| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Mallinckrodt Institute of Radiology, Washington University School of Medicine,
510 S Kingshighway Blvd., St. Louis, MO 63110.
2 Present address: University of Pittsburgh Medical Center, Pittsburgh, PA
15213.
3 Present address: Radiology Department, Chungnam National University School of
Medicine, Dae Jon, Korea 301-721.
Received August 25, 2005;
accepted after revision December 7, 2005.
Address correspondence to D. S. Gierada
(gieradad{at}wustl.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. The study population consisted of 36 men and 20 women who were current or former heavy smokers and underwent standard-dose (effective tube current, 100-250 mAs) chest CT at our institution within 6 months of having undergone low-dose (effective tube current, 30-60 mAs) chest CT. All CT scans were reconstructed at 5-mm slice thickness with a smooth filter. CT-measured lung volume, mean and median lung attenuation, and percentage of lung volume with attenuation lower than multiple thresholds (emphysema index values) were compared by Pearson correlation, two-tailed and paired Student's t tests, and regression analysis.
RESULTS. There were no significant differences in mean attenuation
(-848 vs -846 H, p > 0.35) for the low dose and the standard dose
or in median lung attenuation (-879 vs -878 H, p > 0.66). Low- and
standard-dose emphysema indexes were correlated at all attenuation thresholds
(r = 0.86-0.97). Mean emphysema indexes were higher on the low-dose
scans, but the mean difference at all thresholds was less than 3%. The
differences were significant (p < 0.05) only at the lower index
thresholds, correlated with differences in lung volume (r 
0.86),
and increased with greater differences in dose.
CONCLUSION. Low-dose technique has minimal effect on CT quantification of emphysema.
Keywords: CT technique • lung diseases
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
On CT images, lung tissues of different compositions are mapped onto pixels with different attenuation values. The volume of emphysema is typically calculated as the sum of pixels with attenuation values below a specific threshold value. The studies that established the lung attenuation threshold values used for defining emphysema with quantitative CT (QCT) analysis [1, 2, 13-15] were performed before the introduction of low-dose CT techniques. These studies were conducted with X-ray tube currents up to eight times greater than those commonly used in lung cancer screening studies. Although areas of emphysema on low-dose CT images may be as perceptible as those on standard-dose CT images, reduction of radiation dose lowers the signal-to-noise ratio. The effect of this decrease on the frequency distribution of lung attenuation values in varying degrees of emphysema has not been established. This study was performed to compare QCT measurements of emphysema obtained from standard- and low-dose CT scans.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Using our local National Lung Screening Trial and radiology databases, we identified for inclusion in the study 50 subjects who met the criteria of having undergone a low-dose MDCT screening examination and a standard-dose MDCT examination at our institution within 6 months of the low-dose examination. For 48 of these subjects, the standard-dose CT examination was performed for follow-up of a focal pulmonary lesion 4-18 mm in greatest transverse dimension detected at screening. For the two other subjects, the standard-dose CT examination was performed during evaluation of acute chest pain. The other six subjects included in this study underwent low-dose MDCT for an imaging-pathology correlation study of chronic obstructive pulmonary disease and had undergone standard-dose MDCT within the 6 months before the low-dose CT examination. For one of these patients CT was performed before lobectomy for T1 lung cancer. The other five patients underwent CT before bilateral lung transplantation for end-stage chronic obstructive pulmonary disease.
CT Scans
The local human studies committee approved use of existing CT images for
this study. Written informed consent had been obtained from all subjects
before low-dose CT, although it was not required by the human studies
committee for the retrospective quantitative measurements in this study. The
standard-dose CT scans were performed as part of the patients' clinical
care.
CT was performed with either a 4-MDCT system (Volume Zoom, Siemens Medical Solutions) or a 16-MDCT system (Sensation 16, Siemens Medical Solutions) at full inspiration without IV contrast material. Inspiratory status was not monitored. The scans were obtained at 120 kVp and a pitch of 1.5 (16-MDCT) or 2.0 (4-MDCT). The effective tube current was 30-60 mAs for the low-dose scans and 100-250 mAs for the standard-dose scans. All low-dose and standard-dose scans were reconstructed at 5-mm slice thickness with a smooth reconstruction algorithm.
Image Analysis
QCT analysis was performed with a desktop computer running the Pulmonary
Analysis Software Suite (University of Iowa Division of Physiologic Imaging)
[16,
17]. This program
automatically segments the lung, allowing manual modification if necessary to
correct erroneous inclusion of nonpulmonary structures (such as the trachea
and abdominal gas) and exclusion of lung regions. For this analysis, the
trachea and main bronchi were manually excluded when necessary. Histogram
analysis of the segmented lung is then performed with display of statistical
parameters. Statistical parameters included total segmented lung volume;
percentage tissue ([lung attenuation - air attenuation]/[tissue attenuation -
air attenuation], where air attenuation is -1,000 H and tissue attenuation is
65 H); tissue volume (total segmented lung volume x percentage tissue);
percentage air (1 - percentage tissue); air volume (total segmented lung
volume x percentage air); mean lung attenuation; median lung
attenuation; SD of mean lung attenuation; and full width at half maximum of
the attenuation histogram. Emphysema index was computed as percentage of lung
volume below a user-defined attenuation threshold.
The air calibration of each scan was checked by the average of three circular region-of-interest (ROI) measurements (0.5-1.0 cm2) of the mean attenuation of air in the trachea between the top of the aortic arch and the carina and of air anterior to the subject at the same scan level. Noise levels were assessed as the average SD of the mean attenuation of three circular ROI measurements at the same two locations and of blood in the central portion of the left atrium and in the descending aorta at the same scan level.
Data Analysis
Previous studies validated attenuation threshold values of -950 H
[2,
13] and -910 H
[1] for quantitative
determination of emphysema indexes at slice thicknesses of 1.0 mm and 10 mm,
respectively. The CT scans for this study were obtained at 5-mm slice
thickness, and the raw data were no longer available for reconstruction of the
scans at 1.0-mm or 10-mm slice thickness. An attenuation threshold midway
between those validated for 1.0- and 10-mm slice thicknesses (-930 H) was
considered a reasonable choice for comparing actual amounts of emphysema.
However, multiple additional attenuation thresholds were also evaluated: -960
to -890 H increasing in 10-H increments, -870 H, and -850 H. An emphysema
index was determined for each of these thresholds.
Comparisons between low-dose and standard-dose scan data were made by Pearson correlation, two-tailed paired Student's t tests and linear regression and performed with Excel 2000 software (Microsoft). Variables potentially influencing differences between low-dose and standard-dose scan data were tested by backward stepwise multiple regression analysis performed with JMP software (SAS Institute). Comparisons for which p < 0.05 were considered statistically significant.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There was no significant difference in mean or median lung attenuation, or in the SD, between the low- and standard-dose scans, and correlations for these parameters were high (Table 1). The full width at half maximum of the attenuation histogram was slightly greater on the low-dose scans. Mean emphysema index (percentage of lung volume below a given attenuation threshold) was 1.1-2.6% higher on the low-dose scans than on the standard-dose scans over the range of attenuation thresholds assessed, although this finding was statistically significant only at thresholds of -920 H and lower (Fig. 1). The absolute differences were greatest at the -930 H threshold (mean, 17.8% for low dose vs 15.3% for standard dose; p < 0.005) and the -920 H threshold (mean, 23.8% vs 21.2%; p = 0.03) and decreased at lower and higher thresholds. The number of subjects in whom the low-dose emphysema index was greater than the standard-dose index ranged from 39 to 48 of the 56 subjects for attenuation thresholds in the range of -960 to -900 H. A linear relation between low- and standard-dose emphysema indexes was observed, the Pearson correlation coefficients ranging from 0.87-0.97 at the various attenuation thresholds.
|
|
There were no statistically significant differences in mean air volume, tissue volume, or total volume between the low-dose and the standard-dose scans (Table 1). However, although mean tissue volume was almost identical for the low-dose and standard-dose scans (0.25% difference), mean air volume and total volume were slightly lower in the standard-dose group (2.3% and 2.0% lower, respectively; p = 0.11 for both). Because of this trend, the relation between emphysema index and lung volume during scanning was examined more closely.
Correlations between the differences in lung volume on the low- and standard-dose scans and the differences in emphysema indexes were relatively strong (r = 0.75 for the -930 H threshold). Backward stepwise multiple regression was performed to identify the variables with the strongest influence on the difference in emphysema indexes between low- and standard-dose scans. The independent variables tested included difference in total lung volume, percentage difference in total lung volume, standard-dose to low-dose exposure ratio, standard-dose-low-dose exposure difference, time between low- and standard-dose scans, scanner type, and voxel size. In this analysis, the difference in total lung volume was strongly associated with the difference in emphysema index (p < 0.0001). None of the other variables had a statistically significant relation in the model. The model r2 of 0.59 for the -930 H threshold with all independent variables was only slightly higher than the r2 of 0.56 for the difference in total lung volume alone.
To eliminate the potential influence of differences in lung volume and to isolate the effects of radiation dose, repeat analysis was performed in which comparisons were limited to the subjects in whom the difference in total lung volume between low- and standard-dose scans was 3% or less (Table 2, Fig. 2) and were further subdivided by exposure level (Table 2). In this analysis, lung volume and mean attenuation were almost identical between the low-dose and standard-dose scans. Among all subjects in this group, emphysema indexes at the various thresholds were 0.4-1.6% higher in patients scanned with the low-dose technique. When this group was subdivided into those for whom the standard-dose to low-dose exposure ratio was less than 3.5 and those for whom it was 3.5 or higher, the differences in emphysema indexes were found to be larger in those with the higher exposure ratios (up to 2.3% difference at the -930 H threshold; Table 2).
|
|
The mean of the tracheal air ROI measurements was within 0.2% of the CT-defined value for air of -1,000 H for the low-dose (-1,001 ± 5 H; range, -1,012 to -988 H) and standard-dose (-1,002 ± 8 H; range, -1,016 to -970 H) scans. The mean of the external air ROI measurements anterior to the subjects was within 0.1% of the CT-defined value for air for the low-dose (-1,001 ± 2 H; range, -1,007 to -998 H) and standard-dose (-1,001 ± 2 H; range, -1,010 to -997 H) scans. Mean noise level was 1.6-1.9 times greater on the low-dose scans, depending on the location sampled (Table 3).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
One practical implication of these results is that precise consistency of exposure factors is not necessary when CT scans are used for comparative studies of emphysema. This finding is in contrast to differences related to slice thickness [1, 2, 13, 14] and reconstruction filter [18], which can have substantial effects on emphysema index. As long as the latter two variables are held constant, valid comparisons of cohorts in which exposure factors are not constant should be possible.
Substantial variation in emphysema indexes between low- and standard-dose CT scans was found for some of the individual subjects in this study, as shown in Figure 2. Differences in the lung volume of these subjects during CT accounted for most of this variation. Lung volume, or the degree of inspiration during CT, is well-recognized as a critical factor influencing CT measurements of emphysema [19-22]. In ideal circumstances, a standard method for achieving consistent lung volumes should be used. Because spirometric control of lung volume tends to be cumbersome and is not widely available, most investigators standardize lung volume by coaching subjects to inhale to total lung capacity. In this retrospective study, CT was not performed for emphysema quantification, so there was no prospective attempt to optimize consistency of inspiratory levels. Therefore the effect of lung volume was closely scrutinized.
Although there was no statistically significant difference in CT-measured lung volume for the group of all subjects, a trend toward larger total lung volume was found for the low-dose scans compared with the standard-dose scans. Such a larger inspiratory volume may account for or contribute to an increase in emphysema index. Multivariate analysis showed lung volume was the only variable with a major effect on the differences in emphysema indexes between low- and standard-dose scans. Nevertheless, when we eliminated the effect of lung volume differences by repeating the comparisons using only cases in which the difference in lung volume was 3% or less, the pattern of higher emphysema index on the low-dose scans persisted, although the differences were smaller. This effect of radiation dose on emphysema index was further supported by the finding of a larger difference between mean low- and standard-dose emphysema indexes in subjects with a higher ratio of standard-dose to low-dose exposure.
The results of this study are similar to the results of a more limited study [23] in which 20-80% reductions in CT milliamperage were compared on multiple repeated single-level scans. Using 2-mm slice thickness, the authors found an increase in the emphysema index (-960 H threshold) with a decrease in radiation, of up to 3.6% among subjects with an emphysema index less than 30%. Possible effects of lung volume were not considered. We did not find an obvious trend related to the magnitude of emphysema index (i.e., emphysema severity).
The optimal attenuation threshold for defining emphysema with the slice thickness used for the scans in this study is uncertain. Although thresholds of -910 H for 10-mm-thick, contiguous, contrast-enhanced scans [1] and -950 H for unenhanced 1.0-mm slices obtained every 10 mm [2, 13] have been shown most accurate at reflecting the amount of emphysema present at histologic analysis of lung specimens, the optimal threshold for 5-mm-thick scans has not been defined. Considering the results of the previous studies, however, it is likely that the optimal threshold for defining emphysema with standard dose, 5-mm scans is between -950 and -910 H. This reference range may not be completely comparable, because the standard dose in this study was approximately one half of the dose used in one of the reference studies [2, 13] (the report of the other reference study does not show the exposure factors). In addition, the reports of the reference studies do not specify the reconstruction algorithm used. For these reasons, the effects of a reduced dose on attenuation frequency distribution were assessed at multiple attenuation thresholds. The mean emphysema indexes at the low dose remained slightly but consistently higher at all attenuation thresholds.
There were several limitations to this study. One was that it was retrospective, so the variables were not precisely controlled. A prospective study in which subjects had undergone sequential imaging at low and standard doses on the same scanner with verbal coaching to obtain reproducible inspiratory breath-holds would have been preferable but would have required additional radiation. Because the reconstruction filters for the 4-MDCT and 16-MDCT scanners used in this study are technically equivalent (M. Milite, H. Vestner, Research and Development Group, Siemens Medical Solutions, personal communication) and because scanner type was not a factor in the multivariate model, we do not believe scanner type had an effect on the comparisons. Another limitation was that the low- and standard-dose scans compared were obtained after an average interval of 3 months. Although we are unaware of any longitudinal imaging studies addressing rates of emphysema progression, it would be unlikely for the actual amount of emphysema to change measurably in that interval. Although lung volume increases over time as emphysema progresses, we do not believe this change was a factor given the trend toward greater volume on the low-dose scans, which were obtained first in nearly all cases. Finally, our study does not provide reasons for the small differences found. Measurements of air and blood attenuation obtained for this study along with daily water phantom checks performed for clinical quality assurance excluded improper scanner calibration and drift in attenuation value as explanations; measurements of the SD of the mean attenuation (Table 3) confirmed the expected relation in which noise decreases according to the square root of the increase in radiation exposure (e.g., if exposure is doubled, SD decreases by the square root of 2) [24, 25].
The results of this study apply only to effective tube current values in the range of 30-250 mAs. The effect of a greater reduction in radiation dose is unknown. It is likely that a critical effective tube current level or range exists below which signal-to-noise levels are too low for accurate measurement of emphysema index. The effect of dose with smaller reconstructed slice thickness, which would increase noise levels for both low-dose and standard-dose scans, also is unknown. Studies of isolated lungs or an animal model with the same scanner and multiple doses and slice thicknesses would be useful for defining the effects of lower doses and smaller slice thicknesses. Without this information, it is nonetheless recognizable that reconstruction techniques should be kept constant in comparative quantitative CT studies of emphysema.
We conclude that in addition to its potential value as a tool for early detection of lung cancer, low-dose MDCT has substantial potential value for quantitative studies of emphysema. The differences in emphysema index due to radiation dose were found to be minimal. Although statistically significant, these differences are likely to be of little clinical significance, particularly in relation to the variation due to differences in lung volume that can occur. Therefore it is much more important to try to optimize the consistency of breath-holding efforts to minimize variation due to technical factors. The clinical acceptability of performing CT primarily to quantify emphysema would likely be greater with low-dose technique, because of the substantially reduced radiation exposure. Possible investigational uses include cross-sectional and longitudinal studies to learn more about the interactions between cigarette smoking and other risk factors, genotype-phenotype relations, the natural history of the earliest stages of emphysema in younger smokers, and individual rates of emphysema progression. Low-dose scanning is also of potential value in clinical trials assessing pharmacologic therapies for emphysema as an alternative to standard-dose techniques that have been used [6, 7] and may help define the appropriate point to begin any pharmacologic intervention proved effective.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  B. Siewert, J. Sosna, A. McNamara, V. Raptopoulos, and J. B. Kruskal Quality Initiatives: Missed Lesions at Abdominal Oncologic CT: Lessons Learned from Quality Assurance RadioGraphics, May 1, 2008; 28(3): 623 - 638. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
