Journal of Applied Physiology Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Appl Physiol 98: 2259-2267, 2005. First published January 7, 2005; doi:10.1152/japplphysiol.00245.2004
8750-7587/05 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow All Versions of this Article:
98/6/2259    most recent
00245.2004v1
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 Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Månsson, S.
Right arrow Articles by Golman, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Månsson, S.
Right arrow Articles by Golman, K.

3He MRI-based assessment of posture-dependent regional ventilation gradients in rats

Sven Månsson,1 Anselm J. Deninger,2 Peter Magnusson,2 Göran Pettersson,2 Lars E. Olsson,2 Georg Hansson,2 Per Wollmer,3 and Klaes Golman2

1Department of Experimental Research, Malmö University Hospital, 2Amersham Health Research and Development, Medeon, and 3Department of Clinical Physiology, Malmö University Hospital, Malmö, Sweden

Submitted 8 March 2004 ; accepted in final form 4 January 2005


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
A recently developed method for quantitative assessment of regional lung ventilation was employed for the study of posture-dependent ventilation differences in rats. The measurement employed hyperpolarized 3He MRI to detect the build-up of the signal intensity after increasing numbers of 3He breaths, which allowed for computation of a regional ventilation parameter. A group of six anesthetized rats was studied in both supine and prone postures. Three-dimensional maps of the ventilation parameter were obtained with high spatial resolution (voxel volume ~2 mm3). Vertical (dorsal-ventral) gradients of the ventilation index, defined as the regional ventilation normalized by the average ventilation within the whole lung, were investigated. Variations in the regional distribution of the ventilation parameter, as well as of the ventilation index, could be detected, depending on the posture of the rats. In supine posture, ventilation was elevated in the dependent parts of the lungs, with a linear gradient of the ventilation index of –0.11 ± 0.03 cm–1. In prone posture, the distribution of ventilation was more uniform, with a significantly (P < 0.001) smaller gradient of the ventilation index of –0.01 ± 0.02 cm–1. It is concluded that the 3He MRI-based method can detect and quantify regional ventilation gradients in animals as small as the rat and that these gradients depend on prone or supine posture of the animal.

hyperpolarized gas magnetic resonance imaging; helium-3; lung function; posture dependence

IT IS WELL KNOWN THAT THE distribution of regional pulmonary ventilation depends on the posture of an air-breathing animal or human (27). This has classically been attributed to effects of gravity on pleural pressure and alveolar expansion (24). Regional ventilation values change dramatically between prone and supine body postures, with predominantly dorsal ventilation in supine posture and a more uniform distribution of ventilation in prone posture (15, 16, 20, 31, 40, 41). Oxygenation and gas exchange improve in prone posture, but the exact mechanism remains unclear (27). This is not fully explained by effects of gravity, and other important factors have been suggested, e.g., dorsoventral differences in lung structure balancing out the gravitational forces in the prone posture (37). Various studies on the influence of posture on the regional distribution of ventilation have been conducted on humans and large animals like dogs, sheep, and pigs (28), but no data are available for small species like the rat.

In the past, quantitative measurements of regional lung ventilation have been obtained with invasive techniques or with radioisotope imaging (2, 3, 5), but these methods have been limited in their spatial and temporal resolution. Improved spatial resolution has been realized with xenon-enhanced computed tomography, where the ventilation is determined from the wash-in and wash-out rates of stable xenon (13, 14, 18, 20). Using a MRI-based method, regional ventilation has been evaluated qualitatively from the variations in lung parenchyma signal when a volunteer was breathing air and pure oxygen, respectively (10).

The possibility to use hyperpolarized 3He gas for magnetic resonance imaging (9, 19, 23) has opened new possibilities for the examination of lung ventilation and flow patterns (6). Mata et al. (21, 22) qualitatively investigated the dependence of ventilation on posture by imaging healthy humans in prone and supine posture. In the most inferior (dependent) parts of the lung, they found local ventilation defects when a subject was imaged in supine posture. These defects, which resolved when the subject was imaged in prone posture, were attributed to posture-dependent atalectasis (closure of small airways). Similar findings were observed in a study of healthy and asthmatic subjects by Altes et al. (1).

Recently, Deninger et al. (8) demonstrated a quantitative method to assess regional ventilation in guinea pigs with high spatial resolution. The method employed a mathematical model to describe the buildup of the magnetic resonance (MR) signal after administering a certain number of 3He breaths. Thus a regional ventilation parameter, defined as the relative exchange of gas per breath, could be determined with an accuracy of 2–5% and a spatial resolution better than 1 mm. However, the previous study (8) was restricted to two-dimensional imaging of a single thin slice of the lungs.

In the present study, the 3He-MRI-based method was used to calculate high-resolution three-dimensional (3D) maps of the regional ventilation parameter in rats in both prone and supine posture. The objective was to investigate whether posture-dependent gradients of ventilation could be detected in the rat lung.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
Ventilation model.   Our ventilation model, published by Deninger et al. (8), treats ventilation as a discontinuous, cyclic process. After an inspiration of pure 3He, a fraction r of a given volume element is replaced by fresh 3He, whereas a fraction q contains gas remaining from previous inspirations:

(1)
where Vold and Vfr denote the volumes of old and fresh gas within the volume element, respectively. We assume that all breaths are identical, which is well fulfilled during mechanical respiration, as employed in this study. The unitless parameter r then gives the relative exchange of fresh gas per breath, which serves as a measure of lung ventilation.

To measure the ventilation parameter r, a breathing cycle consisting of N air breaths followed by n breaths of pure 3He is repeated several times with increasing values of n. The duration between consecutive breaths is denoted {tau}. After the jth inspiration of 3He in the nth cycle, the available magnetization M(j,n) of 3He per unit volume consists of 1) the magnetization Mnew of fresh 3He from the jth inspiration, and 2) magnetization remaining from previous breaths of the same cycle (for j ≥ 2). Fraction 1 is subject to spin-lattice relaxation in the external 3He reservoir, T1,ext, caused by surface relaxation (11) and field gradient relaxation (36). Fraction 2 is mainly affected by oxygen-induced relaxation in the lung, T, during the interval {tau} (35). Other relaxation mechanisms play an insignificant role and are neglected here (7). Thus M(j,n) can be written recursively:

(2)
with a start value M(0,n) {equiv} 0 and an exponentially decaying "reservoir magnetization" Mnew(n):

(3)
Here, B counts the total number of inspirations between the first 3He breath of cycles 1 and n, and C is a proportionality factor.

Whenever the inspiratory gas is switched to 3He, oxygen is washed out from the lungs. After the jth inspiration of 3He, the oxygen partial pressure (PO2) has decreased to PO2(j) = P0qj, where P0 denotes the PO2 before the first 3He breath of a cycle. The relaxation time of the oxygen-induced depolarization is related to PO2 via

(4)
with a proportionality factor {xi} {approx} 2.6 bar/s at body temperature (35).

Insertion of Eqs. 3 and 4 into Eq. 2 yields

(5)
where E {equiv} exp(–{tau}/T1,ext). Equation 5 can be rearranged to explicitly express the signal intensity S(n) after the nth 3He inspiration, which is proportional to the magnetization M(n,n):

(6)
A nonlinear least squares fit of Eq. 6 to the measured signal intensities is used to compute the ratio q and thus the ventilation parameter r = 1 – q.

Polarization and administration of 3He.   3He was polarized using spin-exchange optical pumping (IGI 9600 Helium Polarizer, Amersham Health, Durham, NC). A quantity of 1.1 bar/l was polarized to 35% within 15 h. The hyperpolarized 3He was collected in a 300-ml Tedlar bag (Jensen Inert, Coral Springs, FL) at 1-bar pressure. The bag was connected to an in-house-built respirator, described elsewhere (8).

Animal preparation.   Six Sprague-Dawley rats (male, 350–400 g; breeder: M&B, Ry, Denmark) were anesthetized subcutaneously with a mixture of fentanyl (0.1 mg/kg) and fluanisone (5 mg/kg) (Hypnorm, Janssen Animal Health, Saunderton, UK) and midazolam (2.5 mg/kg; Dormicum, Hoffman-La Roche, Basel, Switzerland). The left jugular vein was catheterized for intravenous administration of anesthesia and a neuromuscular blocking agent (pancuronium, Pavulon, Organon Teknika, Boxtel, Netherlands). After tracheal intubation, the animals were placed in the MR scanner and ventilated by the respirator mentioned above. During imaging, the body temperature of the animals was monitored continuously and was kept within the range 37–38°C by circulating a heated fluid through the animal bed. Two animals were imaged in supine posture first, followed by imaging in prone posture. The other four animals were imaged in prone posture first. The breathing rate was set to 30 breaths/min. The peak inspiratory pressure was 20 cmH2O, resulting in tidal volumes (VT) in the range of 9–12 ml. Inspiration and expiration times were 0.5 and 1.0 s, respectively. Positive end-expiratory pressure was not used. The reproducibility of gas volumes administered by the ventilator was tested in separate experiments and was better than 2% (8). The experiments were approved by the local ethics committee (Malmö/Lunds djurförsöksetiska nämnd; appl. no. M4–01).

3He imaging.   All experiments were performed on a 2.35-T scanner (BioSpec 24/30, Bruker BioSpin, Ettlingen, Germany) using a bird-cage radio-frequency coil (Bruker BioSpin) with 72 mm diameter and 110 mm length, tunable to the Larmor frequencies of 1H (100.1 MHz) and 3He (76.6 MHz). Coronal, sagittal, and axial proton scout images were acquired for proper localization of the 3He images.

The breathing pattern used for the experiment is shown in Fig. 1. An experiment consisted of four ventilator cycles, each cycle comprising 15 air breaths followed by an increasing number {1, 2, 3, 4} of helium breaths. After the last helium inspiration of each cycle, breathing was suspended for 3.5 s. To minimize effects of gas motion and turbulence, image acquisition was started 0.5 s after the onset of the breath hold. A 3D fast low-angle shot sequence was employed (repetition time, echo time, and flip angle = 2.1 ms, 0.9 ms, and 2.5°, respectively; field of view = 6 x 5 x 4.5 cm3; acquisition matrix =48 x 48 x 24). The images were interpolated by zero-filling to a final size of 96 x 96 x 48 voxels. The 3D image set was positioned to cover all parts of the lungs and the trachea. Before and after the air/3He cycles sketched in Fig. 1, 3He reference images were obtained to monitor the T1 relaxation of the hyperpolarized gas in the plastic bag. The experiment was repeated twice on each animal, in prone and supine postures, respectively.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 1. Illustration of breathing pattern and image acquisition. Each ventilator cycle consisted of 15 air breaths followed by an increasing number of 1, 2, 3, or 4 3He breaths. Imaging was performed during inspiratory apnea after the last 3He breath of a cycle.

 
Data analysis.   Data analysis was carried out using software implemented in MATLAB (MathWorks, Natick, MA). First, the noise level of an image was calculated from the mean background signal, Sback, using the relation noise = ·Sback (12). The signal-to-noise ratio (SNR) of an image was assessed by dividing the mean signal intensity within a large region of the lung by the noise level.

The external relaxation time T1,ext in the 3He reservoir was computed from the decay of the average image intensity of the reference images. The regional ventilation parameter r was then calculated in each voxel (interpolated voxel volume: 0.3 mm3) by fitting the function given by Eq. 6 to the signal intensities S(n) in images acquired after n = 1, 2, 3, and 4 3He breaths. Background voxels were excluded by selecting a threshold SNR of 20 in the last (n = 4) image. Values of {tau} = 2 s and N = 15 were used in the experiments. In the calculation, a value of P0 = 135 mbar served as approximation of the intrapulmonary oxygen partial pressure (26, 29).

Calculation of posture-dependent ventilation gradients.   To compare interindividual results independent of the absolute value of the ventilation parameter, maps of the ventilation index (VI) were generated for each animal. The VI was calculated by dividing the ventilation parameter of each voxel by the average value of all voxels, excluding the trachea and the bronchi. The latter regions were omitted from the VI calculation by excluding all voxels with ventilation parameter >0.8. The total end-inspiratory lung volume was estimated by multiplying the voxel volume with the number of voxels included in the VI calculation.

Next, the dependence of the VI on the vertical coordinate y was investigated. The coordinate axis y was chosen antiparallel to the direction of gravity, perpendicular to the coronal plane. Negative y-coordinates thus denote lower (more inferior) parts of the lung. The analysis of VI(y) was performed in columns containing ≤30 voxels (apical region of lungs) or ≤40 voxels (midbase region), corresponding to vertical distances of 2.8–3.8 cm. To increase the accuracy of the computation, only columns with ≥5 data points in the ventral-dorsal direction were analyzed. For computation of the VI gradient ({partial}VI/{partial}y), the coronal planes of the VI map were smoothed by averaging regions of 5 x 5 voxels. The effective resolution of the VI gradient maps (see Fig. 5) is thus ~3 mm.



View larger version (45K):
[in this window]
[in a new window]
  		Fig. 5. A: color-coded representations of the vertical gradient {partial}VI/{partial}y (cm–1) for animals 1–6. B: histograms of the gradients. The numbers above each histogram indicate the median gradient. Blue areas and negative values correspond to regions where the ventilation is increased in the dependent parts of the lung.

 
As pulmonary ventilation decreases with increasing distance from the hilus (8, 34), a nonlinear variation of the VI(y) is expected. For this reason, a quadratic rather than a linear function was chosen to model the dependence of VI(y):

(7)
The function VI(y) was fitted to the ventilation parameter data using a least squares fit, with a, b, and c being the free parameters of the fit. ytop, ybottom, and ymid are the coordinates of the uppermost, lowermost, and middle voxel of each column, respectively. The parameter ymid was introduced to render the analysis independent of translations along the y-axis. The linear coefficient b was used as a measure of the linear vertical ventilation gradient. In addition, the reduced {chi}2 value (4) was computed to evaluate the quality of the fit.

For each animal, median values of the gradient b were calculated globally, as well as in apical (apex to midlung) and midbase (midlung to top of diaphragm) regions. To evaluate statistically significant differences between prone and supine postures in each animal, P values were calculated according to Student's paired t-test.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
Images were obtained with average SNR values ranging from 35 (one 3He breath) to 75 (four 3He breaths). Total end-inspiratory lung volumes were 21.0 ± 1.1 and 20.9 ± 0.7 ml in prone and supine postures, respectively. Representative examples of the signal build-up in voxels with high and low values of the regional ventilation parameter r are shown in Fig. 2. Overall, values of r varied between ~0.2 in the peripheral regions and nearly 1 in the trachea and the major bronchi, indicating a near-complete per-breath renewal of gas in the latter regions (Fig. 3). Averaged over all animals, the ventilation parameter was 0.49 ± 0.03 (mean ± SD) in prone posture and 0.52 ± 0.05 in supine posture (see Table 1).



View larger version (11K):
[in this window]
[in a new window]
  		Fig. 2. Typical signal build-up in the 3He images, shown for voxels with low (r = 0.25) and high ventilation parameter (r = 0.65), respectively. The lines represent the fit of Eq. 6 to the signal intensities after 1 ... 4 3He breaths. a.u., Arbitrary units.

 


View larger version (55K):
[in this window]
[in a new window]
  		Fig. 3. Coronal, sagittal, and axial slices of three-dimensional acquisitions after n helium breaths in prone (A) and supine (B) postures (animal 4). The rightmost column shows the map of the regional ventilation parameter, calculated from the 4 columns on the left. The arrows indicate the dependent direction (direction of gravity; negative y-direction). In supine posture, higher ventilation in the dependent parts can be seen in the sagittal and axial ventilation maps.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Ventilation parameter r for animals 1–6

 
In supine posture, a clear tendency toward increased values of r in the dependent (most inferior) parts of the lung (direction of gravity indicated by arrows in Fig. 3) was discernible in the sagittal and axial views of the 3D ventilation maps. Considerably lower values (decrease by up to ~40%) were found in the nondependent (most superior) parts. By contrast, the corresponding maps computed from images taken in prone posture appeared more homogeneous and depicted a more uniform distribution of the ventilation parameter. Within the trachea and the major airways, the ventilation parameter did not differ systematically in prone and supine posture.

Figure 4 shows plots of the VI(y), measured in single-voxel columns in the apical and midbase regions of the lungs of a representative animal (animal 4), in both prone and supine postures. The solid line is a fit of Eq. 7 to the data points, and the dashed line represents the linear component (the coefficient b of Eq. 7). The nonlinear course of VI(y) is visible in particular in the apical lung region. Consistently, this is seen in both prone and supine postures (Fig. 4, A and C). In the examples of Fig. 4, A and C, the reduced {chi}2 values are 15–20 times lower when using the quadratic fit instead of a linear fit. The ventilation enhancement in inferior lung regions in supine posture, qualitatively apparent from Fig. 3, is manifest in considerably different values of the linear gradient coefficient b (prone posture: b = 0.00 cm–1, supine posture: b = –0.10 cm–1). This dependence on posture is still more prominent in midbase lung regions (Fig. 4, B and D). Again, the linear coefficient is zero (b = 0.00 cm–1) in prone posture and negative in supine posture (b = –0.19 cm–1). The data points are well described by the quadratic fit function of Eq. 7. Within the entire lung, this fit function decreased the {chi}2 values by a factor of ~3, compared with a linear fit.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 4. Representative plots of the ventilation index (VI) as function of vertical coordinate in prone posture (A and B) and in supine posture (C and D). Solid arrowhead, dependent; open arrowhead, nondependent. The solid lines show the nonlinear fit to the VI data ({circ}). The dashed lines show the general (linear) trend of the VI (given by the coefficient b in Eq. 7). Data are taken from animal 4.

 
Color-coded maps and histograms of the linear gradient b (cm–1) of the VI are shown in Fig. 5, A and B, respectively. Blue colors indicate negative gradient values and thus a decrease of the VI toward nondependent lung regions. For animals imaged in prone posture, the VI gradients are close to zero everywhere in the lungs, except for some areas near the edges of the gradient maps. The absence of vertical VI gradients is also seen in the corresponding histograms, where the distributions are centered at about zero (mean ± SD of all animals = –0.02 ± 0.01 cm–1). In supine posture, more pronounced vertical VI gradients are found. Gradients are noticed particularly in the midbase region of the lungs (–0.18 ± 0.03 cm–1). In the apical regions, the magnitudes of the VI gradients are smaller (–0.07 ± 0.03 cm–1) but still considerably larger than in prone posture. Consequently, in the histogram representation (Fig. 5B), the distributions of VI gradients center at about –0.14 cm–1 in supine posture. These findings are summarized in Table 2, which lists the calculated VI gradients in prone and supine postures, measured both globally and regionally within areas near the apex and the lung base, respectively. Significant differences of the median VI gradients are found in all regions of the lung (P < 0.01), depending on whether the animal was placed in prone or supine posture.


View this table:
[in this window]
[in a new window]
  		Table 2. Gradient of ventilation index for animals 1–6, measured globally and within regions near the apex and the lung base

 
To test the significance of the measured VI gradients, the quadratic term of Eq. 7 was subtracted from the measured VI data points, and the correlation between the "linearized" VI values and the vertical distance y was computed. Figure 6 shows maps of the squared correlation coefficient R2 between VI and vertical location for prone and supine image sets. In prone posture, the correlation was generally very low (R2 < 0.4), whereas a high correlation was found in supine posture (R2 {approx} 0.7). Viewed regionally, the steep gradients measured in basal regions yield correlation values of R2 > 0.8, whereas the somewhat lower values in apical regions (R2 {approx} 0.6) correspond to shallower gradients (cf. Fig. 5A). Results of the correlation coefficient are summarized in Table 3.



View larger version (38K):
[in this window]
[in a new window]
  		Fig. 6. Maps of linear correlation (R2 maps) between measured VI (after subtraction of the quadratic term) and vertical coordinate, in animals 1–6.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Correlation coefficient (R2) between measured ventilation index values (after subtraction of the quadratic term of Eq. 7) and vertical coordinate, for animals 1–6, measured globally and within regions near the apex and the lung base

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ACKNOWLEDGMENTS
 REFERENCES
 
In this study, the signal build-up after increasing numbers of 3He inspirations is used to calculate a parameter r, which denotes the relative exchange of gas per breath within a local volume element of the lungs. This quantity is related, but not equal, to the traditional definition of alveolar ventilation, which is the total volume of fresh gas entering the respiratory zone each minute (38). However, there exists a simple relationship between the parameter r and the specific ventilation, s, defined as the ventilation normalized to the regional lung air content. Averaged over the whole lung, the alveolar s and ventilation parameter r are given by

(8)
where f is the breathing rate (min–1), VT is in ml, VDS is dead space volume (ml), and FRC is functional residual capacity (ml). From Eq. 8, it follows that

(9)
On a regional scale, the same relation is obtained by using the definition given in Eq. 1. For small values of r (r < 0.15–0.20), the relationship between r and s is thus approximately linear.

In the experiments, FRC was not measured, but, given the high inspiratory pressure of 20 cmH2O and large VT (~25 ml/kg), the computed end-inspiratory lung volumes of {approx}21 ml can be regarded as a measure of the total lung capacity. The FRC can thus be estimated by subtracting the VT from the end-inspiratory lung volume, yielding an average value of FRC of ~10 ml. Neglecting the dead space, the average s can then be estimated, according to Eq. 8, to ~33 min–1. This is in good agreement with the value of 30 min–1, which is obtained by inserting an average value of r = 0.5 in Eq. 9. In peripheral lung regions, where r ~ 0.2, Eq. 9 yields s = 7.5 min–1. The latter value can be compared with s {approx} 4.6 min–1 reported for a rabbit study with VT = 12.7 ml/kg (32).

Since the ventilation parameter is measured after inspiration of pure 3He, a gas with higher diffusion coefficient than air, the absolute values of the ventilation parameter can be expected to be somewhat higher compared with the gas exchange of air, at least in terminal airways where the gas transport is driven by diffusion rather than flow. On the other hand, in these regions, the diffusion of helium is strongly restricted by the alveolar walls, which reduces the difference in diffusion velocity between 3He and air.

Our main finding is that, in supine posture, the regional ventilation parameter increases toward dependent parts of the lungs, whereas a more homogenous distribution is found in prone posture. In other words, in supine posture, dorsal regions are better ventilated. This finding is in general agreement with results obtained in large animals (20, 37) and humans (30), yet, to our knowledge, the present study is the first to describe the phenomenon in an animal as small as the rat.

The method of measurement has been adopted from a previous proof-of-principle study (8), which was able to show that hyperpolarized 3He MRI offers the possibility to assess regional lung ventilation in small animals quantitatively and with high accuracy. However, the experiments described in Ref. 8 were restricted to imaging of a single thin coronal slice of the lungs. In the present study, the technique was employed to acquire high-resolution 3D images in both prone and supine postures. An advantage of the 3D acquisition and analysis is that data from the entire lungs are utilized.

Compared with the earlier study (8), a modified breathing scheme was used. To compensate for the smaller voxel volume in the present study, VT were increased by a factor of 3–4 in order to optimize the image SNR and thus improve the measurement accuracy. The average ventilation parameter of r {approx} 0.5 implies that the VT approximately equals FRC (cf. Eq. 8). For a human, this corresponds to a VT of ≥2 liters. Hence, the ventilation distribution observed in this study may differ from more normal breathing patterns. In particular, due to the increased gas intake, the assumption of an initial intrapulmonary oxygen pressure of P0 = 135 mbar may be an underestimation. However, the error of the ventilation parameter r due to an underestimation of P0 is at most ~2% and becomes even smaller with increased ventilation, as oxygen is washed out more rapidly from the lungs (8). Furthermore, efforts were made to limit the consumption of 3He: the maps of the ventilation parameter were calculated from {1, 2,..., 4} 3He breaths only, instead of using {1, 2,..., 7} breaths. Also, the number of air breaths was reduced from 20 to 15, thereby shortening the total duration of a measurement to ~2.5 min. From SNR considerations and error analysis as outlined in Ref. 8, we estimate the relative uncertainty ({Delta}r/r) of the calculated ventilation parameter r to be <15% in voxels with r > 0.3, <9% for r > 0.4, and <6% for r > 0.5. Clearly, this uncertainty alone does not account for the observed intraindividual variability of r (Table 1): if the scatter of r was caused by noise only, a standard deviation of 0.03 rather than ~0.11 would be expected for r. We thus presume that the measured variability is of physiological origin and only to a small extent influenced by noise. In poorly ventilated regions, e.g., under pathological conditions or when low VT are used, the precision of the method will obviously degrade. In such cases, SNR may be regained by sacrificing spatial resolution or by optimizing the scheme of helium inspirations [e.g., {1, 3, 5} inspirations are superior to {1, 2, 3, 4} (8)]. The polarization level of the 3He is certainly crucial: in this study, the polarization was ~35%. We are confident that our polarizer may be optimized to yield ~50%, which would increase the SNR by >40%.

As the ratio of VT and total lung volume varied in different animals, the absolute values of the per-breath gas replacement differed substantially (relative scatter of mean value, 20%). Therefore, the VI was computed for each animal to enable an interindividual comparison of the results. VI is a measure of regional lung ventilation relative to global lung ventilation, excluding regions such as the trachea and the bronchi, which do not take part in the alveolocapillary gas exchange.

To quantify posture-dependent ventilation changes, the vertical gradient {partial}VI/{partial}y of the VI was calculated. For this step, each point of a VI map was averaged over a 5 x 5 pixel region in the coronal plane to reduce the effect of fluctuating r values. The aforementioned uncertainty in the ventilation parameter propagates to an uncertainty in the VI gradient, which decreases with an increasing number of voxels used for the calculation of {partial}VI/{partial}y. Based on Monte Carlo simulations, we estimate that the uncertainty in the VI gradient in this study is at most ±0.02 cm–1 (r = 0.2, VI gradient calculated across 5 voxels), and typically ±0.01 cm–1 (r = 0.3, VI gradient calculated across 15 voxels).

In the literature, vertical pressure gradients in the lung have been expressed as linear gradients (25, 39). Although somewhat of an oversimplification, a linear slope is an easily understood method to describe a general trend in the data. Plots of the VI as a function of the vertical coordinate y (Fig. 4), however, reveal an obvious nonlinear dependence, with elevated VI in the central part of the lung. Robertson et al. (34) reported reduced ventilation in the periphery of the lungs in pigs, but the central-to-peripheral gradient in that study (0.060 cm–1) is clearly smaller than in our measurement, as shown in Fig. 4, A and C. The strong nonlinearity in Fig. 4 is most likely explained by the presence of 3He in conducting airways: the ventilation parameter r in the conducting airways (the anatomical dead space from trachea to the terminal bronchioles) is {approx}1 (8). In central voxels, the signal contribution from the anatomical dead space is presumably significant, whereas it is negligible in the most peripheral voxels. Thus the ventilation parameter is supposed to be a relevant measure of alveolar ventilation only in the periphery of the lung. To account for the elevated ventilation parameter within the center of the lungs, a quadratic rather than linear fit function (Eq. 7) was chosen to describe VI(y). The {chi}2 value was reduced by a factor of 3, on average, by using the quadratic fit, demonstrating the superiority of this fit function to model the course of VI(y).

The quadratic fit thus allows for distinguishing between the parabolic, gas-replacement-induced variation of the VI and purely linear variations, which may have different physiological interpretations (27). One way of probing the significance of the linear VI gradients is to compute the correlation coefficient between VI values and vertical distance y after subtraction of the quadratic term of Eq. 7. The low correlation observed in prone posture indicates that vertical VI gradients are zero or at most very small. (In a few cases, high R2 values are observed in prone posture near the edges of the lungs, where the 3D data set comprises only a small number of voxels along the vertical axis, increasing the likelihood of an accidental alignment of data points.) In supine posture, on the other hand, high correlations are found within large regions of the lungs, in particular in the midbase regions (see Fig. 6).

The vertical VI gradients determined in this study (Table 2) are within a similar range as results described in the literature for larger animals. In prone pigs, Robertson et al. (34) measured a nonsignificant ventilation gradient of –0.03 cm–1. Marcucci et al. (20) investigated gradients of the VI in dogs, using xenon-enhanced computed tomography. In their study, global vertical VI gradients of –0.07 and –0.008 cm–1 were measured in supine and prone posture, respectively. Very similar values were found by Hubmayer et al. (17) in supine (–0.07 cm–1) and prone (–0.002 cm–1) dogs, measured by means of a fluoroscopic technique. The corresponding values of the present small-animal study are –0.11 and –0.01 cm–1. All of these studies indicate that ventilation gradients are present in supine posture but virtually absent in prone posture. Tentatively, higher gradient values in a smaller animal could be explained by the shorter distance, across which the regional variation of ventilation takes place.

Our measurement technique, which has been optimized with respect to small-animal imaging, would require more severe modifications if applied to larger animals or even humans. Even though the number of helium inspirations has been reduced compared with the previous guinea pig study (8), the number of helium inspirations is still fairly large. The impendence of hypoxia in human subjects (33) could be alleviated by adding oxygen to the inspiratory gas admixture. In this case, however, the depolarizing effect of molecular oxygen has to be considered in the mathematical model. A more general limitation is the need for absolutely reproducible inspiration volumes, which calls for the use of a mechanical ventilator.

In summary, hyperpolarized 3He was used to quantify a regional ventilation parameter, defined as per-breath gas replacement, in rat lungs, employing 3D MRI with high spatial resolution. Regional gradients of the ventilation parameter along the vertical (dorsal-ventral) direction were measured when the animals were placed in supine posture, indicating increased ventilation in dorsal lung regions. The gradients were significantly smaller, and close to zero, in prone posture. While the presence of ventilation gradients in rat lungs may be of limited physiological significance, its magnitude corresponds to findings reported in larger animals and may thus hint toward the existence of common mechanisms determining the distribution of ventilation in mammals of different sizes. We conclude that ventilation gradients can be detected in animals as small as the rat and that the described method could provide a novel alternative to other techniques, such as xenon-enhanced computed tomography, for ventilation investigations with high spatial resolution.


   ACKNOWLEDGMENTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
REFERENCES
 
The authors thank Kerstin Thyberg and Birgit Persson for excellent technical assistance.


   FOOTNOTES
 

Address for reprint requests and other correspondence: S. Månsson, Dept. of Experimental Research, Malmö Univ. Hospital, SE-205 02 Malmö, Sweden (E-mail: Sven.Mansson{at}rontgen.mas.lu.se)

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 

  1. Altes TA, Powers PL, Knight-Scott J, Rakes G, Platts-Mills TA, de Lange EE, Alford BA, Mugler JP 3rd, and Brookeman JR. Hyperpolarized 3He MR lung ventilation imaging in asthmatics: preliminary findings. J Magn Reson Imaging 13: 378–384, 2001.
  2. Ball WC, Stewart PB, Newsham LGS, and Bates DV. Regional pulmonary function studied with Xenon-133. J Clin Invest 41: 519–531, 1962.
  3. Berdine GG, Lehr JL, McKinley DS, and Drazen JM. Nonuniformity of canine lung washout by high-frequency ventilation. J Appl Physiol 61: 1388–1394, 1986.
  4. Bevington PR and Robinson DK. Data Reduction and Error Analysis for the Physical Sciences. New York: McGraw-Hill, 2002.
  5. Bunow B, Line BR, Horton MR, and Weiss GH. Regional ventilatory clearance by xenon scintigraphy: a critical evaluation of two estimation procedures. J Nucl Med 20: 703–710, 1979.
  6. Chen BT, Brau AC, and Johnson GA. Measurement of regional lung function in rats using hyperpolarized 3helium dynamic. MRI Magn Reson Med 49: 78–88, 2003.
  7. Deninger AJ, Eberle B, Ebert M, Grossmann T, Heil W, Kauczor HU, Lauer L, Markstaller K, Otten E, Schmiedeskamp J, Schreiber W, Surkau R, Thelen M, and Weiler N. Quantification of regional intrapulmonary oxygen partial pressure evolution during apnea by 3He MRI. J Magn Reson 141: 207–216, 1999.
  8. Deninger AJ, Månsson S, Petersson JS, Pettersson G, Magnusson P, Svensson J, Fridlund B, Hansson G, Erjefeldt I, Wollmer P, and Golman K. Quantitative measurement of regional lung ventilation using 3He MRI. Magn Reson Med 48: 223–232, 2002.
  9. Ebert M, Grossmann T, Heil W, Otten WE, Surkau R, Leduc M, Bachert P, Knopp MV, Schad LR, and Thelen M. Nuclear magnetic resonance imaging with hyperpolarised 3He. Lancet 347: 1297–1299, 1996.
  10. Edelman RR, Hatabu H, Tadamura E, Li W, and Prasad PV. Noninvasive assessment of regional ventilation in the human lung using oxygen-enhanced magnetic resonance imaging. Nat Med 2: 1236–1239, 1996.
  11. Fitzsimmons WA, Tankersley LL, and Walters GK. Nature of surface-induced nuclear-spin relaxation of gaseous He3. Physiol Rev 179: 156–165, 1969.
  12. Gudbjartsson H and Patz S. The Rician distribution of noisy MRI data. Magn Reson Med 34: 910–914, 1995.
  13. Gur D, Drayer BP, Borovetz HS, Griffith BP, Hardesty RL, and Wolfson SK. Dynamic computed tomography of the lung: regional ventilation measurements. J Comput Assist Tomogr 3: 749–753, 1979.
  14. Gur D, Shabason L, Borovetz HS, Herbert DL, Reece GJ, Kennedy WH, and Serago C. Regional pulmonary ventilation measurements by xenon enhanced dynamic computed tomography: an update. J Comput Assist Tomogr 5: 678–683, 1981.
  15. Hoffman EA. Effect of body orientation on regional lung expansion: a computed tomographic approach. J Appl Physiol 59: 468–480, 1985.
  16. Hoffman EA and Ritman EL. Effect of body orientation on regional lung expansion in dog and sloth. J Appl Physiol 59: 481–491, 1985.
  17. Hubmayer RD, Walters BJ, Chevalier PA, Rodarte JR, and Olson LE. Topographical distribution of regional lung volume in anesthetized dogs. J Appl Physiol 54: 1048–1056, 1983.
  18. Kreck TC, Krueger MA, Altemeier WA, Sinclair SE, Robertson HT, Shade ED, Hildebrandt J, Lamm WJ, Frazer DA, Polissar NL, and Hlastala MP. Determination of regional ventilation and perfusion in the lung using xenon and computed tomography. J Appl Physiol 91: 1741–1749, 2001.
  19. MacFall JR, Charles HC, Black RD, Middleton H, Swartz JC, Saam B, Driehuys B, Erickson C, Happer W, Cates GD, Johnson GA, and Ravin CE. Human lung air spaces: potential for MR imaging with hyperpolarized 3He. Radiology 200: 553–558, 1996.
  20. Marcucci C, Nyhan D, and Simon BA. Distribution of pulmonary ventilation using Xe-enhanced computed tomography in prone and supine dogs. J Appl Physiol 90: 421–430, 2001.
  21. Mata JF, Altes TA, Christopher JM, Mugler JP 3rd, Brookeman JR, and de Lange EE. Positional Dependence of Small Inferior Ventilation Defects Seen on Hyperpolarized Helium-3 MR of the Lung. Glasgow: ISMRM, 2001, p. 949.
  22. Mata JF, Altes TA, Christopher JM, Mugler JP 3rd, Brookeman JR, and de Lange EE. Transient, Position Dependent Ventilation Defects of the Lung in Healthy Volunteers: Demonstration With Hyperpolarized Helium-3 MR Imaging. Honolulu, HI: ISMRM, 2002, p. 2027.
  23. Middleton H, Black RD, Saam B, Cates GD, Cofer GP, Guenther R, Happer W, Hedlund LW, Johnson GA, and Juvan K. MR imaging with hyperpolarized 3He gas. Magn Reson Med 33: 271–275, 1995.
  24. Milic-Emili J. Ventilation. In: Regional Differences in the Lung, edited by West JB. New York: Academic, 1977, p. 167–199.
  25. Milic-Emili J, Henderson JA, Dolovich MB, Trop D, and Kaneko K. Regional distribution of inspired gas in the lung. J Appl Physiol 21: 749–759, 1966.
  26. Möller HE, Hedlund LW, Chen XJ, Carey MR, Chawla MS, Wheeler CT, and Johnson GA. Measurements of hyperpolarized gas properties in the lung. III. 3He T1. Magn Reson Med 45: 421–430, 2001.
  27. Mure M, Domino KB, Lindahl SG, Hlastala MP, Altemeier WA, and Glenny RW. Regional ventilation-perfusion distribution is more uniform in the prone position. J Appl Physiol 88: 1076–1083, 2000.
  28. Mure M and Lindahl SG. Prone position improves gas exchange-but how? Acta Anaesthesiol Scand 45: 150–159, 2001.
  29. Olsson LE, Magnusson P, Deninger AJ, Pettersson G, kesson P, Petersson JS, and Golman K. Intrapulmonary PO2 Measured by Low Field MR Imaging of Hyperpolarized 3He. Honolulu, HI: ISMRM, 2002, p. 2021.
  30. Pelosi P, Caironi P, Taccone P, and Brazzi L. Pathophysiology of prone positioning in the healthy lung and in ALI/ARDS. Minerva Anestesiol 67: 238–247, 2001.
  31. Pelosi P, D'Andrea L, Vitale G, Pesenti A, and Gattinoni L. Vertical gradient of regional lung inflation in adult respiratory distress syndrome. Am J Respir Crit Care Med 149: 8–13, 1994.
  32. Porra L, Monfraix S, Berruyer G, Le Duc G, Nemoz C, Thomlinson W, Suortti P, Sovijarvi AR, and Bayat S. Effect of tidal volume on distribution of ventilation assessed by synchrotron radiation CT in rabbit. J Appl Physiol 96: 1899–1908, 2004.
  33. Ramirez MP, Sigaloff KC, Kubatina LV, Donahue MA, Venkatesh AK, and Albert MS. Physiological response of rats to delivery of helium and xenon: implications for hyperpolarized noble gas imaging. NMR Biomed 13: 253–264, 2000.
  34. Robertson HT, Glenny RW, Stanford D, McInnes LM, Luchtel DL, and Covert D. High-resolution maps of regional ventilation utilizing inhaled fluorescent microspheres. J Appl Physiol 82: 943–953, 1997.
  35. Saam B, Happer W, and Middleton H. Nuclear relaxation of 3He in the presence of O2. Physiol Rev A 52: 862–865, 1995.
  36. Schearer LD and Walters GK. Nuclear spin-lattice relaxation in the presence of magnetic-field gradients. Physiol Rev 139: A1398–A1402, 1965.
  37. Treppo S, Mijailovich SM, and Venegas JG. Contributions of pulmonary perfusion and ventilation to heterogeneity in A/ measured by PET. J Appl Physiol 82: 1163–1176, 1997.
  38. West JB. Respiratory Physiology (6th ed.). Philadelphia, PA: Lippincott Williams & Wilkins, 2000, p. 11–19.
  39. West JB and Matthews FL. Stresses, strains, and surface pressures in the lung caused by its weight. J Appl Physiol 32: 332–345, 1972.
  40. Wilson TA and Beck KC. Contributions of ventilation and perfusion inhomogeneities to the A/ distribution. J Appl Physiol 72: 2298–2304, 1992.
  41. Yang QH, Kaplowitz MR, and Lai-Fook SJ. Regional variations in lung expansion in rabbits: prone vs. supine positions. J Appl Physiol 67: 1371–1376, 1989.



This article has been cited by other articles:


Home page
 J. Appl. Physiol.Home page
K. Emami, R. V. Cadman, J. M. Woodburn, M. C. Fischer, S. J. Kadlecek, J. Zhu, S. Pickup, R. A. Guyer, M. Law, V. Vahdat, et al.
Early changes of lung function and structure in an elastase model of emphysema--a hyperpolarized 3He MRI study
J Appl Physiol, March 1, 2008; 104(3): 773 - 786.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
W. W. Lam, D. W. Holdsworth, L. Y. Du, M. Drangova, D. G. McCormack, and G. E. Santyr
Micro-CT imaging of rat lung ventilation using continuous image acquisition during xenon gas contrast enhancement
J Appl Physiol, November 1, 2007; 103(5): 1848 - 1856.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
S. R. Hopkins, D. L. Levin, K. Emami, S. Kadlecek, J. Yu, M. Ishii, and R. R. Rizi
Advances in magnetic resonance imaging of lung physiology
J Appl Physiol, March 1, 2007; 102(3): 1244 - 1254.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow All Versions of this Article:
98/6/2259    most recent
00245.2004v1
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 Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Månsson, S.
Right arrow Articles by Golman, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Månsson, S.
Right arrow Articles by Golman, K.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Visit Other APS Journals Online
Copyright © 2005 by the American Physiological Society.