Published in Medical Imaging 1996: Physiology and Function
from Multidimensional Images, Eric A. Hoffman, Editor, Proc. SPIE 2709, 40-54
(1996).
Xenon Enhanced CT Imaging of Local Pulmonary Ventilation
Jehangir K. Tajik, Binh Q. Tran and Eric A. Hoffman
Division of Physiologic Imaging
Department of Radiology
University of Iowa College of Medicine
Iowa City, IA 52242
[ home
| search
| contact us
]
Table of Contents
- Abstract
- Introduction
- Methods
- Results
- Discussion
- Conclusion
- Acknowledgements
- References
We are using the unique features of electron beam CT (EBCT) in
conjunction with respiratory and cardiac gating to explore the use of
non-radioactive xenon gas as a pulmonary ventilation contrast agent. The goal
is to construct accurate and quantitative high-resolution maps of
local pulmonary ventilation in humans. We are evaluating
xenon-enhanced computed tomography in the pig model with dynamic
tracer washout/dilution and single breath inhalation imaging
protocols. Scanning is done via an EBCT scanner which
offers 50 msec scan aperture speeds. CT attenuation coefficients
(image gray scale value) show a linear increase with xenon
concentration (r=0.99). We measure a 1.55 Hounsfield Unit (HU)
enhancement (kV=130, mA=623) per percentage increase in xenon gas
concentration giving an approximately 155 HU enhancement with 100%
xenon gas concentration as measured in a plexiglass
super-syringe. Early results indicate that a single breath (from
functional residual capacity to total lung capacity) of 100%
xenon gas provides an average 32
1.85 (SE) HU enhancement in the
lung parenchyma (maximum 50 HU) and should not encounter unwanted
xenon side effects. However, changes in lung density occuring during
even short breath holds (as short as 10 seconds) may limit using a
single breath technique to synchronous volumetric scanning, currently
possible only with EBCT. Preliminary results indicate
close agreement between measured regional xenon concentration-time
curves and theoretical predictions for the same sample. More than 10
breaths with inspirations to as high as 25cmH2O airway pressure were
needed to clear tracer from all lung regions and some regions had
nearly linear rather than mono-exponential clearance curves. When
regional parenchymal xenon concentration-time curves were analyzed,
vertical gradients in ventilation and redistribution of ventilation at
higher inspiratory flow rates were consistent with known pulmonary
physiology. We present here a works in progress, showing results from
two pigs illustrating the high resolution and detailed regional
information obtainable with careful attention to cardiac and
respiratory gating during a multi-breath washout protocol.
Keywords: xenon, pulmonary imaging, computed tomography,
regional ventilation, physiologic imaging, electron beam CT, lung,
pulmonary physiology
Calculating regional pulmonary ventilation requires measuring the rate
of delivery (wash-in) of a tracer to lung regions or the clearance
rate (washout) of a tracer from lung regions. Early studies of
pulmonary ventilation used radioactive isotopes of chemically inert
gases such as 133xenon as the tracer and measured either the buildup of
radioactivity (wash-in study) or the decrease in radioactivity
(washout study) of lung zones over multiple breaths
[2]. More recently, the enhanced spatial resolution provided
by dynamic x-ray CT has been used clinically in conjunction with
inhalation of stable (non-radioactive) 131xenon gas as a contrast agent
to measure local pulmonary ventilation
[14,1,15,16,20,8,5,6] and
regional cerebral blood flow [11,19,7,13]. Though there
have been several studies using dynamic x-ray CT in conjunction with
stable xenon to calculate local pulmonary ventilation, there has been
little if any information generated as to the accuracy and validity of
these measurements or studies utilizing accurately controlled
simultaneous respiratory and cardiac gating.
Stable (non-radioactive) xenon gas has a K-edge similar to iodine and
should, therefore, be a potent x-ray attenuator and provide good
contrast enhancement when used in conjunction with CT
scanning[17]. Though xenon gas is soluble in blood and
tissues[18], it is chemically inert and thus has minimal
physiologic effects. Xenon gas has been demonstrated to have an
anesthetic effect when concentrations of 40% or greater are breathed
for prolonged periods[4].
We hypothesized that it may be possible to ascertain the relative
distribution of pulmonary ventilation by having the subject take a
single deep breath of a high xenon gas concentration mixture
(i.e. 80% xenon/20% oxygen or 100% xenon). A single breath of a
higher xenon gas mixture should not encounter unwanted xenon side
effects as mean alveolar xenon concentration from a single breath of
this gas mixture should not exceed 40%. Subtraction of identically
gathered xenon enhanced and unenhanced images should yield information
regarding the relative distribution of local pulmonary
ventilation. The enhancement due to xenon normalized to regional
alveolar gas volume, yields direct assessments of well ventilated
vs. poorly ventilated lung regions. It is crucial to reproduce
the same lung volume for the enhanced and unenhanced scans for proper
image registration. Early xenon studies were poorly controlled for lung
volume and thus suffered from serious mis-registration artifacts which,
when uncorrected, leave the results uninterpretable for regional
physiologic assessment.
Another method for quantitating regional pulmonary ventilation
involves sampling regional tracer delivery or clearance curves followed by
mathematical analysis of the concentration-time curves to calculate
regional kinetic rate constants or turnover rates.
Previous work by others regarding x-ray CT assessments of local pulmonary
ventilation have assumed the kinetics of xenon tracer gas filling or clearing
the alveoli to follow the mono-exponential model described by
Kety[12]. Kety's model of inert gas exchange at the lungs has been used
because it describes the buildup or clearance of an inert gas as a
simple exponential function whose slope, when plotted on semi-log
scale, is directly proportional to the fractional ventilation
(V/Vgas). Using this relationship, some previous CT studies of
local pulmonary ventilation were limited to simply
measuring two time points, before and after xenon inhalation,
rather than accurately sampling the wash-in and/or wash-out curves.
where Ct is the concentration of inspired gas at time t,
is the concentration (amount) of gas present in the
lungs at time infinity (i.e., at equilibrium) and is related to volume,
and k is the rate constant which relates alveolar ventilation to
alveolar volume (V/Vgas).
The Kety model assumes that the inert tracer gas has zero solubility
in blood. Though this assumption is obviously incorrect for xenon gas
as demonstrated by its anesthetic effects, previous CT assessments
have ignored xenon solubility. Moreover, the Kety model has several
assumptions and gives an imperfect fit to experimental ventilation
data from radioactive 133Xe studies, leading some investigators
to develop complicated multi-compartment models of inert gas uptake
and exchange to better fit the data. We believe that it may be
possible to correct for xenon solubility in blood without resorting to
complicated multi-compartment analysis by using the high temporal
resolution capabilities unique to electron beam CT[3]
(EBCT) to dynamically follow the xenon concentration in
blood over time simultaneously with the alveolar xenon
wash-in/wash-out study. Correction for xenon solubility in blood can
be made by knowing the mixed venous xenon concentration as a function
of time (i.e., xenon blood concentration entering the pulmonary
circulation, sampled in the right ventricle) and the local pulmonary
blood content (either directly from a corresponding perfusion scan or
estimated by knowledge of the relationship between regional alveolar
gas volume and regional blood volume if a perfusion scan is not
done). Another possibility would be to assume that xenon is soluble in
everything that is not air (i.e. pulmonary parenchyma tissue and
blood) and this volume of air and non-air fraction can be directly
computed from an unenhanced scan.
Our goal is to develop technique(s) using stable xenon gas as a
contrast agent in conjunction with the unique capabilities of dynamic
x-ray CT imaging that will allow constructing accurate, quantitative
and valid high resolution maps of regional pulmonary ventilation in
humans. One of the major drawbacks of previous x-ray CT assessments
of local pulmonary ventilation has been the lack of accurate lung
volume control over multiple breaths, so that misalignment between
unenhanced and enhanced scans is considerable and a possible source of
significant error. In contrast to previous work by others, we are
developing the capability to sample the majority of the local alveolar
xenon filling and/or clearance curves by accurately tracking and
controlling lung volume in patients over multiple breaths during a
stable xenon wash-in or washout study. Using the two modes of the EBCT
scanner (high spatial resolution and high temporal resolution), a
dynamic washout study is designed to image regional tracer gas
clearance (time-intensity data) during tidal breathing and a single
breath protocol gathers data for subtraction of a xenon enhanced scan
from an identical scan prior to xenon inhalation. The multi-breath
washout study provides information on the regional distribution of
pulmonary ventilation under dynamic/tidal breathing conditions while
the single breath protocol provides information on the distribution of
ventilation under static conditions. The physiology of these two
conditions is clearly dissimilar.
In order to generate accurate high-resolution maps of pulmonary
ventilation from detailed regional measurements throughout the lungs,
both imaging protocols employ respiratory and cardiac gating during
image acquisition to reduce density changes due to cardiogenic motion
artifact or respiratory movement rather than due to xenon
enhancement. Both methods rely heavily on accurate registration of the
multi-time point data and hence on accurate lung volume control to
assure that scanning occurs at identical lung volumes.
In parallel with these studies we are developing the lung volume
control equipment and software needed for human ventilation
studies. In these initial animal studies, we were able to control
airway pressure (hence lung volume) either via our computer controlled
ventilator or by using a calibrated super-syringe during scanning.
Scanning protocols utilize an Imatron C-150L EBCT
scanner[3] (figure 1). This scanner differs
from conventional x-ray CT scanners in that a focused electron beam is
swept along tungsten targets surrounding the subject instead of
mechanically rotating an x-ray source around the subject. This setup
allows for fast tomographic image acquisition with scan apertures on
the order of 50-100 milliseconds as opposed to approximately 0.6-1.0
second scan apertures from most other commercial x-ray CT scanners.
The fast scan times available with EBCT allow stop action images of
the heart to be obtained and allow us to minimize lung density changes
induced by the effect of cardiac motion.
Figure 1:
Cross-sectional view of the Imatron EBCT scanner. See text for
details. (With permission of Imatron, Inc.)
The EBCT scanner operates in a high temporal resolution mode
(multi-time point imaging) or in a high spatial resolution mode
(volumetric imaging). In the high temporal resolution mode, the
electron beam is magnetically steered sequentially along 4 tungsten
target rings encompassing 210 degrees about the subject. X-rays
produced from each target ring are focused onto two detector rings so
that up to 8 spatial levels may be acquired at each time point. Each
pair of images is gathered within 54 milliseconds followed by a 8
millisecond pause as the beam is reset for the next target ring. All
8 spatial levels are therefore acquired within 224 milliseconds. The
axial resolution for the multi-time point imaging mode is 1.5-3.0 mm
if only a single tomographic slice is to be acquired over multiple
time points or 7 mm if the multi-slice mode (up to 8 spatial
tomographic levels covering 7cm) is used in conjunction with multi-time
point imaging. This mode of scanning was used to acquire regional
tracer clearance data (i.e. concentration-time data) during the
multi-breath washout studies (see below).
In the high spatial resolution mode, a single tomographic slice
(1.5-3.0 mm thickness) is acquired after which the patient table is
advanced to the next tomographic image level. The higher spatial
resolution of this mode is achieved by focusing the generated x-rays on
a double-populated detector ring during a double sweep of a single target
ring. The scan aperture in this mode is 100 milliseconds and images
can be reconstructed using a high spatial frequency algorithm to
enhance anatomic detail. This mode of scanning was used during the
single breath imaging protocols (see below).
To ascertain the relationship between xenon concentration and x-ray
attenuation (CT number or Hounsfield Unit), we imaged xenon
concentrations of 100%, 80%, 60%, 30%, 15% and 0% (room air) in
a plexiglass super-syringe with EBCT scanner settings of kV=130 and
mA=623 during imaging. Mean HU was plotted against xenon gas
concentration and the slope of the relationship determined by linear
regression. This relationship is used to predict regional xenon
enhancement after equilibration to a given percent xenon concentration
as a verification of our initial measurements as is discussed below.
We report preliminary results here from two pigs. All animal studies
were performed within guidelines for animal care adhered to by the
American Physiological Society and the American Heart Association. The
animal use protocols were approved in advance by the University of
Iowa Institutional Animal Care and Use Committee.
Pigs were premedicated with I.M. ketamine (1 mg/kg) and acepromazine
and anesthesia induced with bolus (2.5 mg/kg) propofol (Diprivan) in
an ear vein. A cuffed endotracheal tube was placed and the pig
mechanically ventilated (10 ml/kg) with a respirator. Under
fluoroscopic guidance, pressure catheters (Millar) were placed in the
right ventricle (pressure monitoring), right carotid artery (systemic
blood pressure, blood gas monitoring) and right ventricular outflow
tract (pulmonary artery pressure, non-ionic contrast injection, wedge
pressure) for measurement and recording of physiologic
information. After catheterization, pigs were transported to the EBCT
facility and placed supine on the scanner table and connected to a CWE
9000 computer controlled respirator. ECG, airway pressure, systemic
and right ventricular pressures were monitored using an IBM compatible
computer equipped with a standard laboratory A/D converter board and
recording software. The scanner x-ray on pulse was also recorded to
allow exact correlation of physiologic and scanner
events. Throughout the catheterization procedures and scanning
protocols, deep anesthesia was maintained with propofol (12 mg/kg/hr)
titrated to heart rate and reflexes. At the conclusion of scanning,
the pig was euthanized with sodium pentobarbital (30 mg/kg) and
concentrated KCl.
Immediately preceeding each scanning protocol, the lungs
were inflated to 25 cmH2O airway pressure 3-5 times to assure a
constant volume history.
The imaging protocol for the single breath study is illustrated in
figure 2. The EBCT scanner is used in the high-spatial
resolution mode (see above) so that a volumetric image spanning the
apex to base extent of the lungs is gathered by advancing the patient
table 3mm after the acquisition of each tomographic slice. Prior to
the acquisition of the first slice, the subject inhales a single, deep
breath of ambient air (unenhanced or baseline scan) after which
respiration is suspended at a pre-selected lung inflation level
(i.e. ambient, 15cmH2O, 25cmH2O airway pressure). After the onset of
apnea, image acquisition begins with each tomographic level (slice)
gathered with a 100 millisecond scan aperture ECG gated to the peak of
the QRS complex. In the animal studies reported here, respiration
remained suspended during the image acquisition period (approximately
40 seconds) and resumed immediately following the acquisition of
the last image. The identical scan protocol is repeated substituting
a single deep breath of a high xenon gas mixture (xenon enhanced scan)
for the ambient air breath prior to imaging. In both instances, a
stack of 40 3mm contiguous tomographic slices were gathered to
generate a volumetric image data set of the lungs.
Figure 2:
Sample imaging protocol for the single breath inhalation
method. Top trace: scanner on pulses; Middle trace: ECG; Bottom
trace: airway pressure along with scanner on signal (square
wave). Note that scanning is triggered to the ECG signal
during a period of extended apnea.
We sampled the washout curve at 10 or 20 time points with 1, 4 or 8
spatial levels acquired at each time point. It is of interest to
determine the minimum and optimum number of time points (scans) needed
to accurately represent the xenon clearance curve so that x-ray
exposure/dosage can be minimized in human studies. It is critical that
each time point (scan) be acquired at the same lung volume (i.e. slice
registration) so that density changes within a parenchymal sample are
due to local changes in xenon gas concentration rather than
contaminated with density changes due to cardiogenic or ventilatory
motion artifact.
For this multi-breath/dynamic protocol, both respiratory and cardiac
gating is a requirement for proper image alignment at all time
points. An example imaging protocol is illustrated in
figure 3. For these studies, the EBCT scanner is used
in the high temporal resolution mode, and the first image is acquired
at equilibration of the lungs with a known xenon gas mixture. Cardiac
gated image acquisition is accomplished by programming the scanner to
acquire a multi-level data set every N+3 heart beats (where N is the
number of beats required to accomplish a tidal breath) with a scan aperture
of 50 milliseconds gated to 40% of the R-R interval. Concomitant
respiratory gating was accomplished in the animal studies reported
here via our customized computer controlled ventilator programmed to
take a single breath during the N R-R intervals between successive
image acquisitions. The timing (N+3) of the scanning was adjusted so
that the animal had always returned to end expiration before the next
set of images were gathered. This timing required customization for
the animals passive expiration and for the tidal volume used in a
given protocol.
Figure 3:
Sample imaging protocol for the multi-breath washout method.
Top trace: scanner on pulses; Middle trace: ECG; Bottom
trace: airway pressure along with scanner on signal (square
wave). Note that scanning is triggered to both the respiratory
signal and to the QRS complex of the ECG.
Questions being addressed include cardiac mixing effects and temporal
stability of the single breath tracer distribution, repeatability of
the regional washout measurements and matching of the measured
ventilation distributions with known pulmonary physiology.
Since the ECG gated volumetric data set in the single breath imaging
protocol is gathered sequentially over roughly 40 seconds (largely
determined by the heart rate and the scanners ability to follow a
given heart rate), tracer xenon gas could be redistributed do to
cardiogenic mixing leading to errors in the observed tracer
distribution (i.e. tracer now redistributed due to mixing action of the
heart on the lung parenchyma rather than representing the static
distribution of ventilation.) To explore the possible effects of this
phenomenon on the single breath change in HU distribution, we repeated
the single breath imaging protocol using the high temporal resolution
mode of EBCT scanning to follow possible density changes in a single
tomographic section over time during breath hold at FRC after a breath
of 100% xenon. Scanning was again gated to the ECG waveform during
apnea so that any changes in CT number (HU) seen during such an
experiment should represent changes due to redistribution of the
radiopaque xenon gas.
We repeated the multi-breath imaging study (with peak airway pressure
of 15 cmH2O for the intervening breaths) three consecutive times to
test the repeatability of the regional tracer washout/dilution
measurements. During these studies we increased the temporal
resolution to include 3 baseline points acquired during ECG gated
image acquisition during a brief period of apnea at equilibration
followed by gathering the regional time-intensity data over 17
breaths.
Matching of the measured ventilation distributions to expected pulmonary
physiology was evaluated by ascertaining the presence of vertical
(ventral-dorsal axis in a supine pig) ventilation gradients and by repeating
the muti-time point imaging protocol with 3 different inspiratory flow
rates.
- We analyzed regional xenon dilution curves from the dependent and
non-dependent lung zones to see whether the relationship between the
washouts in the two regions matched with known physiology. Data from
11 regional clearance curves in the dependent lung zone were averaged
and plotted on the same scale as the average xenon clearance curve
from 11 samples in the non-dependent lung region.
- The inspiratory flow rate has been previously shown to affect
the distribution of ventilation within the lungs . To examine
whether our washout technique could detect a redistribution in
ventilation at different flow rates, we repeated a washout imaging
protocol with inspiratory flow rates for the intervening breaths of 15
liters/min, 20 liters/minute and 25 liters/minute. The peak airway
pressure for the intervening breaths during each washout study was 15
cmH2O, so that the tidal volume remained constant across all 3
studies. Clearance data were gathered for 11 samples in the dependent
lung region and 11 samples in the non-dependent region and the average
clearance curve for each region plotted. The identical sampling
locations were analyzed across all 3 data sets.
Figure 4:
Data showing the linear relationship between radiopacity and
xenon conentration as measured in a plexiglass super-syringe. Data
was collected using xenon concentrations of 100%, 80%,
60%,30%, 15% and 0% (room air). Scanner mA and kV settings
were 623 and 130 respectively.
As shown in figure 4, CT x-ray attenuation is linearly
related to xenon concentration over the entire range of xenon
concentration. Linear regression analysis shows a slope of 1.55 HU
enhancement per percentage increase in xenon concentration when
calibrated using the EBCT scanner (kV=130, mA=623).
A representative subtraction image obtained from a single breath
protocol using 100% xenon as the ventilation tracer gas is depicted
in figure 5. In this imaging study, the lung was
held at 25 cmH2O airway pressure (i.e. approximately at total lung
capacity) during both the xenon-enhanced and un-enhanced scans. The
un-enhanced scan was digitally subtracted from the xenon enhanced
scan. The subtraction image shows minimal edge effects (some edge
effects are seen around the diaphragmatic border) indicating that the
enhanced and unenhanced scans were correctly registered. The
resulting subtraction image, therefore, should depict the distribution
of change in Hounsfield units through the lung field due to a single
deep breath of 100% xenon gas and therefore regional ventilation.
The mean enhancement from the single deep breath of 100% xenon gas is
321.85 (SE) HU in the lung parenchyma (maximum 50 HU).
Figure 5:
Subtraction image from an anesthetized pig pre- and post-inhalation of
a single breath of 100% xenon gas (both scans at 25
cmH2O airway pressure). There are two major features of
this image: 1) The lack of edge misalignment artifact
demonstrates our ability to exquisitely control the return to
a fixed lung volume following intervening breaths, and 2) The
significant enhancement of the lung field after a single
breath of 100% xenon gas suggests that a single breath of a
high xenon concentration gas mixture provides an adequate
signal for regional ventilation measurements to be made.
The results from the temporal stability of the single breath
distribution study are depicted in figure 6 with the
time-intensity data from representative samples from the dependent
(dorsal) and non-dependent (ventral) lung regions in the adjacent
panel. Notice that, in the dependent lung areas, the lung density
progressively increases over time while the density of areas in the
non-dependent lung regions remain relatively stable over the same
time. Some of the noise in the time-intensity curves can be attributed
to improper cardiac gating by the scanner. At first look, the increase
in CT number in the dependent lung regions could indicate that the
dense gas xenon is migrating to the dependent (lower regions of the
image) regions as time progresses. However, there are no concomitant
change in density in the non-dependent regions to indicate such an
event. Another possibility is that during the breath hold, oxygen is
being taken up faster in the dorsal lung regions (due to a greater
blood flow) without an equal exchange for carbon dioxide and thus
concentrating the xenon present within the region. Playing the multi-timepoint
image set as a cine loop showed that lung volume was diminishing
in the dependent regions. We feel that the increase in density over
time in the dependent lung regions during a breath hold after inhaling
a single breath of 100% xenon is most likely due to the well
recognized tendency towards dependent atelectasis during breath hold
due to normal oxygen/carbon dioxide exchange characteristics and due
to associated concentration of xenon gas.
Figure 6:
Following the indicated regions of interest (ROI's) on the 3mm thin-
slice image during apnea reveals appreciable
Hounsfield Unit (i.e. density) changes over time. In the
dependent regions, the observed changes indicate a monotonically
inceasing change in HU over time, whereas the nondependent zones
show minor fluctuations in HU over the same time.
As an initial validation of the equilibrium and washout data obtained
from the multi-breath imaging protocol, we used the equation below to predict
the equilibrium HU value of each sampled lung region based on regional
lung air content, xenon gas concentration used during a
given equilibration study and the linear relationship between xenon
concentration and HU shown in figure 4.
HUeq = (Airfraction * C * %Xenon) + HUbaseline1
where HUeq is the predicted HU at equilibration, Air fraction
is the fraction of the sample that is air[9] (proportional
to HUbaseline), C is the slope of the HU vs. [Xe] relationship,
and HUbaseline is the HU after all tracer has been cleared from
the sample.
This equation was applied to data from washouts studies after
equilibration with 33% or 66% xenon gas mixtures. Using measured
values of Air fraction=0.52 of a given sample and C=1.55 HU/%Xe, we
computed and matched the following predictions to measured
observations:
- 33% xenon: HUeq = 0.52*1.55*33% + 522 = 549 HU
(measured = 550 HU)
- 66% xenon: HUeq = 0.52*1.55*66% + 522 = 575 HU
(measured = 574 HU)
The measured washout data match closely with the predicted values.
Figure 7 illustrates the high resolution and
detailed regional information obtainable with careful attention to
cardiac and respiratory gating during a multi-breath washout protocol.
For this imaging study, the lung was equilibrated with 66% xenon gas
as the ventilation tracer and the lung was imaged at 0 cmH2O airway
pressure (i.e. end expiration or functional residual capacity) at all
time points. The intervening breath between consecutive images was to
25 cmH2O airway pressure (i.e. TLC breaths). The first time point
acquired is the equilibration image before any tracer has been washed
out (diluted) due to inspirations.
Figure 7:
ECG and respiratory gated multi-breath imaging allows
sampling washout curves with high-resolution (approximately 6mm by 6mm
samples shown) in both dependent and non-dependent lung
regions. Clearance curves from the samples depicted indcate that
washout occured faster in the dependent lung regions than in the
non-dependent lung regions. Note the nearly linear clearance curve
for sample 2 in the nondependent lung zone.
Sample regions used to generate data shown in
figure 7 represent 6x6 pixel boxes on a 7mm thick
slice. In figure 7, change in HU is plotted
against breath number during the imaging study. Notice that the sample
from the dependent lung region has a faster clearance/dilution of
xenon relative to the two samples from the non-dependent region of the
lung, suggesting greater ventilation in the dependent regions. An
important finding from the time-intensity (concentration) data is that
samples in the non-dependent areas of the lung are not cleared (or
tracer not fully diluted) of tracer within the ten breaths scanned
(see figure 9). Another important point from this
image data is that some samples in the non-dependent lung regions have
almost linear rather than exponential clearance (dilution) curves as
assumed by the Kety model and assumed in all previous x-ray CT
assessments of pulmonary ventilation.
We repeated an identical imaging study as above (with peak airway
pressure of 15 cmH2O) three consecutive times to test the
repeatability of the regional measurements. The data are displayed in
figure 8 for four distinct samples in the lung
field. To better understand the time it takes lung regions to fully
washout to baseline, and to assure a good measure of the equilibrium
enhancement, during these studies we increased the temporal resolution
to include 3 baseline points acquired (ECG gated) during a brief
period of apnea at equilibration followed by gathering the regional
time-intensity data over 17 breaths. In all four sample locations
depicted in figure 8, xenon dilution curves from the
three repeat scans are nearly coincident, indicating that respiratory
and cardiac gated image acquisition during a washout study allows
reproducible detailed regional ventilation measurements throughout the
lungs.
Figure 8:
Repeatable washout curves were obtained from both dependent and
non-dependent lung regions when the same multi-breath
imaging protocol was repeated 3 consecutive times. Graphs A-D
correspond to ROI's A-D in the image. For each ROI, the 3 washout
curves from the thrice repeated experiment have good
correspondence with one another, demonstrating that our
multi-breath imaging protocol can give repeatable data.
An interesting finding from these repeatability scans is the increase
in density seen in the dependent lung zones during the brief apnea at
equilibrium (panels B and C of figure 8)
that parallels the findings from the single breath study
(figure 6). It appears that the
lung volume
diminishes appreciably within 5-10 seconds of breath hold. Also, note
that in panel D, xenon clearance appears linear rather than exponential.
We analyzed regional xenon dilution curves from the dependent and
non-dependent lung zones to see whether the relationship between the
washouts in the two regions matched with known physiology. Data from
11 regional clearance curves in the dependent lung zone were averaged
and plotted on the same scale as the average xenon clearance curve
from 11 samples in the non-dependent lung region. As can be seen in
figure 9, in the dependent lung
region, xenon tracer
gas is diluted much more rapidly than the non-dependent lung zones.
There is, therefore, evidence of a vertical gradient in ventilation in
this supine pig (ventral-dorsal gradient) which is consistent with
known pulmonary physiology.
Figure 9:
To examine the existence of vertical ventilation gradients, we
analyzed the relationship between the average clearance curves
of 10-11 ROI's placed in random locations (areas containing
major airways and blood vessels were avoided) in each the
dependent and non-depenedent lung regions in a supine
pig. Examining these ROI's during a washout study revealed a
significantly steeper washout curve in the dependent lung
regions than in the nondependent zones. Notice that the average
clearance curve from the nondependent lung region does not appear
to have returned to baseline (i.e fully cleared of tracer)
within 10 breaths.
As seen in figure 10, inspiratory
flow rate variations
used in this study had no effect on the distribution of air flow (as
indicated by xenon washout rates) to non-dependent lung regions in the
supine pig. In the dependent regions, however, washout is enhanced at
25 liters/minute inspiratory flow rate while there was no difference
between 15 and 20 liters/minute inspiratory flow rate (tidal volume
held constant). Since at high flow rates, inertial factors and hence
pathway geometry becomes more important in the distribution of air
flow (ventilation), our data indicate that the dorsal, diaphragmatic
region of the pig lung (dependent lung region in a supine pig)
receives an enhanced air flow distribution at higher flow rates
perhaps do to favorable airway structure.
Figure 10:
Examining the effects of inspiratory flow rate on ventilation
distribution reveals that washout in non-dependent lung regions
in the supine pig are not affected by the rate of inspiration
over the 15-25 liter/minute inspiratru flow rate range. Tidal
volume was held constant in all 3 studies during this
experiment. The dependent lung zones of the supine pig, however,
show a faster clearance at 25 liters/minute inspiratory flow rate
and negligible differences between clearance rates at 15 and 20
liters/minute inspiratory flow.
Stable (non-radioactive) xenon gas is chemically inert and thus has
minimal physiologic effects while providing good radiopacity and
contrast as a ventilation tracer during CT imaging. We have
implemented a dynamic washout imaging protocol where we image regional
tracer gas clearance (time-intensity data) during tidal breathing and
a single breath protocol where we subtract a xenon enhanced scan from
an identical scan prior to xenon inhalation. The multi-breath washout
study provides information on the regional distribution of pulmonary
ventilation under dynamic/tidal breathing conditions while the single
breath protocol provides information on the distribution of
ventilation under static conditions.
Figure 11:
Picture of a pig airway cast. The main airway in a supine pig
traverses towards the dorsal (dependent) regions in the base of
the lung. We believe the geometric structure of the airways plays
an important role in preferential distribution of ventilation to
dorsal regions of the pig lung at higher inspiratory flow
rates. (From Amirav et. al., J. Appl. Physiol 75(5):
2239-2250, 1993). With permission from the American
Physiological Society.
While the single breath protocol is attractive due to the simplicity
of the imaging protocol and quantitative image analysis, our results
indicate that even with close attention to respiratory and ECG gated
image acquisition, density changes begin soon after the onset of a
breath hold, confounding assessment of regional ventilation. However,
adequate concentrations of xenon (HU changes) are achieved with a
single breath approach. Assessment of regional ventilation using a
single breath approach, therefore, requires near synchronous volume
scanning following a single breath of xenon, currently achievable
using the EBCT 50 msec mode of scanning.
While the contaminating density changes during breath hold may not
always be as dramatic as shown in figure 6
, they are
manifested within the first 5-10 seconds of apnea after a single
inhalation of high concentration xenon gas. Because of the time
course of this effect, using commercial scanners (other than the EBCT
scanner) currently available, where a volumetric image is gathered via
table advancements past a fixed scanning plane, use of the single
breath ventilation protocol should be limited to volumetric
acquisitions taking no longer than 5 seconds. If a volumetric data set
of the lungs cannot be obtained synchronously, calculated density
changes from digitally subtracting the enhanced and un-enhanced scans
of the single breath protocol need to be interpreted with caution. It
is of interest to note that this is not a problem using radioactive
tracers scans as in nuclear medicine since quantitation in that arena
is based on regional radioactive count rates rather than regional
density changes.
Unlike previous investigators who have used stable xenon gas in
conjunction with x-ray CT imaging for measurement of regional
ventilation, we sampled the entire regional tracer clearance
curve. Moreover, the previous
investigators[14,1,15,16,20,8,6,5]
were unable to use cardiac gated image acquistion because of long scan
times. One study did use an EBCT scanner, however, they did not report
cardiac gating[14]. Moreover, the previous studies imaged
only a very limited number of time points and assumed the clearance or
wash-in curves to be mono-exponential.
Accurate sampling of the entire washout curve indicates that more than
10 breaths are needed to clear tracer from regions with low
ventilation. Of even greater importance, is the finding from our
studies that some non-dependent lung regions demonstrate nearly linear
rather than mono-exponential clearance curves as assumed in all
previous investigations of regional pulmonary ventilation using x-ray
CT and stable xenon gas. Measurements based on limited time points
modeled as a mono-exponetial may not, therefore, yield physiologically
accurate information.
At high flow rates, inertial factors and hence pathway geometry become
more important in the distribution of air flow (ventilation). Our data
indicate that the dorsal, diaphragmatic region of the pig lung
(dependent lung region in a supine pig) receives an enhanced air flow
distribution at higher flow rates, perhaps do to favorable airway
structure. A pig airways cast, shown in figure
11,
illustrates that due to the monopodial branching pattern of the pig
airways, there is a major airway traversing to the dorsal,
diaphragmatic region of the pig lung. It is our hypothesis that this
airway branching pattern when coupled with high flow rates would be
expected to produce greater tunrover rates in the dependent lung
region of a supine pig. Functional x-ray CT imaging may, therfore,
allow us to explore pulmonary structure-function relationships in
vivo.
Our preliminary results indicate that, with careful attention to cardiac and
ventilatory gating during image acquisition, x-ray CT based imaging of
stable xenon gas can yield detailed maps of regional lung
function. Single breath methods are possible if scanning times are
rapid. Additional work is needed to evaluate the prevelance of near
linear washout regions before we are able to recommend appropriate
sampling intervals and total sampling duration. Used in conjunction
with thin-slice volumetric imaging contrast injection for perfusion
imaging[10], x-ray CT can provide here-to-fore unavailable
information relating pulmonary structure and function.
We are currently automating the image analysis and integrating
ventilation measurements into our customized image analysis
package[10] to allow rapid analysis of the clearance curves
and construction of a corresponding accurate and quantitative color
coded functional maps. Our future investigations include further
validation of the washout data and developing a model-free method of
calculating regional ventilation based on the measured breath to
breath tracer dilution for a regional sample and is thus free from
possible oversimplfying assumptions.
Acknowledgements: We would like to thank Steve DeJong and Tony
Smith for their help with animal preparation. This study was funded in
part by a grant from Mallinckrodt Medical, Inc. and
NIH-HL-42672. Xenon gas was generously donated by PraxAir.
7 References
- 1
-
Awai, K., K. Ito, S. Katsuta, K. Fujikawa, S. Nakamura, T. Ichiki, K. Yamane,
M. Fujie, K. Hasegawa, O. Nishida, et al.
``Xenon enhanced dynamic computed tomography of the lung; regional
ventilation measurement''.
Rinsho Hoshasen--Japanese Journal of Radiology, 35(7):847--853,
1990.
- 2
-
Ball, W. C., P. B. Stewart, L. G. S. Newsham, and D. V. Bates.
``Regional pulmonary function studied with xenon133''.
J Clin Invest, 41(3):519--631, 1962.
- 3
-
Boyd, D. P. and M. J. Lipton.
``Cardiac computed tomography''.
Proceedings of the IEEE, 71:298--307, 1983.
- 4
-
Cullen, S. C. and E. G. Gross.
``Anesthetic properties of xenon in animals and human beings with
additional observation on krypton''.
Science, 113:580--581, 1951.
- 5
-
Gur, D., B. P. Drayer, H. S. Borovetz, B. P. Griffith, R. L. Hardesty, and
S. K. Wolfson.
``Dynamic computed tomography of the lung: regional ventilation
measurements''.
J Comput Assist Tomogr, 3(6):749--753, 1979.
- 6
-
Gur, D., L. Shabason, H. S. Borovetz, D. L. Herbert, G. J. Reece, W. H.
Kennedy, and C. Serago.
``Regional pulmonary ventilation measurements by xenon enhanced
computed tomography: an update''.
J Comput Assist Tomogr, 5(5):678--683, 1981.
- 7
-
Gur, D., H. Yonas, and W. F. Good.
``Local cerebral blood flow by xenon-enhanced CT: current status,
potential improvements, and future directions''.
Cerebrovasc Brain Metab Rev, 1(1):68--86, 1989.
- 8
-
Herbert, D. L., D. Gur, L. Shabason, W. F. Good, J. E. Rinaldo, J. V. Snyder,
H. S. Borovitz, and M. C. Mancini.
``Mapping of human local pulmonary ventilation by xenon enhanced
computed tomography''.
J Comput Assist Tomogr, 6(6):1088--1093, 1982.
- 9
-
Hoffman, E. A.
``Effect of body orientation on regional lung expansion: a computed
tomographic approach''.
J Appl Physiol, 59:468--480, 1985.
- 10
-
Hoffman, E. A., J. K. Tajik, and S. D. Kugelmass.
``Matching pulmonary structure and perfusion via combined dynamic
multislice CT and thin-slice high-resolution CT''.
Computerized Medical Imaging and Graphics, 19(1):101--112,
1995.
- 11
-
Holden, J. E., S. S. Winkler, D. C. Flemming, W. R. Ip, and J. F. Sackett.
``Imaging methods in the transmission computed tomographic
measurement of regional xenon kinetics in the brain''.
Radiology, 135:501--505, 1980.
- 12
-
Kety, S. S.
``The theory and applications of the exchange of inert gas at the
lungs and tissues''.
Pharmacol Rev, 3:1--42, 1955.
- 13
-
Kishore, P. R., G. U. Rao, R. E. Fernandez, R. L. Keenan, G. D. Arora,
P. Gadisseux, L. M. Stewart, A. O. Wist, P. P. Fatouros, D. Dillard, et al.
``Regional cerebral blood flow measurements using stable xenon
enhanced computed tomography: a theoretical and experimental evaluation''.
J Comput Assist Tomogr, 8(4):619--630, 1984.
- 14
-
Murphy, D. M., J. T. Nicewicz, S. M. Zabbatino, and R. A. Moore.
``Local pulmonary ventilation using nonradioactive xenon-enhanced
ultrafast computed tomography''.
Chest, 96(4):799--804, 1989.
- 15
-
Snyder, J. V., B. Pennock, D. Herbert, J. E. Rinaldo, J. Culpepper, W. F. Good,
and D. Gur.
``Local lung ventilation in critically ill patients using
nonradioactive xenon-enhanced transmission computed tomography''.
Crit Care Med, 12(1):46--51, 1984.
- 16
-
Tomiyama, N., N. Takeuchi, H. Imanaka, N. Matsuura, S. Morimoto, J. Ikezoe,
T. Johkoh, J. Arisawa, and T. Kozuka.
``Mechanism of gravity-dependent atelectasis. analysis by
nonradioactive xenon-enhanced dynamic computed tomography''.
Invest Radiol, 28(7):633--638, 1993.
- 17
-
Winkler, S. S., J. E. Holden, J. F. Sackett, D. C. Flemming, and S. C.
Alexander.
``Xenon and krypton as radiographic inhalation contrast media with
computerized tomography: Preliminary note''.
Invest Radiol, 12(1):19--20, 1977.
- 18
-
Winkler, S. S., J. F. Sackett, J. E. Holden, D. C. Flemming, S. C. Alexander,
M. Madsen, and R. I. Kimmel.
``Xenon inhalation as an adjunct to computerized tomography of the
brain: Preliminary study''.
Invest Radiol, 12(1):15--18, 1977.
- 19
-
Wist, P. P. F. A. O., P. R. Kishore, D. S. DeWitt, J. A. Hall, R. L. Keenan,
L. M. Stewart, A. Marmarou, S. C. Choi, and H. A. Kontos.
``Xenon/computed tomography cerebral blood flow measurements. methods
and accuracy''.
Invest Radiol, 22(9):705--712, 1987.
- 20
-
Yonas, H., W. F. Good, D. Gur, S. K. W. Jr, R. E. Latchaw, B. C. Good,
R. Leanza, and S. L. Miller.
``Mapping cerebral blood flow by xenon-enhanced computed tomography:
clinical experience''.
Radiology, 152(2):435--442, 1984.
©1994-2000 Division of Physiologic Imaging, Dept.
of Radiology, Univ. of Iowa
Papers |
DPI Homepage |
VIDA |
NLM |
Contact Us |
Search
Last modified: Fri May 26 16:10:38 CDT
