Using small animal PET with 68Ga-radiolabeled human albumin microspheres (Ga-68-microspheres), we investigated the effect of posture on regional pulmonary blood flow (PBF) in normal rats. This in vivo method is noninvasive and quantitative, and it allows for repeated longitudinal measurements. The purpose of the experiment was to quantify spatial differences in PBF in small animals in different postures. Two studies were performed in anesthetized, spontaneously breathing Wistar rats. Study 1 was designed to determine PBF in the prone and supine positions. Ga-68-microspheres were given to five prone and eight supine animals. We found that PBF increased in dorsal regions of supine animals (0.75) more than in prone animals (0.70; P = 0.037), according to a steeper vertical gradient of flow in supine than in prone animals. No differences in spatial heterogeneity were detected. Study 2 was designed to determine the effects of tissue distribution on PBF measurements. Because microspheres remained fixed in the lung, PET was performed on animals in the position in which they received Ga-68-microsphere injections and thereafter in the opposite posture. The distribution of PBF showed a preference for dorsal regions in both positions, but the distribution was dependent on the position during administration of the microspheres. We conclude that PET using Ga-68-microspheres can detect and quantify regional PBF in animals as small as the rat. PBF distributions differed between the prone and supine postures and were influenced by the distribution of lung tissue within the thorax.
- pulmonary blood flow
- positron emission tomography
- 68Ga radiolabel
- human serum albumin microspheres
- prone position
- supine position
- small animal
colored or fluorescent microspheres have primarily been used for blood flow measurements in small animals. Intravenously administered microspheres larger than 10 μm are predicted to become trapped in the pulmonary capillaries. Their spatial distribution therefore reflects pulmonary blood flow (PBF; Refs. 11, 13, 14, 20, 32, 38). Data analysis for colored or fluorescent microspheres requires lung excision and post mortem processing. In contrast, by using radiolabeled microspheres, data for the blood flow distribution could be achieved noninvasively and in vivo (22, 35). A major advantage of positron emission tomography (PET) is its high sensitivity and the ability to quantify the concentration of radiotracers in the lung (24). The use of a dedicated small animal PET could allow this method to be translated to rodents. Therefore, we tested the feasibility of measuring PBF in rats using biodegradable, 68Ga-radiolabeled human serum microspheres (Ga-68-microspheres) and micro-PET.
Subsequently, and with a variety of radiotracers, positional changes in regional blood flow from upper to lower regions have been found in large animals and humans (4, 15, 23, 25, 33, 36). However, to the best of our knowledge there are no data regarding the influence of posture (prone and supine) on regional pulmonary blood flow in small animals. We performed two studies: the purpose of study 1 was to quantify the regional distribution of Ga-68-microspheres in prone and supine positions during spontaneous breathing in anesthetized rats.
Radiotracers that remain fixed in the lung vascular bed allow for repeated measurements in different positions, as shown with single photon emission computed tomography (SPECT; Ref. 27). When regional blood flow is imaged in different postures, differences between the images may result from different blood flow distribution patterns, from a shift in the distribution of parenchyma within the thorax, or a combination of both. It has been demonstrated that a change from supine to prone posture caused a change in the vertical distribution of lung tissue in humans. The effect on the vertical distribution of PBF was much less within the lung parenchyma (27). In a second study, the tracer distribution was measured twice, once after tracer administration and at another time point in the opposite position without a new injection of Ga-68-microspheres. Because the Ga-68-microspheres were entrapped in the pulmonary capillary bed, study 2 was performed to analyze how the measurement of blood flow distribution is affected by pure positional changes in the animal and may, therefore, be influenced by tissue attenuation.
Preparation and Radiolabeling of Microspheres
The preparation of chelator-bound microspheres was previously described in detail (34). Human serum albumin microspheres (HSAM; mean diameter: 20 μm) were provided by ROTOP Pharmaka, Radeberg, Germany. Gallium-68 was eluted as 68GaCl3 with 0.1 M HCl from a 68Ge/68Ga-Generator (Obninsk, Russia) and concentrated via an ion exchange column to a volume of 200 μl. HSAM were conjugated with p-SCN-Bn-DOTA. Dried 1,4,7,10-tetraazacyclododecane-tetraacetic acid-modified HSAM (DOTA-HSAM) were stored under argon at −18°C. DOTA-HSAM (1 mg) were suspended in 1 ml 0.5 M ammonium acetate buffer (pH 5). To this suspension, 100 μl of the 68Ga eluate was added, resulting in a final pH of 4.5. The suspension was shaken in a Thermomixer at 90°C for 15 min. Labeling yields were determined by comparing the activity of the supernatant after centrifugation, with that of the 68Ga-DOTA-HSAM washed three times using electrolyte solution E153 (Serumwerk Bernburg, Germany). This method resulted in reproducibly high yields of ≥95% microsphere-associated radioactivity. The 68Ga-DOTA-HSAM were washed immediately before application, reaching a radiochemical purity >99%.
Animals, Feeding, Husbandry, and Animal Preparation
The local animal research committee at the Landesdirektion Dresden approved the animal facilities and the experiments according to institutional guidelines and the German animal welfare regulations. The experimental procedure used conforms to the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes (ETS No. 123), to the Deutsches Tierschutzgesetz, and to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health [DHEW Publication No. (NIH) 82-23, Revised 1996, Office of Science and Health Reports, DRR/NIH, Bethesda, MD 20205]. Wistar rats (Wistar Unilever, HsdCpb: Wu, Harlan Winkelmann, Borchen, Germany) were housed under standard conditions with free access to standard food and tap water.
Thirteen male Wistar-Unilever rats weighing 340 (320–398) g [median (interquartile range)] were anesthetized with isoflurane (1.5–2.0%, first 6 animals) or furthermore (because of technical reasons) with desflurane (7.0–12.0%) and with a single intraperitoneal injection of ketamine (75 mg/kg). The guide value for breathing frequency was 65 breaths/min. Animals were put in the supine position and placed on a heating pad to maintain body temperature. The spontaneously breathing rats were treated with 100 units/kg heparin (Heparin-Natrium 25.000-ratiopharm, Ratiopharm, Germany) by subcutaneous injection to prevent blood clotting on intravascular catheters. After local anesthesia by injection of Lignocain 1% (Xylocitin loc, Mibe, Jena, Germany) into the right groin, catheters were introduced into the right femoral artery (0.9 mm leader cath, Vygon, Ecouen, France) for blood gas analysis and arterial blood pressure measurements, and a second catheter into the right femoral vein (0.8 mm Umbilical Vessel Catheter, Tyco Healthcare, Tullamore, Ireland) was used for continuous infusion (1 ml/h) of a balanced electrolyte solution (E153) and administration of the tracer. A tracheostomy tube was placed under additional local anesthesia (Lignocain 1%). Femoral arterial pressure was referenced to mid-chest level and was continuously monitored using a Component Monitoring System (Hewlett-Packard Model 54S, Saronno, Italy). Blood gases were measured with an automated blood-gas analyzer (model ABL 50, Radiometer, Copenhagen, Denmark).
Anesthetized, spontaneously breathing animals were allowed to stabilize for 30 min after preparation. The animals started in prone (n = 5) or supine (n = 8) position. A 10-min transmission scan was recorded during this time for each subject by using a rotating point source of 57Co. The transmission scan was used to correct the emission scan for g-ray attenuation caused by body tissues and supporting structures; it was also used to demarcate the lung field. After the transmission scans, physiological data [mean arterial blood pressure (MAP), heart rate (HR), respiration rate (RR), and body temperature (T)] were measured and arterial blood samples were taken for pH, PaO2, PaCO2, hemoglobin (Hb), and hematocrit (Hct).
The radioactivity of the injection solution was measured in a well counter (Isomed 2000, Dresden, Germany) cross-calibrated to the microPET P4 scanner (Siemens preclinical solutions, Knoxville, TN). The PET acquisition of 30-min emission scan was started and the infusion of the Ga-68-microspheres was initiated with a delay of 30 s. A solution of 0.5 cm3 of ∼10 MBq (272 μCi) Ga-68-microspheres was infused over 1 min (with a Harvard apparatus 44 syringe pump) into the right femoral vein. After completion of the PET scan, the assessment of physiological data was repeated and arterial blood samples were obtained; see also Fig. 1.
Seven male rats weighing 305 (263–338) g [median (interquartile range)] were anesthetized with desflurane (9.0–12.0%). The anesthetized animals breathed spontaneously.
Three animals were initially studied prone and, after a 10-min transmission scan, Ga-68-microspheres were administered into the tail vein followed by a 30-min emission scan as described above. Animals were turned over and the emission and transmission scans were repeated in supine position without a new administration of Ga-68-microspheres.
Four animals received administration of Ga-68-microspheres and PET measurement in the supine position, followed by a change to the prone position. Subsequently, assessment of the emission and transmission scan was repeated. Figure 1 provides an overview of the experiment.
At the end of the experiment, the animals were deeply anesthetized and killed by an intravenous injection of potassium chloride.
PET Image Analysis
Data acquisition was performed in 3D list mode. Emission data were collected continuously and list mode data were sorted into a single sinogram. The data were corrected for decay and attenuation, and frames were reconstructed by Ordered Subset Expectation Maximization applied to 3D sinograms (OSEM3D) with 14 subsets, six OSEM3D iterations, two maximum a posteriori iterations, and 0.05 beta-value for smoothing. We used the list mode to extract the mean values of the emission data. The voxel size was 0.8 by 0.8 by 1.2 mm3, and the image resolution in the center of field of view (CFOV) was 1.85 mm for the full width half maximum (FWHM) in both the tangential and radial directions. At a radial distance of 2.5 cm from the CFOV, the resolution decreased to 2.5 mm FWHM. Given the spatial resolution of the scanner, no recovery coefficient was applied to compensate for possible partial-volume averaging errors. The image volume data were converted to Siemens ECAT7 format for further processing.
The image files were then processed using ROVER software (ABX, Radeberg, Germany). Although the animals were carefully placed prone or supine in standard positions, differences between the animals in posture that could arise from factors such as distortion or bend of the spine could not be excluded completely. Coregistration of all datasets to one position had the advantage of standardizing conditions for further image processing. All the images were user guided and roughly coregistered to one master image with the best alignment in the supine position using a transformation algorithm. The supine position was chosen because of the preponderance of animals in supine posture, as seen in Fig. 2. The ROVER coregistration module corrected the residual differences in the relative positioning of the animals automatically, using a mutual information algorithm with the conjugate gradient maximization method. The data were transformed so that the lungs (ROI total) had the same dimensions and the same alignment in all three axes. As a result, we achieved 3D images in identical positions. Therefore, it was possible to use the same ROI mask for a specific lung region (i.e., ventral) for all images.
ROIs for the entire lung image (ROItotal) were obtained by definition of the maximum border of the organ in the coronal, sagittal, and transversal axis. Division of the imaged lung in the middle of ROItotal along the ventrodorsal and craniocaudal axis, and along the right and left lung, yielded three pairs of ROIs; see Fig. 2. Masks for defining regions of interest (ROI) were set in these coregistered image data sets for the right and left, the dorsal and ventral, and the basal and apical parts of the lung. As a result, PBF could be assessed from geometrically identical ROIs for all lung images. The ROIs were further defined by thresholding with 38% of the maximum activity in the total lung. The pulmonary radioactivity concentration and the volume for each ROI were determined. This activity concentration is proportional to the number of microspheres lodged in that region. The number of microspheres reaching the region is N × p, where N is the number of microspheres injected (corresponds to the total injected activity and the ROItotal) and p is the proportion of total flow to the organ, which perfuses the region. Thus p is also the expected proportion of microspheres captured by a specific region (28), which is equal to the fractional flow to ROI/ROItotal. The PBF value within each specific region was expressed as fractional flow (PBF/total PBF of the ROI, with the denominator calculated as the mean PBF of the region × the volume of the region, based on voxel dimensions) according to Schuster et al. (35).
Further data processing, such as linear regression analyses, were performed to extract a maximum of information from the study and to compare both analyzing methods through the use of software specifically developed in the Matlab environment (Matlab R2007b, The Mathworks, Natick, MA). The mean normalized voxel intensity in the imaged lung was regressed vs. the vertical distance from the most dependent point of the imaged lung. The vertical gradient was measured as the slope of the regression line and expressed in percent per centimeter, relative to the mean. The spatial heterogeneity of the images was assessed as the coefficient of variation (CV) of the voxel data within the lung field, defined as the SD normalized by the mean value of the data (23, 36).
Values are expressed as median and interquartile range (IQR). IQR is defined as the range of the values extending from the 25th percentile to the 75th percentile. Values were compared using an unpaired Student's t-test with Welch's correction and an F-test to compare the variances (GraphPad Prism 5.02 for Windows, GraphPad Software, San Diego, CA). The nonparametric Wilcoxon signed rank test and the D'Agostino-Pearson normality test were used for statistical evaluation for some of the data. Linear regression was used to compute vertical gradients (positive gradients indicate an increase from the dependent to the nondependent regions). A P value of <0.05 was considered significant and is indicated by an asterisk.
The mean values for physiological data are reported in Table 1. Differences between the prone and supine groups and between the time points in each group did not reach statistical significance, which demonstrates the physiological stability of the animals throughout the experiment.
Regional pulmonary blood flow.
Basal lung regions received about two-thirds of PBF compared with the apical regions in both groups. We registered slightly less activity in the left lung than in the right lung in both the prone and supine postures. Dorsal regions received more PBF per identical volume along the ventrodorsal axis. However, the regional blood flow was more pronounced in the supine posture [0.75 (0.74–0.78)] than in the prone posture [0.70 (0.66–0.74)]. Consequently, PBF was found to be lower in ventral ROIs. In the prone position, PBF in ventral ROIs were less reduced [0.30 (0.26–0.34)] than in supine-positioned animals [0.25 (0.22–0.26) (P = 0.037)]. The total lung volume in the prone animal group was 6.6 ± 2.4 vs. 5.5 ± 0.7 cm3 in the supine group. The approach of halving the dimension of the lung for each axis for ROI definition resulted in different lung volumes caused by the cone-shaped lung structure. Therefore, most lung volumes were obtained in the dorsal and basal ROIs. Lung volumes showed no differences in each corresponding ROI between the prone and supine animals.
PBF and lung volume for each ROI are presented in Table 2.
The vertical gradient (dependent to nondependent) of PBF was significantly steeper in the supine than in the prone animals, as shown by the data in Fig. 3. Supine animals presented a vertical gradient of −0.165%/cm (−0.171 to −0.126%/cm), whereas prone animals showed a gradient of 0.028%/cm (0.006–0.052%/cm; P = 0.0043; Fig. 3). The CV were equal with 0.20 (0.19–0.22) and 0.20 (0.17–0.21) in the supine and prone positions (P = 0.35), respectively.
The animals received an infusion of Ga-68-microspheres in the prone (group P, n = 3) and supine (group S, n = 4) positions. After initial measurements, the positions of the animals were switched and the measurements were repeated with no further addition of microspheres. The PBF and lung volumes for ventral and dorsal ROIs for each group showed no differences between the initial position and the opposite position, or between group P and group S, as presented in Table 3.
According to the data shown in Fig. 3, the vertical gradients of regional perfusion were significantly steeper in supine-positioned animals (group S) than in prone-positioned animals (group P). Supine-positioned animals presented a vertical gradient of −0.133%/cm (−0.142 to −0.119%/cm) in group S, whereas the prone-positioned animals had a gradient of −0.061%/cm (−0.064 to −0.040%/cm) in group P (P = 0.002). Changes to the opposite position resulted in a vertical gradient of −0.084%/cm (−0.126 to −0.071%/cm) in group S (prone position), and the vertical gradient in the supine position of group P reached a similar value, −0.087%/cm (−0.110 to −0.083%/cm). Although the change in posture to the opposite position diminished the difference in the vertical gradient between group P and group S, this effect did not produce significant changes in the vertical gradient within group S and group P. The difference in the vertical gradient between group P and group S in supine position (P = 0.035) indicates a dependence of blood flow distribution on the position during the administration of microspheres (Fig. 3).
No effect of position on CV within or between groups P and S could be observed. Group P demonstrates a CV of 0.207 (0.179–0.227) in prone and 0.227 (0.221–0.240) in supine position. CV in group S was 0.225 (0.213–0.238) in the prone and 0.225 (0.220–0.238) in supine positions.
We introduced a novel methodology using small animal PET and Ga-68-microspheres for noninvasive and repeatable quantification of PBF in vivo. The relatively high spatial resolution enables regional PBF differences to be assessed. In the present studies, the effects of posture on regional PBF in laboratory animals as small as the rat could be recorded.
The PBF was vertically distributed differently and exhibited a significant dependence on the posture, prone or supine, of the anesthetized, spontaneously breathing animals. PBF was more pronounced in the dorsal regions, even more when the animals were in the supine position.
We found that variance between the vertical gradients for prone and supine postures were the result of both the distribution of PBF within the lung and the distribution of lung tissue within the thorax.
Methodological Issues and Critique of the Technique
Most issues related to the use of microspheres for the measurement of regional blood flow in small animals have been discussed in the literature (9, 10, 13, 14, 20). Assessments of regional PBF were achieved by measuring trapped radionuclide-labeled microspheres in the arterial pulmonary circulation with PET (3, 32). Therefore, the number of microspheres trapped in the lung was proportional to the blood flow in these regions. Microspheres are carried by the blood flow into prealveolar capillaries. As shown recently in rats, polystyrene microspheres do not completely occlude the prealveolar capillaries in which they are lodged. Red blood cells can flow around the deposited microspheres and subsequent microspheres can take the same preferential pathways (20). When used in adequate numbers, microspheres do not alter regional flow or vascular tone in the rat pulmonary circulation. Vascular resistance is not significantly increased after injection of <100,000 microspheres and does not cause vasodilatation in rat circulation (13, 37). We used 70,000–100,000 microspheres per injection. The quality of the images was found to decrease with smaller numbers of injected Ga-68-microspheres, which is consistent with previous findings (12).
HSA microspheres, labeled with a suitable radionuclide, have been an approved drug for research purposes and lung scanning in clinical nuclear medicine for many years (31). Diameter, smoothness, density, chemical composition, and amount must be taken into account when considering microsphere biodistribution (11, 28, 34). These microspheres are in routine use for perfusion studies because of their biocompatibility as well as their uniform size and biokinetics (41). In addition, several studies have demonstrated their excellent in vivo stability. Within 24 h, <10% of rare earth activity was released from albumin microspheres (6). HSA microspheres, such as those used in this study, showed high in vitro stability in human serum, with only 8% loss of radioactivity from the surface after 48 h. The in vivo stability measured in healthy rats was below 1% loss of activity at 1 and 12 h after injection, and below 9% at 24 and 48 h after injection of 86Y-labeled microspheres (34). The stability of Ga-68-microspheres was far above the observation time in this study.
HSA microspheres of diameters 10–30 μm were found suitable for lung imaging in a canine (7). As determined in a pilot study, all Ga-68-microspheres with a diameter of 20 μm were completely trapped in the lung. The distribution of 20 μm sized Ga-68-labeled macroaggregates correlates well with pulmonary perfusion, as determined by 15O-water in dogs (22) and humans (35). HSA microspheres of this magnitude (average diameter 19.5 μm) localized selectively in the capillary bed and blood vessels of the lung tissue with no adverse side effects in mice and rats (30). After administration of Ga-68-microspheres in our study, no changes in physiological variables were found, suggesting hemodynamic stability.
Data Acquisition and Analysis
PET is the preferred method to achieve quantitative data of PBF (22, 35) presented as a three-dimensional image. Advances in detector technology have made it possible to produce quality images in small animals, with an intrinsic/image spatial resolution of 1.85 mm. Given the small size of our animals, we used the highest spatial resolution of the PET camera compatible with an appropriate signal-to-noise ratio. The image volumetric resolution in center of field of view was 6.4 μl, according to a 1.15 mm spherical volume radius in the center of the field. This resolution is analogous to the spherical volumes (radius = 1 mm) in rat lungs obtained by using an imaging cryomicrotome in a previous high-resolution measurement of PBF (5). In addition to the numbers of injected HSAM, the length of data acquisition, and the reconstruction mode, PET scans were influenced by the β+ range of the radionuclide and by the lung movements.
We used 68Ga, whose physical half-life (67,629 min) is well suited for PET studies (22, 35, 38). The β+ range of 68Ga does not impair spatial resolution by more than 2.4 mm, and the fraction of additional γ-radiation is low (3%), which suggests a comparatively low radiation dose. The advantage of using 68Ga, a generator-produced positron emitter, is the independence from a cyclotron.
With the assumption of having unchanged conditions, the length of data acquisition (30 min) was chosen to receive a high yield of activity and to optimize the quality of the images. During the observation time, the influence of factors such as formation of atelectasis cannot be excluded with this method.
PET images were all acquired in vivo during spontaneous breathing; therefore, we received images representing the PBF during the complete breathing cycle. These PET images were expected to be affected by the elongations of the lungs along the ventral-dorsal and apical-basal axis during the data acquisition. However, posture-mediated changes in ventilation were found primarily to be oriented along the caudocranial axis (2). Along this axis, expansion of the lung parenchyma during breathing could result in greater gas content, especially in different positions. Depression of ventilation in spontaneously breathing anesthetized rats was reported to be primarily due to a decrease in the respiratory rate (19, 21), not in tidal volume. The respiratory rate remained equal (see Table 1), suggesting stable respiration during this study. The effect of posture on lung dimension was minimized by coregistration of the lung images to one “master” image in supine position, resulting in nearly identical lung dimensions. Because the entire amount of the Ga-68-microspheres was trapped in the lung exclusively, the edge of the organ could be defined easily and all given Ga-68-microspheres could be included in the analysis. Spatial averaging in PBF could not be excluded along the border of adjacent ROIs (i.e., basal/apical) without gating correction. Minimizing the effect of breathing, we used ROIs with a large length scale.
Geometrical lung expansion was used to define ROI. This might have produced an overrepresentation of lung volume in dorsal ROIs, which could have resulted in an underestimation of differences in PBF in dorsal and ventral ROIs in the prone and supine postures. Linear regression analysis was performed over the total lung volume, without using subdivisions, in several ROIs. This may explain the absence of differences in PBF between ROIs and position in study 2, with a smaller number of animals, whereas we found detectable differences in vertical gradients between postures.
Body Position and PBF
In this study, posture had no influence on physiological variables. This is consistent with previous studies comparing healthy animals (2, 33). Pulmonary perfusion was found to be similar between the rat and other larger species (14). Posture-mediated changes were primarily oriented along the ventrodorsal axis for blood flow, as seen in CT studies (8, 17, 29). The finding of our study that substantial perfusion was preserved in dorsal lung regions in the prone position is consistent with previous observations in healthy large animals (2, 4, 15, 36, 39). The gravitational gradient was different in the supine and prone positions according to a PET study in dogs (36), contrary to PET measurements in humans (23). The difference between the gradients for prone and supine (p/s) in study 1 was greater than the difference between prone and supine (pp/ss) in study 2, as shown in Fig. 3. The small number of animals used in study 2 may account for this difference. Additionally, studies 1 and 2 were performed differently than each other. The Ga-68-microspheres were administered by different routes (central venous line/tail vein), and the animals in study 1 breathed through a tracheostomy tube, which may have resulted in differences in the work of breathing and airway resistance between the two study groups. Another difference between studies 1 and 2 is the time of anesthesia before the Ga-68-microspheres were injected. Taking the preparation of the animal into account, the time before the Ga-68-microspheres could be administered was ∼1 h later in study 1, with enhanced impact of anesthesia and its effects on the lungs in that group.
We found a further increase in PBF in the supine position in dorsal regions, in accordance with studies in pigs (2), dogs (36), and sheep (39). This finding is difficult to explain by a gravitational mechanism in the rat lung. Therefore, other important factors such as the vascular geometry (15), higher vascular conductance in these regions (4, 26), and the weight of the heart (1, 8, 29, 40) might play a role in small animals. Although all corresponding volumes in the several ROIs were not significantly different between the supine and prone positioned animals, the lung is an elastic structure and gravity causes the lung parenchyma to stretch at the top and compress at the bottom (16, 29). Changing posture modified the distribution of the lung volume, thus a portion of the lung tissue was compressed by the heart weight in the supine position and released while prone. The extent to which these changes occurred was not assessed by CT or MRI in this experiment. Consequently, study 2 was performed to investigate the impact of such changes in lung structure on PBF.
Effect of Posture on Tissue Distribution
Once fixed in the parenchyma, the Ga-68-microspheres serve as tissue markers. If administration of Ga-68-microspheres is followed by image acquisition in both postures, the image change in the distribution of microspheres serves as an indication of the redistribution of lung tissue. The effect of a true vertical gradient in prone and supine positions could be preserved after correction for tissue volume in humans (17). As shown in Fig. 3, the vertical gradient changed to some degree by turning the animals from prone to supine position, reflecting an increased distribution of Ga-68-microspheres toward the dorsal regions. By changing the position from supine to prone, the vertical gradient was more pronounced toward the ventral region, as a result of a shift of lung tissue within the thorax. These data support previous findings from computer tomography scans, which showed that variations in lung density affect the interpretation of vertical gradients (16, 29). Although lung tissue is compressible, the assumption of greater tissue density in dependent regions due to gravitational effects should be even smaller in rats, because of their smaller lung volumes. In contrast, in study 2, after correction for a shift in lung tissue, we found that an effect of posture on vertical gradient resulted in a dorsal predominance in supine (but imaged prone) identical to that in prone (but imaged supine), probably caused by a greater number of alveoli and blood vessels per unit volume in the dorsal region. These data from spontaneously breathing healthy rats support a recent SPECT study in healthy humans (27) that suggested that microparticle distribution is determined by the posture during imaging, not the posture during the flow measurement. Otherwise, the difference in the vertical gradient between the prone injected group and the supine injected group in supine position indicate a dependence of blood flow distribution on the position during the administration of microspheres (Fig. 3).
The heterogeneity of pulmonary blood flow did not change between supine- and prone-positioned animals. The fact that a pronounced heterogeneity in supine animals did not reach significant levels, despite clear differences in the flow distribution, reflects a large variability in perfusion along the vertical axis, similar to results previously reported in humans (29), but in contrast to results in dogs (36).
The heterogeneity of PBF was normalized differently in the literature. However, CV values reported by Glenny et al. (CV 0.44 supine, CV 0.39 prone; Ref. 15), Chon et al. (CV 0.32 supine, CV 0.28 prone; Ref. 8), and Prisk et al. (CV 0.71 supine, CV 0.70 prone; Ref. 29) were greater than those from our experiment (CV 0.20 supine, CV 0.20 prone). Treppo et al. (36), using PET, reported the value of spatial heterogeneity of PBF as CV 0.41 supine, CV 0.21 prone. Their results in the prone position were similar to ours, but in the supine position, our results were again smaller than their findings. The fact that blood flow heterogeneity was smaller in our experiment was not surprising. The expansion of the lung parenchyma during breathing may have had an effect on CV, resulting in an underestimation of regional heterogeneity because of spatial averaging. Otherwise, the spatial flow heterogeneity from our experiment in anesthetized, spontaneously breathing rats was relatively close to that from an experiment in awake spontaneously breathing sheep (CV 0.29; Ref. 39).
The heterogeneity of flow has been shown to depend on the sample size. The absolute lung volumes from our experiment ranged between 4.5 and 6.5 cm3, which were much higher than the volume resolution of the system in the center of the field of view (0.0064 cm3) and 2.5 cm away from this point (0.0156 cm3). In rats, at a sampling volume of 200 mm3, CVs of 0.102 (18) and 0.113 (14) were reported. By reducing the sample size to 0.52 mm3, the heterogeneity was increased to a CV of 0.31 (14). Taking the + energy range of 68Ga into account, the smallest sampling unit in our study was 7.24 mm3. Hence, our sample size was between the values reported by Glenny et al. (14), and flow heterogeneities from our experiment were between these values as well.
In summary, we demonstrated the utility of measuring pulmonary blood flow in intact rat lungs with Ga-68-microspheres and small animal PET. This technique represents an additional tool for in vivo measuring of PBF in small animals. Further optimization of this novel method might be achieved by using tracers with a smaller + energy range and by gating.
In addition, we showed that posture plays a role in blood flow distribution studies in healthy, spontaneously breathing, and anesthetized rats. Compared with the supine position, the prone position reduced regional blood flow in dorsal lung regions without changes along the caudocranial axis. This effect of posture on PBF was influenced by a shift in lung tissue within the thorax.
No conflicts of interest are declared by the authors.
The authors thank J. Schlesinger for preparation of the radioisotope, J. G. Venegas for technical support with software, and T. Koch for administrative assistance.
- Copyright © 2010 the American Physiological Society