DOI:10.2214/AJR.06.1153
AJR 2007; 188:1683-1690
© American Roentgen Ray Society
Characterization of the Relation Between CT Technical Parameters and Accuracy of Quantification of Lung Attenuation on Quantitative Chest CT
Brian M. Trotta1,
Alexander V. Stolin1,
Mark B. Williams1,
Spencer B. Gay1,
Alan S. Brody2 and
Talissa A. Altes3
1 Department of Radiology, University of Virginia Medical Center,
Charlottesville, VA.
2 Department of Radiology, Cincinnati Children's Hospital Medical Center,
Cincinnati, OH.
3 Department of Radiology, Children's Hospital of Philadelphia, 34th St. and
Civic Center Blvd., Philadelphia, PA 19104.
Received August 28, 2006;
accepted after revision January 15, 2007.
Address correspondence to T. A. Altes.
Supported by the Rare Lung Disease Consortium and National Institutes of
Health grant NIH RR019498.
Abstract
OBJECTIVE. The purpose of this study was to assess the compromise
between CT technical parameters and the accuracy of CT quantification of lung
attenuation.
MATERIALS AND METHODS. Materials that simulate water (0 H), healthy
lung (–650 H), borderline emphysematous lung (–820 H), and
severely emphysematous lung (–1,000 H) were placed at both the base and
the apex of the lung of an anthropomorphic phantom and outside the phantom.
Transaxial CT images through the samples were obtained while the effective
tube current was varied from 440 to 10 mAs, kilovoltage from 140 to 80 kVp,
and slice thickness from 0.625 to 10 mm. Mean ± SD attenuation within
the samples and the standard quantitative chest CT measurements, the
percentage of pixels with attenuation less than –910 H and 15th
percentile of attenuation, were computed.
RESULTS. Outside the phantom, variations in CT parameters produced
less than 2.0% error in all measurements. Within the anthropomorphic phantom
at 30 mAs, error in measurements was much larger, ranging from zero to 200%.
Below approximately 80 mAs, mean attenuation became increasingly biased. The
effects were most pronounced at the apex of the lungs. Mean attenuation of the
borderline emphysematous sample of apex decreased 55 H as the tube current was
decreased from 300 to 30 mAs. Both the 15th percentile of attenuation and
percentage of pixels with less than –910 H attenuation were more
sensitive to variations in effective tube current than was mean attenuation.
For example, the –820 H sample should have 0% of pixels less than
–910 H, which was true at 400 mA. At 30 mA in the lung apex, however,
the measurement was highly inaccurate, 51% of pixels being below this value.
Decreased kilovoltage and slice thickness had analogous, but lesser,
effects.
CONCLUSION. The accuracy of quantitative chest CT is determined by
the CT acquisition parameters. There can be significant decreases in accuracy
at less than 80 mAs for thin slices in an anthropomorphic phantom, the most
pronounced effects occurring in the lung apex.
Keywords: chest CT • emphysema • lung disease • lung volume reduction surgery • radiation
Introduction
The severity and progression of emphysema are assessed on the basis of
clinical symptoms and spirometric findings, which are measures of global
airflow obstruction. Both of these parameters are relatively insensitive to
small changes in the amount of emphysematous tissue in the lung; in the
absence of treatments that alter the course of the disease, they have been
adequate for diagnosis and monitoring of emphysema progression. A treatment
has become available, however, that may modify the course of emphysema related
to 
1-antitrypsin deficiency
[1–5].
For smoking-related emphysema, lung volume reduction surgery (LVRS) improves
lung function in some but not all patients
[6–8].
Because of the morbidity and mortality of the procedure, it would be desirable
to select only patients likely to respond well to LVRS. Results of preliminary
studies suggest that quantitative CT may have a role in patient selection for
LVRS
[9–13].
A variety of pharmacologic and endobronchial treatments of patients with
smoking-related emphysema also are under development
[1,
5,
14–19].
With the advent of these new treatments, it is important to have a sensitive
and accurate test for assessing the degree of pulmonary emphysema to promote
early detection of disease, monitoring of response to treatment, and initial
drug validation. Quantitative chest CT has been proposed as a sensitive test
for quantifying emphysematous change within the lung
[20–25].
It is estimated that between 15% and 25% of smokers will develop
symptomatic emphysema [14,
26–28].
It is plausible that early detection and early treatment with drugs being
developed to modify the course of disease may provide the best outcome among
these smokers. It therefore is important to accurately quantify lung
attenuation on chest CT. Although accuracy and repeatability of the CT
measurements used to quantify emphysema are essential, the radiation dose must
be minimized to prevent excessive radiation exposure of patients, particularly
if multiple CT scans are needed to track disease progression and response to
treatment. Data comparing the radiation exposure of persons undergoing annual
low-dose CT examinations with that of atomic bomb survivors suggest a small
but statistically significant increase in the number of cases of cancer
resulting directly from the radiation dose of annual low-dose chest CT scans
[29]. To minimize risk to
patients, the radiation dose from quantitative chest CT should be decreased to
the minimum that produces accurate results. Use of lower doses, however, is
associated with noise, artifacts, and unreliable assessment of tissue contrast
enhancement [30]. Although it
is desirable to minimize the radiation exposure of patients, the dose must be
high enough for discrimination between two relatively low-contrast
tissues—normal and emphysematous lung.
The purpose of this study was to quantify the compromise between the
accuracy of lung attenuation measurements and the CT technical parameters used
to determine radiation dose. Because of concerns about radiation exposure of
human subjects, an anthropomorphic phantom was used for repeated measurements
at varying radiation doses. The goal was to determine the minimum CT values
for accurate quantification of emphysematous tissue in the lung.
Materials and Methods
Samples simulating the CT attenuation characteristics of fluid, healthy
lung, borderline emphysematous lung, and severely emphysematous lung were
obtained. The fluid and severely emphysematous lung samples were tap water and
air, respectively, inside thin-walled 40-mm-diameter plastic spheres (Ping
Pong Ball, Wilson Sporting Goods). The normal and borderline emphysematous
lung samples were 3 x 3 x 2 cm blocks of bubbled urethane with a
mean attenuation of approximately –650 and –820 H, respectively
(Fig. 1). Attenuation of
–650 H is approximately the upper limit for healthy lung on an
expiratory CT scan, and –820 H is approximately the upper limit of early
emphysematous change. The lower upper values were selected to minimize the
measured inaccuracies in 15th percentile of attenuation and percentage of
pixels with attenuation less than –910 H. Had samples with the mean
values been used, the errors in CT measurements would have been greater. We
therefore biased the study toward finding the minimum error.
Transaxial CT images through the centers of the samples were obtained with
the samples in three environments: approximately 5 cm above the CT table on a
very thin plastic holder, in the base of an anthropomorphic thoracic phantom,
and in the apex of the phantom (Fig.
2A,
2B). The samples in the apex of
the phantom were arranged in a common transaxial plane with two samples in
each of the hemithoraces. When the samples were placed in the base of the
phantom, the samples also were arranged in a common transaxial plane with two
samples in each hemithorax. Outside of the phantom, the four samples were
aligned in a common transaxial plane.
A 1-second gantry rotation speed was used so that the tube current was
numerically equal to effective tube current in the results presented.
Transaxial slices in each of the three environments were obtained at varying
effective tube currents, kilovoltages, and slice thicknesses on the 16-MDCT
scanner (LightSpeed, GE Healthcare). Images were acquired in a step-and-shoot
mode with a constant field of view of 36 x 36 cm. The physical aperture
size was 1.25 mm for all detectors, referenced to the isocenter. Slices
measuring 0.625 mm are obtained by collimation of the beam down to the inner
halves of the two central detector rows. For slice thicknesses greater than
1.25 mm, data from adjacent detector rows were summed.
Tube current modulation was disabled for this investigation. The standard
(soft-tissue) reconstruction kernel was used for all images. For measurement
of the effects of tube current alone, the effective tube current was varied
from 440 to 10 mAs at a constant 120 kVp and slice thickness of 0.625 mm. Tube
voltage was varied from 140 to 80 kVp with a constant effective tube current
of 200 mAs and slice thickness of 0.625 mm. The effects of slice thickness
were studied between 0.625 and 10 mm at a constant 120 kVp and both 200 and 60
mAs. The CT scanner automatically switched from a focal spot size of 1.2 to
0.7 mm when the effective tube current was changed from 300 and 200 mAs. The
focal spot size remained constant at 0.7 mm for all effective tube current
values less than 200 mAs.
A rectangular region of interest (ROI) with an area of approximately 2
cm2 was drawn manually within the center of each of the sample
materials for each image obtained at each of the various technical parameters.
The mean ± SD of the CT attenuation of the pixels within the ROI was
calculated. Most quantitative chest CT analysis programs perform
histogram-based analysis of the lung attenuation values and produce metrics
such as the 15th percentile of attenuation and percentage of the lung volume
with attenuation below a threshold, typically approximately –910 H
[25]. Thus a program (IDL,
Research Systems) was developed to perform similar analyses within the ROIs,
and the 15th percentile of attenuation and percentage of the pixels with
attenuation less than –910 H were calculated within each ROI for each
sample and each set of technical parameters. The percentage error (percentage
difference) for each CT measurement in each sample and for each set of CT
parameters was calculated as the absolute value of difference between the
reference value and the measured value divided by the mean of the correct and
measured values. The reference value for each metric was that obtained outside
of the phantom at the highest effective tube current (440 mAs) because this
current consistently had the lowest SD of mean attenuation and thus the lowest
noise level.
View larger version (51K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with high-dose (200 mA) technique outside phantom (A), at lung
base (B), and at lung apex (C).
|
|
View larger version (71K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with high-dose (200 mA) technique outside phantom (A), at lung
base (B), and at lung apex (C).
|
|
View larger version (87K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with high-dose (200 mA) technique outside phantom (A), at lung
base (B), and at lung apex (C).
|
|
View larger version (54K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3D —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with low-dose (30 mA) technique outside phantom (D), at lung
base (E), and at lung apex (F). Degraded image quality is
apparent in F.
|
|
View larger version (107K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3E —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with low-dose (30 mA) technique outside phantom (D), at lung
base (E), and at lung apex (F). Degraded image quality is
apparent in F.
|
|
View larger version (122K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3F —Sample materials in three environments tested in phantom.
Slice thickness is 0.625 mm and tube voltage, 120 kVp. Axial CT images
acquired with low-dose (30 mA) technique outside phantom (D), at lung
base (E), and at lung apex (F). Degraded image quality is
apparent in F.
|
|
Results
Typical high- and low-dose CT images are shown in Figure
3A,
3B,
3C,
3D,
3E,
3F. Outside the phantom there
was no apparent difference between high-dose and low-dose scans. At the base
of the lungs, additional noise was evident on the 30-mA scan. At the apex of
the lungs, there was significant degradation in image quality, streak
artifacts, and increased noise being apparent on the low-dose image.
The variation in mean attenuation in relation to tube current is shown in
Figure 4A,
4B,
4C,
4D for each of the three
environments. Outside the phantom, mean attenuation did not change appreciably
as the effective tube current was decreased for the range of values evaluated.
At 30 mAs, the percentage error ranged from zero to 1.5% for the three lung
samples. At the lung base in the phantom, the effects of a lower tube current
were evident, a slight decrease in mean attenuation occurring at lower
effective tube current. The mean attenuation in the normal lung sample
decreased from –651 to –662 H as the tube current was decreased
from 200 to 30 mAs, giving a percentage error of 2.5% at 30 mAs. At the apex
of the lung, the mean attenuation was strongly dependent on tube current at
currents less than approximately 60 mAs. The mean attenuation of the
borderline emphysematous lung sample decreased from –815 to –860 H
as the effective tube current was decreased from 200 to 30 mAs, moving the
mean attenuation of the borderline emphysematous lung sample into the
emphysematous range. This change represented a percentage error of 4.6% at 30
mAs.
The change, or bias, in mean attenuation was greatest in the apex of the
lung for all samples. In the apex, the bias was also greatest for the water
sample, a change of –99 H occurring as the effective tube current was
decreased from 440 to 30 mAs. The bias was the least for the severely
emphysematous lung sample (air), a change of 29 H occurring as the effective
tube current was decreased from 440 to 30 mAs. In some of the samples, bias
occurred in the opposite direction: The mean attenuation of water, normal
lung, and borderline emphysematous lung decreased as effective tube current
was decreased, and that of the severely emphysematous lung increased with a
decrease in effective tube current (see Discussion).
Outside of the phantom at high effective tube current, the SD of the
measured attenuation of the pixels within the samples was very small. This
finding indicated that the materials were very homogeneous in CT attenuation
and that quantum noise was low (Fig.
4A,
4B,
4C,
4D). Inside the phantom,
however, the SD of attenuation within the samples was even more dependent on
tube current than was mean attenuation (Fig.
4A,
4B,
4C,
4D). Again, this dependence
was greatest at the lung apex. For example, outside the phantom, the SD of the
borderline emphysematous sample increased only 11 H as effective tube current
was decreased from 440 to 10 mAs, most of the increase in SD occurring at
effective tube currents less than 80 mAs. At 30 mAs outside the phantom, the
maximum percentage error was 1.5% for the three lung samples. In the lung
apex, the SD of the borderline emphysematous sample increased 99 H as
effective tube current was decreased from 440 to 10 mAs, and the percentage
error ranged from 4.2% to 12.8% for the three lung samples at 30 mAs. The
increase in SD versus decreasing effective tube current appeared to be highly
nonlinear at all three locations, an approximately exponential increase in SD
occurring at low effective tube current
(Fig. 4D).
The more commonly used metrics in quantitative CT, 15th percentile of
attenuation and percentage of pixels with attenuation less than –910 H,
proved more sensitive to variations in tube current than was mean attenuation.
A larger error, or bias, was found with 15th percentile of attenuation than
with mean attenuation for all samples and environments (Fig.
5A,
5B,
5C). This difference was most
pronounced in the lung apex, where the 15th percentile of attenuation of the
normal lung sample decreased from –687 to –957 H as the effective
tube current was decreased from 200 to 30 mAs, changing the appearance from
normal lung to emphysematous lung. In the lung apex, the 15th percentile of
attenuation of the borderline emphysematous lung sample decreased from
–848 to –999 H as the effective tube current was decreased from
200 to 30 mAs, changing the appearance from borderline emphysema to severe
emphysema. The 15th percentile of attenuation of the severely emphysematous
sample remained –1,000 H for all technical parameters in all three
environments. At 30 mAs, the maximum percentage error for the three lung
samples was 1.5% outside the phantom, 12.9% at the lung base, and 38.5% in the
lung apex.
For percentage of pixels with attenuation less than –910 H, there was
little change with changes in effective tube current. The samples outside the
phantom for all three lung samples had 0% error at 30 mAs (Fig.
6A,
6B,
6C). The measurement was,
however, quite sensitive to variations in tube current when the samples were
in the lung base or apex and became highly inaccurate at lower tube currents.
For normal lung (–650 H) and borderline emphysematous lung (–820
H), percentage of pixels with attenuation less than –910 H should be
zero, as was found in samples outside the phantom. However, a large number of
pixels had less than –910-H attenuation within these samples at lower
tube currents. In a comparison of the three environments for the borderline
emphysematous sample, the percentage of pixels within the ROI with values less
than –910 H was 1% at 80 mA, 12% at 30 mA, and 39% at 10 mA at the lung
base and 9% at 80 mA, 51% at 30 mA, and 85% at 10 mA at the lung apex.
The sensitivity of the mean ± SD of attenuation to variations in
tube voltage and slice thickness was assessed. The mean ± SD of
attenuation was found to vary slightly less with changes in tube voltage and
slice thickness than with changes in tube current for the range of parameters
evaluated. As tube voltage was decreased from 140 to 80 kVp, the mean
attenuation for the borderline emphysematous sample decreased, and the SD of
mean attenuation increased. The 15th percentile of attenuation also decreased
as tube voltage was decreased, and the percentage of pixels with attenuation
less than –910 H increased (Table
1).
View this table:
[in this window]
[in a new window]
|
TABLE 1: Changes in Value of Four Measurements at Three Locations for Borderline
Emphysematous Lung Sample with Decrease in Tube Kilovoltage from 140 to 80
kVp
|
|
The effects of slice thickness on mean attenuation were measured at two
effective tube currents, 200 and 60 mAs. At the higher current, mean
attenuation increased as the slice thickness was increased from 0.625 to 10 mm
(Table 2). At the lower
effective tube current of 60 mAs, mean attenuation increased even more as
slice thickness was increased from 0.625 to 10 mm
(Table 3). There was an
approximately linear relation between slice thickness and mean attenuation.
Linear regression with the equation mean attenuation = (15.5 x slice
thickness) – 853 resulted in r = 0.91 for the borderline
emphysematous sample. As expected, the SD decreased as slice thickness was
increased from 0.625 to 10 mm, the lower effective tube current of 60 mAs
being more sensitive to changes in slice thickness (Tables
2 and
3). The 15th percentile of
attenuation also showed an approximately linear relation to slice thickness.
Accuracy improved as slice thickness was increased. Linear regression with the
equation 15th percentile = (10.7 x slice thickness) – 872 resulted
in r = 0.86 for the borderline emphysematous sample.
View this table:
[in this window]
[in a new window]
|
TABLE 2: Changes in Value of Four Measurements at Three Locations in Borderline
Emphysematous Lung Sample with Increase in Slice Thickness from 0.625 to 10
mm
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 3: Changes in Value of Four Measurements at Three Locations in Borderline
Emphysematous Lung Sample with Increase in Slice Thickness from 0.625 to 10
mm
|
|
Focal spot size automatically changed from 1.2 to 0.7 mm when the effective
tube current was changed from 300 and 200 mAs. There were no discontinuities
in any of the results as effective tube current was decreased from 300 to 200
mAs, so variation in focal spot size did not appear to influence the
results.
Discussion
We found that the accuracy of quantification of lung attenuation on CT of
an anthropomorphic thoracic phantom varies with slice position. Noise and
artifacts increase at the lung apex, where a number of high-attenuation
structures (clavicles, scapula, upper ribs) surround the lung. This dependence
on slice position is not surprising given the known increase in noise and
artifacts in areas adjacent to high-attenuation structures on CT. The new
result is that we quantified how this phenomenon affects the accuracy of
quantitative chest CT and found that bias is introduced into mean attenuation
as tube current is decreased to the values commonly used for low-dose chest
CT. As expected, the SD of the measured attenuation increased with decreases
in tube current. The increase was not linear, however, as predicted in theory
but was approximately exponential, increasing at low effective tube current,
likely because of the presence of reconstruction artifacts. Furthermore, the
commonly used quantitative CT measurements were more sensitive to decreases in
tube current than was mean attenuation.
Because smoking-related lung disease frequently is apically predominant,
our findings have implications for the use of quantitative chest CT in the
detection and quantification of smoking-related emphysema. At a low dose,
errors in the apex tend to result in overestimation of the degree of
emphysematous change, which biases whole lung–based metrics. A patient
with apically predominate disease will appear to have more severe emphysema
than a patient with an equal amount of emphysematous tissue with uniform or
basally predominate disease. Furthermore, the errors will likely depend on the
exact position of the shoulders and ribs relative to the lung parenchyma.
Differences in the position of the patient's arms and shoulders during CT may
result in differences in the measured amount of emphysematous tissue,
particularly in the apices.
Inaccuracy of measured CT attenuation values in the lung apex is thought to
be due to the surrounding bone structures, which decrease the number of
photons detected with attenuation of the X-ray beam, reducing signal intensity
and signal-to-noise ratio. The image thus has the appearance of increased
noise. These structures also create artifacts such as streak and beam
hardening, which are accentuated at low tube currents. Beam-hardening artifact
was thought to cause bias in the mean attenuation we found at lower effective
tube current because the bias was toward lower mean values for all samples
except severely emphysematous lung (pure air) sample. The positive bias for
the pure air sample occurred because the minimum attenuation reported by the
CT scanner was set to –1,000 H, and noise in a material with an
attenuation of –1,000 H can only increase the measured attenuation
values. This phenomenon has implications for differentiating intrinsically
low-contrast-enhancement and low-attenuation materials such as emphysematous
and normal lung. The bias for severely emphysematous lung is positive, whereas
the bias for normal or borderline emphysematous lung is negative. This
phenomenon causes the measured values to approach each other, further
decreasing the contrast between the tissues at low effective tube current and
potentially rendering them indistinguishable. This phenomenon also likely
causes overestimation of the amount of emphysema in patients with mild
disease, particularly in the lung apices.
In the absence of surrounding structures, the SD of CT attenuation in
homogeneous media should increase linearly as effective tube current is
decreased, which was our finding for sample materials outside the phantom.
However, we found an approximately exponential increase in SD in the lung
apex, likely the result of the increasing severity of reconstruction
artifacts. Furthermore, the commonly used quantitative CT measurement of 15th
percentile of attenuation and percentage of pixels with attenuation less than
–910 H were even more sensitive to reductions in tube current, becoming
highly inaccurate at tube currents less than approximately 40 mA.
Studies have been conducted to assess the repeatability of low-dose
quantitative chest CT. Stolk et al.
[31] determined the
repeatability of lung density measurements on low-dose CT in the
quantification of emphysema in human subjects. Using CT settings of 140 kVp,
20 mAs, and a slice thickness of 2.5 mm, Stolk et al. found excellent
repeatability of results of quantitative lung density analysis in 10 subjects
with emphysema. Gierada [32]
reached the same conclusion in a study of the repeatability of CT indexes in
patients evaluated for LVRS. Although these studies of low-dose quantitative
CT showed excellent reproducibility, none was designed to determine the
accuracy of low-dose CT in the quantification of emphysema. In a more recent
study, Shaker et al. [33]
examined the reproducibility of CT measurement at tube currents of 8, 16, and
32 mA and with various reconstruction algorithms in patients with emphysema
due to smoking and 1-antitrypsin deficiency. The results
indicated good reproducibility regardless of the type of emphysema, radiation
dose, or reconstruction algorithm. In accordance with our results, Shaker et
al. concluded that the use of very low radiation doses results in
overestimates of the amount of emphysematous tissue.
Reproducibility–repeatability and accuracy are two different, but
essential, attributes of a good biomarker. That whole-lung quantitative CT
measurements derived from low-dose CT scans are repeatable is not surprising
given the law of large numbers and the relatively small fraction of the lung
parenchyma in the lung apex. Our results, however, suggest that the accuracy
of low-dose CT metrics is poor, especially on a regional basis.
In a recent study, Zaporozhan et al.
[34] added simulated noise to
the CT scans of 30 patients with severe emphysema to assess the influence of
noise on the accuracy of CT emphysema metrics. Those authors found degraded
accuracy at simulated effective tube currents less than 50 mAs. We found that
the SD of measured attenuation did not increase linearly with decreasing tube
current, as would be expected if noise from reduced tube current were the only
factor degrading image quality; instead, the SD increased approximately
exponentially. Thus the results of this simulation may represent a best-case
scenario. Because our study was conducted with materials of known attenuation,
we were able to directly measure the compromise between accuracy and dose.
Although it is repeatable, CT quantification at tube current less than 80 mA
becomes progressively more inaccurate at lower tube currents. This effect has
important implications for the use of quantitative CT as a biomarker for
emphysema.
Quantitative chest CT has been used as a biomarker for emphysema and is the
only outcome measure that has shown a positive effect in patients with
1-antitrypsin deficiency treated with
1-antitrypsin replacement therapy. That study
[15] was conducted with a
high-dose CT technique. Concerns regarding radiation risk, however, make the
use of low-dose technique at least theoretically attractive. We found that the
accuracy of quantitative CT was greatly reduced at the tube currents commonly
used with low-dose techniques but that this inaccuracy can be largely
ameliorated with use of thick slices. However, advanced quantitative CT
techniques used to automatically assess the airways require thin slices
[35,
36]. A possible compromise is
to acquire thin slices at a low dose and to reconstruct the scan with thicker
slices. The thin slices would be used for the more advanced quantitative CT
techniques, and the thicker slices would be used for the more common
histogram-based metrics.
Because the repeated CT required for this study could not be performed on
human subjects owing to the high radiation dose, an anthropomorphic phantom
was used. The phantom had a very slim body habitus, and inaccuracies in
measured attenuation would likely be greater in an average human patient. Thus
a single set of CT parameters is not adequate for low-dose quantitative chest
CT. To compensate for the increased X-ray attenuation of thicker body tissues,
CT of larger patients would have to be performed at higher effective tube
current for the degree of accuracy achieved with the phantom. Automatic tube
current modulation, whereby the tube current is automatically adjusted to
variations in the amount of X-ray attenuation in different portions of the
body, is available on newer CT scanners. The effect of automatic tube current
modulation on quantitative chest CT appears to be an interesting area for
future research.
In summary, thin-slice low-dose quantitative CT of the chest is highly
inaccurate. This lack of accuracy will likely cause overestimation of the
severity of emphysema, particularly in patients with mild disease. It may be
possible, however, to ameliorate the lack of accuracy by reconstructing
thicker slices for use in analysis of quantitative chest CT scans.
Acknowledgments
We acknowledge Bruce Trapnell for supporting this work. We also thank Kyoto
Kagaku for providing the synthetic lung samples.
References
- Abusriwil H, Stockley RA. Alpha-1-antitrypsin replacement therapy:
current status. Curr Opin Pulm Med 2006;12
: 125–131[Medline]
- Barker AF, Siemsen F, Pasley D, D'Silva R, Buist AS. Replacement
therapy for hereditary alpha1-antitrypsin deficiency: a program for long-term
administration. Chest 1994;105
:1406
–1410[CrossRef][Medline]
- Gadek JE, Klein HG, Holland PV, Crystal RG. Replacement therapy of
alpha 1-antitrypsin deficiency: reversal of protease-antiprotease imbalance
within the alveolar structures of PiZ subjects. J Clin
Invest 1981; 68:1158
–1165[Medline]
- Wewers MD, Casolaro MA, Sellers SE, et al. Replacement therapy for
alpha 1-antitrypsin deficiency associated with emphysema. N Engl J
Med 1987; 316:1055
–1062[Abstract]
- Sandhaus RA. 1-Antitrypsin deficiency 6: new and
emerging treatments for 1-antitrypsin deficiency.
Thorax 2004; 59:904
–909[Abstract/Free Full Text]
- Naunheim KS, Wood DE, Mohsenifar Z, et al. Long-term follow-up of
patients receiving lung-volume-reduction surgery versus medical therapy for
severe emphysema by the National Emphysema Treatment Trial Research Group.
Ann Thorac Surg 2006;82
: 431–443[Abstract/Free Full Text]
- Lim E, Ali A, Cartwright N, et al. Effect and duration of lung
volume reduction surgery: mid-term results of the Brompton trial.
Thorac Cardiovasc Surg 2006;54
: 188–192[CrossRef][Medline]
- Miller JD, Malthaner RA, Goldsmith CH, et al. A randomized clinical
trial of lung volume reduction surgery versus best medical care for patients
with advanced emphysema: a two-year study from Canada. Ann Thorac
Surg 2006; 81:314
–320[Abstract/Free Full Text]
- Screaton NJ, Reynolds JH. Lung volume reduction surgery for
emphysema: what the radiologist needs to know. Clin
Radiol 2006; 61:237
–249[CrossRef][Medline]
- Slone RM, Pilgram TK, Gierada DS, et al. Lung volume reduction
surgery: comparison of preoperative radiologic features and clinical outcome.
Radiology 1997;204
: 685–693[Abstract/Free Full Text]
- Maki DD, Miller WT Jr, Aronchick JM, et al. Advanced emphysema:
preoperative chest radiographic findings as predictors of outcome following
lung volume reduction surgery. Radiology1999; 212:49
–55[Abstract/Free Full Text]
- Cederlund K, Bergstrand L, Hogberg S, et al. Visual classification
of emphysema heterogeneity compared with objective measurements: HRCT vs
spiral CT in candidates for lung volume reduction surgery. Eur
Radiol 2002; 12:1045
–1051[CrossRef][Medline]
- Coxson HO, Whittall KP, Nakano Y, et al. Selection of patients for
lung volume reduction surgery using a power law analysis of the computed
tomographic scan. Thorax 2003;58
: 510–514[Abstract/Free Full Text]
- Snell GI, Holsworth L, Borrill ZL, et al. The potential for
bronchoscopic lung volume reduction using bronchial prostheses: a pilot study.
Chest 2003; 124:1073
–1080[CrossRef][Medline]
- Dirksen A, Dijkman JH, Madsen F, et al. A randomized clinical trial
of alpha(1)-antitrypsin augmentation therapy. Am J Respir Crit Care
Med 1999; 160:1468
–1472[Abstract/Free Full Text]
- Juvelekian GS, Stoller JK. Augmentation therapy for
alpha(1)-antitrypsin deficiency. Drugs2004; 64:1743
–1756[CrossRef][Medline]
- Wan IY, Toma TP, Geddes DM, et al. Bronchoscopic lung volume
reduction for end-stage emphysema: report on the first 98 patients.
Chest 2006; 129:518
–526[CrossRef][Medline]
- Toma TP, Hopkinson NS, Hillier J, et al. Bronchoscopic volume
reduction with valve implants in patients with severe emphysema.
Lancet 2003; 361:931
–933[CrossRef][Medline]
- Yim AP, Hwong TM, Lee TW, et al. Early results of endoscopic lung
volume reduction for emphysema. J Thorac Cardiovasc
Surg 2004; 127:1564
–1573[Abstract/Free Full Text]
- Gevenois PA, Yernault JC. Can computed tomography quantify
pulmonary emphysema? Eur Respir J 1995;8
: 843–848[Abstract]
- Stern EJ, Song JK, Frank MS. CT of the lungs in patients with
pulmonary emphysema. Semin Ultrasound CT MR1995; 16:345
–352[CrossRef][Medline]
- Dirksen A, Friis M, Olesen KP, Skovgaard LT, Sorensen K. Progress
of emphysema in severe alpha 1-antitrypsin deficiency as assessed by annual
CT. Acta Radiol 1997;38
: 826–832[Medline]
- Coxson HO, Rogers RM, Whittall KP, et al. A quantification of the
lung surface area in emphysema using computed tomography. Am J
Respir Crit Care Med 1999;159
: 851–856[Abstract/Free Full Text]
- Bankier AA, Madani A, Gevenois PA. CT quantification of pulmonary
emphysema: assessment of lung structure and function. Crit Rev
Comput Tomogr 2002; 43:399
–417[Medline]
- Goldin JG. Quantitative CT of emphysema and the airways.
J Thorac Imaging 2004;19
: 235–240[CrossRef][Medline]
- Centers for Disease Control and Prevention. Smoking-attributable
mortality and years of potential life lost: United States, 1984.
MMWR Morb Mortal Wkly Rep 1997;46
: 444–451[Medline]
- Centers for Disease Control and Prevention. Annual
smoking-attributable mortality, years of potential life lost, and economic
costs: United States, 1995–1999. MMWR Morb Mortal Wkly
Rep 2002; 51:300
–303[Medline]
- Cummings KM, Stiles J, Mahoney MC, Sciandra R. Health and economic
impact of cigarette smoking in New York State, 1987–1989. N Y
State J Med 1992; 92:469
–473[Medline]
- Brenner DJ, Elliston CD. Estimated radiation risks potentially
associated with full-body CT screening. Radiology2004; 232:735
–738[Abstract/Free Full Text]
- Sprawls P. AAPM tutorial: CT image detail and noise.
RadioGraphics 1992;12
:1041
–1046[Abstract]
- Stolk J, Dirksen A, van der Lugt AA, et al. Repeatability of lung
density measurements with low-dose computed tomography in subjects with
alpha-1-antitrypsin deficiency-associated emphysema. Invest
Radiol 2001; 36:648
–651[CrossRef][Medline]
- Gierada DS. Radiologic assessment of emphysema for lung volume
reduction surgery. Semin Thorac Cardiovasc Surg2002; 14:381
–390[CrossRef][Medline]
- Shaker SB, Dirksen A, Laursen LC, Skovgaard LT, Holstein-Rathlou
NH. Volume adjustment of lung density by computed tomography scans in patients
with emphysema. Acta Radiol 2004;45
: 417–423[CrossRef][Medline]
- Zaporozhan J, Ley S, Weinheimer O, et al. Multidetector CT of the
chest: influence of dose onto quantitative evaluation of severe emphysema: a
simulation study. J Comput Assist Tomogr2006; 30:460
–468[CrossRef][Medline]
- Aykac D, Hoffman EA, McLennan G, Reinhardt JM. Segmentation and
analysis of the human airway tree from three-dimensional X-ray CT images.
IEEE Trans Med Imaging 2003;22
: 940–950[CrossRef][Medline]
- Fetita CI, Preteux F, Beigelman-Aubry C, Grenier P. Pulmonary
airways: 3-D reconstruction from multislice CT and clinical investigation.
IEEE Trans Med Imaging 2004;23
:1353
–1364[CrossRef][Medline]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
, –650 H;
, –820 H; 
, –1,000 H.
