum (P < .008). Maternal oxygen saturation measurements were occasionally influenced by the position of the sensor on the maternal tongue; this factor contributed to the variation in regression coefficients between maternal oxygen saturation and normalized BOLD SI.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Overall Normalized BOLD SI and Maternal Oxygen Saturation and Heart Rate Values during Control and Hypoxic Periods

 
Linear Regression Analysis of Changes in Normalized BOLD SI versus Changes in Maternal Oxygen Saturation
Linear regression analysis and correlation of normalized BOLD SI and maternal oxygen saturation data (Table 3) were performed by using the mean values in the control phase (at which the normalized BOLD SI was 100%) and the hypoxic phase. Correlation coefficients were highly significant for cerebral (r = 0.92, P < .001) and cerebellar (r = 0.91, P < .001) normalized BOLD SI. Slopes were significantly larger for the cerebellum than the cerebrum. In the cerebellum, a decrease in normalized BOLD SI of 1.54% resulted from a reduction in maternal oxygen saturation of 1%, whereas in the cerebrum, normalized BOLD SI decreased 1.27% for every 1% reduction in maternal oxygen saturation (Fig 3).


View this table:
[in this window]
[in a new window]

  		TABLE 3. Results of Linear Regression Analysis of Correlations between Normalized BOLD SI and Changes in Maternal Oxygen Saturation

 


View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a. (a, b) Graphs show results of linear regression analysis of normalized BOLD SI versus maternal oxygen saturation (MatSO2) in (a) the cerebrum and (b) the cerebellum. Mean values at the hypoxic plateaus are marked for each animal (animal 1: {diamond}; animal 2: {blacktriangleup}; animal 3: {circ}; animal 4: {blacksquare}; animal 5: {diamondsuit}). Dotted lines indicate the 95% CI of the regression line.

 


View larger version (22K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b. (a, b) Graphs show results of linear regression analysis of normalized BOLD SI versus maternal oxygen saturation (MatSO2) in (a) the cerebrum and (b) the cerebellum. Mean values at the hypoxic plateaus are marked for each animal (animal 1: {diamond}; animal 2: {blacktriangleup}; animal 3: {circ}; animal 4: {blacksquare}; animal 5: {diamondsuit}). Dotted lines indicate the 95% CI of the regression line.

 
Variations in ROI Placement
So that we could investigate the influence of different ROI settings on the normalized BOLD SI, differences in normalized BOLD SI during hypoxia were calculated. For ROI measurements in different section levels, normalized BOLD SI changed between 0% and 4% in the adjacent sections compared with the value in the reference section. That is, the mean difference in the normalized BOLD SI value was close to zero in the section below the reference section and was 4% in the section above the reference section. These differences were not significant.

The mean difference in normalized BOLD SI varied by up to 12% between different ROI positions within the reference section: The mean normalized BOLD SI was 63% in the frontal region of the left hemisphere, 60% in the frontal region of the right hemisphere, 56% in the temporal region of the left hemisphere, 60% in the temporal region of the right hemisphere, 58% in the occipital region of the left hemisphere, and 68% in the occipital region of the right hemisphere. The difference between the left and right occipital regions was significant (P < .038).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Fetal MR imaging is playing an increasing role in prenatal diagnosis. Fetal MR imaging is used mainly in detecting fetal anomalies with morphologic studies (18–23). In addition to anatomic data, functional data about the fetal situation are of interest. To date, there have been few studies in which functional MR imaging techniques were applied to fetal imaging (16,17,24–27). Although Moore et al (27) were able to demonstrate fetal brain activation due to acoustic stimulation in some cases, no significant brain activation due to visual stimulation has been seen (25). Semple et al (17) tried to detect changes in fetal oxygenation with BOLD MR imaging. T2* measurements of the fetal liver were performed in nine women before and after they breathed 100% oxygen; there was a significant increase in T2* in five cases. Results of a more recent study (16), showed that functional MR imaging was capable of revealing changes in fetal tissue oxygenation during maternal hypoxia in a fetal sheep model. On the basis of those results, we investigated methodologic aspects of fetal BOLD MR imaging and attempted to quantify changes in the cerebral BOLD SI during maternal hypoxia.

Relationship between Fetal Normalized BOLD SI and Maternal Blood Oxygen Saturation
Our findings confirm those of a previous study (16) and clearly indicate that decreases in maternal blood oxygen saturation are reflected by decreases in BOLD SI in the fetal brain. Therefore, recording BOLD SI in the fetal brain appears to be a tool for monitoring changes in the supply of oxygen from the mother to the fetus. The similar time course of maternal oxygen saturation and BOLD SI values suggests that their association is close. Most likely, changes in SI are caused by changes in fetal blood oxygen saturation, and we conclude that functional MR imaging of the fetal brain is capable of depicting changes in fetal oxygen supply. However, in this study, we observed changes in normalized BOLD SI during acute hypoxia in otherwise apparently healthy fetuses. To investigate whether this method is capable of depicting a fetus with chronic ischemia, similar examinations must be performed in fetuses with growth restriction. Different reactions during hypoxia—for example, more pronounced decreases in normalized BOLD SI—might enable differentiation between healthy fetuses and those with growth restriction.

In the present study, a reduction in maternal oxygen saturation led to a reduction in the fetal BOLD SI in all animals. However, the relationship between maternal oxygen saturation changes and fetal relative BOLD SI changes varied considerably between experiments, as was expressed in different slopes. This may reflect different fetal reactions to maternal hypoxia but may also reflect the fact that fetal BOLD SIs depend directly on fetal blood oxygenation and only indirectly on maternal oxygen levels. The long-lasting nonphysiologic experimental situation we devised (involving general anesthesia with artificial ventilation and lateral positioning of the ewe) would have affected maternal-fetal oxygen exchange in the placenta and oxygen distribution in the fetus to various degrees.

Differences between Cerebrum and Cerebellum
Interestingly, there was a difference between the cerebrum and cerebellum in terms of how sensitive they were to hypoxia; linear regression analysis revealed smaller slope values for the cerebral hemispheres than for the cerebellum. It is well known that total brain blood flow may increase with moderate hypoxia ("brain sparing") (28,29). It appears that on a relative scale, cerebellar blood flow increases less than hemispheric blood flow (28); this phenomenon may make local oxygen saturation more susceptible to decreases during hypoxia. This may indicate that the cerebral brain oxygen supply is less dependent on fetal oxygen supply during hypoxia than is the cerebellum. Alternatively, the greater changes in the relative SIs may have been generated simply by the presence of a larger proportion of blood in the cerebellum than in the cerebrum (30). It is known (31,32) that the BOLD SI may decrease, for example, when the proportion of blood in a given organ increases, even when the oxygen saturation does not change. Typically, blood volume changes may be brought about by alterations in the partial pressure of CO2.

Another potential reason why the cerebellum showed a stronger decrease in SI than the cerebrum could be that the ROI in the cerebrum included the lateral ventricles, in which blood flow would not be expected to change over time. However, this influence seems to be of only minor concern, because changes in SI between different section levels were less than 5%. Finally, changes in normalized BOLD SI might have been caused by factors other than oxygenation—for example, if fetal motion occurred, the location of the fetal head may have moved with respect to the surface coil. But the influence of those changes was examined by placing a circular ROI in different positions within a reference section and seems to be of minor concern.

ROI Placement
Neither the position of the ROI within a given section, the position of the ROI within adjacent sections (above and below the reference section), nor the size of the ROI had a major influence (<12%) on the average BOLD SI in a given ROI.

The mild influence of the ROI position on the results is reassuring because small fetal head movements (which change the position of the ROI with respect to the head structures) are unavoidable: They are caused by maternal breathing, maternal bowel movements, uterine contractions, or fetal muscular activity.

Influences on BOLD SI
A major problem in the quantification of BOLD SI is the unknown influence of blood volume changes on the BOLD signal. BOLD signal changes are related to changes in blood flow, blood volume, and oxygen utilization (12,33). Results of activation studies in the human motor cortex (34) indicated that the change in deoxyhemoglobin concentration is the major source of the BOLD signal changes and that the influence of changes in blood volume on the BOLD signal is negligible. On the other hand, changes in tissue blood volume alone can result in substantial changes in T2*, independent of any change in oxygen saturation (31,32). Much effort has been made to distinguish the effects of blood volume and magnetic susceptibility changes on BOLD SI (32,34–37). High correlations between BOLD SI and cerebral blood oxygen saturation during hypoxemic hypoxia (38,39) and low correlations during hypercapnia (40) have been observed, suggesting that blood volume alterations during hypoxia have a minor influence on BOLD SI (40). However, these results cannot easily be transferred to our study because all such investigations of blood volume changes have been based on experiments with adult (human or animal) brains. We consider it unlikely that the BOLD signal changes in our experiment were caused predominantly by blood volume changes instead of changes in blood oxygen saturation because of the good correlation between fetal normalized BOLD SI and maternal oxygen saturation that we observed.

Practical application: We conclude that BOLD MR imaging is a possible tool for detecting changes in tissue oxygenation in the brains of fetal sheep following changes in maternal oxygen saturation during acute hypoxia. A reduction in the fetal BOLD SI with maternal hypoxia was observed in all experiments, but there was a large variation among the regression coefficients, with a difference between the cerebrum and cerebellum. It is to be expected that fetal BOLD SIs will be more closely associated with fetal rather than maternal oxygenation parameters, and we are currently testing this hypothesis. In the future, results of animal studies must show whether and how reliable changes in fetal oxygenation can be predicted with functional MR imaging.


   ACKNOWLEDGMENTS
 
The support of the Animal Care Facility of the University Hospital Hamburg Eppendorf (Jens Dimigen, MVD, and Bastian Tiemann, MVD) is gratefully acknowledged. We thank Hendrik Kooijman, PhD, at Philips Medical Systems, Best, the Netherlands for development and optimization of MR imaging sequences. Development of the computer software used to evaluate MR images was substantially aided by Dietmar Jörn Kreye, PhD, of the Department of Computer Science, University Kiel, Schleswig-Holstein, Germany, and is gratefully acknowledged.


   FOOTNOTES
 

Abbreviations: BOLD = blood oxygen level dependent • CI = confidence interval • ROI = region of interest • SI = signal intensity

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, all authors; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, U.W., M.T., M.W., H.J.S.; experimental studies, U.W., M.T., S.S., A.N.P., M.W., H.J.S.; statistical analysis, U.W., S.S., G.A., H.J.S.; and manuscript editing, U.W., A.N.P., G.A., H.J.S.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Kok JH, den Ouden AL, Verloove-Vanhorick SP, Brand R. Outcome of very preterm small for gestational age infants: the first 9 years of life. Br J Obstet Gynaecol 1998;105:162–168.[Medline]
  2. Dobson PC, Abell DA, Beischer NA. Mortality and morbidity of fetal growth retardation. Aust N Z J Obstet Gynaecol 1981;21:69–72.[Medline]
  3. Pryor J. The identification and long term effects of fetal growth restriction. Br J Obstet Gynaecol 1997;104:1116–1122.[Medline]
  4. Widdowson EM, McCane RA. A review: new thoughts on growth. Pediatr Res 1975;9:154–156.[Medline]
  5. Barker DJ. Growth in utero and coronary heart disease. Nutr Rev 1996;54(2 pt 2):S1–S7.
  6. Nicolaides KH, Economides DL, Soothill PW. Blood gases, pH and lactate in appropriate and small-for-gestational-age fetuses. Am J Obstet Gynecol 1989;161:996–1001.[Medline]
  7. Soothill PW, Nicolaides KH, Bilardo K, Hackett GA, Campbell S. Utero-placental blood velocity resistance index and umbilical venous pO2, pCO2, pH, lactate and erythroblast count in growth-retarded fetuses. Fetal Ther 1986;1:176–179.[Medline]
  8. Macara L, Kingdom JC, Kaufmann P, et al. Structural analysis of placental terminal villi from growth-restricted pregnancies with abnormal umbilical artery Doppler waveforms. Placenta 1996;17:37–48.[Medline]
  9. Nelson KB, Dambrosia JM, Ting TY, Grether JK. Uncertain value of electronic fetal monitoring in predicting cerebral palsy. N Engl J Med 1996;334:613–618.[Abstract/Free Full Text]
  10. Devoe L, Golde S, Kilman Y, Mortom D, Shea K, Waller J. A comparison of visual analyses of intrapartum fetal heart rate tracings according to the new National Institute of Child Health and Human Development guidelines with computer analyses by an automated fetal heart rate monitoring system. Am J Obstet Gynecol 2000;183:361–366.[CrossRef][Medline]
  11. Ogawa S, Menon RS, Tank DW, et al. Functional brain mapping by blood oxygenation level dependent contrast magnetic resonance imaging. Biophys J 1993;64:803–812.[Abstract/Free Full Text]
  12. Ogawa S, Menon RS, Kim SG, Ugurbil K. On the characteristics of functional magnetic resonance imaging of the brain. Annu Rev Biophys Biomol Struct 1998;27:447–474.[CrossRef][Medline]
  13. Ogawa S, Lee TM, Kay AR, Tank DW. Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sci U S A 1990;87:9868–9872.[Abstract/Free Full Text]
  14. Ogawa S, Tank D, Menon RS. Intrinsic signal changes accompanying sensory stimulation: functional brain mapping with magnetic resonance imaging. Proc Natl Acad Sci U S A 1992;89:5951–5955.[Abstract/Free Full Text]
  15. Belliveau JW, Kennedy DN Jr, McKinstry RC, et al. Functional mapping of the human visual cortex by magnetic resonance imaging. Science 1991;254:716–719.[Abstract/Free Full Text]
  16. Wedegärtner U, Tchirikov M, Koch M, Adam G, Schröder H. Functional magnetic resonance imaging (fMRI) for fetal oxygenation during maternal hypoxia: initial results. Rofo 2002;174:700–703.[Medline]
  17. Semple SI, Wallis F, Haggarty P, et al. The measurement of fetal liver T2* in utero before and after maternal oxygen breathing: progress towards a non-invasive measurement of fetal oxygenation and placental function. Magn Reson Imaging 2001;19:921–928.[CrossRef][Medline]
  18. Matsuoka S, Takeuchi K, Yamanaka Y, Kaji Y, Sugimura K, Maruo T. Comparison of magnetic resonance imaging and ultrasonography in the prenatal diagnosis of congenital thoracic abnormalities. Fetal Diagn Ther 2003;18:447–453.[CrossRef][Medline]
  19. Caire JT, Ramus RM, Magee KP, Fullington BK, Ewalt DH, Twickler DM. MRI of fetal genitourinary anomalies. AJR Am J Roentgenol 2003;181:1381–1385.[Abstract/Free Full Text]
  20. Hartung J, Heling KS, Rake A, Zimmer C, Chaoui R. Detection of an aneurysm of the vein of Galen following signs of cardiac overload in a 22-week old fetus. Prenat Diagn 2003;23:901–903.[CrossRef][Medline]
  21. Malinger G, Ben-Sira L, Lev D, Ben-Aroya Z, Kidron D, Lerman-Sagie T. Fetal brain imaging: a comparison between magnetic resonance imaging and dedicated neurosonography. Ultrasound Obstet Gynecol 2004;23:333–340.[CrossRef][Medline]
  22. Shinmoto H, Kuribayashi S. MRI of fetal abdominal abnormalities. Abdom Imaging 2003;28:877–886.[Medline]
  23. Ghi T, Tani G, Savelli L, Colleoni GG, Pilu G, Bovicelli L. Prenatal imaging of facial clefts by magnetic resonance imaging with emphasis on the posterior palate. Prenat Diagn 2003;23:970–975.[CrossRef][Medline]
  24. Borowska-Matwiejczuk K, Lemancewicz A, Tarasow E, et al. Assessment of fetal distress based on magnetic resonance examinations: preliminary report. Acad Radiol 2003;10:1274–1282.[CrossRef][Medline]
  25. Fulford J, Vadeyar SH, Dodampahala SH, et al. Fetal brain activity in response to a visual stimulus. Hum Brain Mapp 2003;20:239–245.[CrossRef][Medline]
  26. Hykin J, Moore R, Duncan K, et al. Fetal brain activity demonstrated by functional magnetic resonance imaging. Lancet 1999;354:645–646.[CrossRef][Medline]
  27. Moore RJ, Vadeyar S, Fulford J, et al. Antenatal determination of fetal brain activity in response to an acoustic stimulus using functional magnetic resonance imaging. Hum Brain Mapp 2001;12:94–99.[CrossRef][Medline]
  28. Gleason CA, Hamm C, Jones MD. Effect of acute hypoxemia on brain blood flow and oxygen metabolism in immature fetal sheep. Am J Physiol 1990;258(4 pt 2):H1064–H1069.
  29. Jensen A, Roman C, Rudolph A. Effects of reducing uterine blood flow on fetal blood flow distribution and oxygen delivery. J Dev Physiol 1991;15:309–323.[Medline]
  30. Foley LM, Picot P, Thompson RT, Yau MJ, Brauer M. In vivo monitoring of hepatic oxygenation changes in chronically ethanol-treated rats by functional magnetic resonance imaging. Magn Reson Med 2003;50:976–983.[CrossRef][Medline]
  31. Jezzard P, Heineman F, Taylor J, et al. Comparison of EPI gradient-echo contrast changes in cat brain caused by respiratory challenges with direct simultaneous evaluation of cerebral oxygenation via a cranial window. NMR Biomed 1994;7:35–44.[Medline]
  32. Kennan RP, Scanley BE, Gore JC. Physiologic basis of BOLD MR signal changes due to hypoxia/hyperoxia: separation of blood volume and magnetic susceptibility effects. Magn Reson Med 1997;37:953–956.[Medline]
  33. Van Zijl PC, Eleff SM, Ulatowshi JA, et al. Quantitative assessment of blood flow, blood volume and blood oxygenation effects in functional magnetic resonance imaging. Nat Med 1998;4:159–167.
  34. Toronov V, Walker S, Gupta R, et al. The roles of changes in deoxyhemoglobin concentration and regional cerebral blood volume in the fMRI BOLD signal. Neuroimage 2003;19:1521–1531.[CrossRef][Medline]
  35. An H, Lin W. Cerebral oxygen extraction fraction and cerebral venous blood volume measurements using MRI: effects of magnetic field variation. Magn Reson Med 2002;47:958–966.[CrossRef][Medline]
  36. Pears JA, Francis ST, Butterworth SE, Bowtell RW, Gowland PA. Investigation of the BOLD effect during infusion of Gd-DTPA using rapid T2* mapping. Magn Reson Med 2003;49:61–70.[CrossRef][Medline]
  37. Scheffler K, Seifritz E, Haselhorst R, Bilecen D. Titration of the BOLD effect: separation and quantitation of blood volume and oxygenation changes in the human cerebral cortex during neuronal activation and ferumoxide infusion. Magn Reson Med 1999;42:829–836.[CrossRef][Medline]
  38. Lin W, Paczynski RP, Celik A, Kuppusamy K, Hsu CY, Powers WJ. Experimental hypoxemic hypoxia: changes in R2* of brain parenchyma accurately reflect the combined effects of changes in arterial and cerebral venous oxygen saturation. Magn Reson Med 1998;39:474–481.[Medline]
  39. Prielmeier F, Nagatomo Y, Frahm J. Cerebral blood oxygenation in rat brain during hypoxic hypoxia: quantitative MRI of effective transverse relaxation rates. Magn Reson Med 1994;31:678–681.[Medline]
  40. Lin W, Celik A, Paczynski RP, Hsu C, Powers W. Quantitative magnetic resonance imaging in experimental hypercapnia: improvement in the relation between changes in brain R2* and the oxygen saturation of venous blood after correction of changes in cerebral blood volume. J Cereb Blood Flow Metab 1999;19:853–862.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 Am. J. Neuroradiol.Home page
M. Castillo
Selective Vulnerability and the Cerebellum in Neonates
AJNR Am. J. Neuroradiol., January 1, 2007; 28(1): 20 - 21.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wedegärtner, U.
Right arrow Articles by Schröder, H. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wedegärtner, U.
Right arrow Articles by Schröder, H. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE