Automated Diseased Lung Volume Percentage Calculation in Quantitative CT Evaluation of Chronic Obstructive Pulmonary Disease and Idiopathic Pulmonary Fibrosis : Journal of Computer Assisted Tomography

Secondary Logo

Journal Logo

Thoracic Imaging

Automated Diseased Lung Volume Percentage Calculation in Quantitative CT Evaluation of Chronic Obstructive Pulmonary Disease and Idiopathic Pulmonary Fibrosis

Kitaguchi, Yoshiaki MD; Fujimoto, Keisaku MD; Droma, Yunden MD; Yasuo, Masanori MD; Wada, Yosuke MD; Ueno, Fumika MD; Kinjo, Takumi MD; Kawakami, Satoshi MD; Fukushima, Kiyoyasu MD§; Hanaoka, Masayuki MD

Author Information
Journal of Computer Assisted Tomography 45(4):p 649-658, 7/8 2021. | DOI: 10.1097/RCT.0000000000001182
  • Open

Abstract

Quantitative computed tomography (CT) analysis is a rapidly growing field of radiomics of practices of extracting, analyzing, and interpreting quantitative data from medical images to aid in disease diagnosis and prognosis. Quantitative analysis can range from simple threshold measurements to texture metrics for evaluations of changes in features over time. Quantitative CT has been applied to various obstructive, infiltrative, and restrictive pulmonary diseases including chronic obstructive pulmonary disease (COPD)/emphysema, cystic fibrosis, asthma, idiopathic pulmonary fibrosis (IPF), hypersensitivity pneumonitis, connective tissue–related interstitial lung disease (ILD), and combined pulmonary fibrosis and emphysema (CPFE).1

In history, the software-based quantification of emphysema that was based on the lung density threshold in CT images was developed more than 3 decades ago.2 After that, the quantitative CT was applied to patients with COPD with increasingly ulilization.3–5 Some of software-based quantitative CT analysis methods are currently used in clinical practice at specialist centers.6 Furthermore, it has been reported that, in approximately 30% of patients, emphysema complicated IPF.7,8 Recently, this disease state was termed CPFE. Emphysema and pulmonary fibrosis can sometimes be superimposed on chest CTs, which makes it difficult to visually assess separately the extents of emphysema and pulmonary fibrosis. Therefore, a software-based quantitative CT analysis that simultaneously detects emphysema and pulmonary fibrosis might be useful.

Several software-based quantitative CT analysis methods have been developed for assessing emphysema and ILD, including the density histogram analysis, the density mask technique, and the texture classification method. Although the texture classification method appeared to be more successful than the other methods,6 the software programs are not commercially available, to our knowledge. In contrast, the density mask technique is the most widely used method. A threshold on the order of −950 Hounsfield units (HU) was identified for the quantification of emphysema in patients with COPD.2,6,9–12 The density mask technique is also convenient for evaluating disease status in patients with ILD.13,14 A threshold value of −950 HU could distinguish emphysema from normal lungs, and a threshold value of −700 or −750 HU could distinguish normal lungs from ground-glass opacity.6 Thus, the density mask technique has universal applicability, and the results are highly consistent with visual assessments, because the CT values provide clear cutoff points.6 A previous study used semiautomatic software to evaluate simultaneously the percentage of low attenuation area (%LAA) and the percentage of high attenuation area (%HAA) in 4 CT slices selected from CT images. They showed that the %LAA and %HAA were independent contributors to the lung diffusion capacity for carbon monoxide (DLCO) in patients with CPFE.15 Moreover, a longitudinal study highlighted the usefulness of the percentage of destructed lung area, defined as %LAA + %HAA.16

The present study aimed to confirm the usefulness of a new software program that applied the density mask technique. Previous reports showed that the texture classification was best for quantitative CT analysis.6 However, we aimed to demonstrate that the density mask technique was also effective for quantitative CT analyses. The software program we tested could perform automatic, simultaneous assessments of the low attenuation volume (LAV) and high attenuation volume (HAV), based on volumetric image data from whole lungs of patients with the clinical diagnosis of COPD or IPF. In the first part of this study, we assessed the validity of the quantitative assessment performed by the software program, compared with a visual assessment. We determined cutoff values for the LAV and HAV that could identify emphysema and pulmonary fibrosis. In the second part of this study, we confirmed the usefulness of the quantitative CT analysis by comparing pulmonary function parameters with quantitative CT assessment parameters in patients with COPD and IPF. We then investigated whether the utility of the quantitative CT analysis could be improved by assessing the diseased lung volume (DLV), defined as the LAV + HAV. In addition, we calculated the composite physiologic index (CPI) based on pulmonary function parameters. It is suggested that the CPI is a more accurate prognostic determinant for patients meeting the clinical definition of IPF as compared with individual pulmonary function parameters.17 The correlations between the CPI and the quantitative CT assessment parameters were investigated.

MATERIALS AND METHODS

Subjects and Protocol

Sixty-two patients with stable COPD were recruited from Shinshu University Hospital before any treatment of COPD in a period from April 2016 to October 2019 for the current study. The COPD diagnosis was based on the GOLD Report with FEV1/FVC < 70% after inhalation of a β2 agonist.18 Twenty patients were excluded due to bronchiectasis, typical asthma, lymphangioleiomyomatosis, lung cancer, cardiovascular diseases, or late sequelae of pulmonary tuberculosis. Two patients were excluded because of respiratory tract infection and COPD exacerbation before 6 months of the study.

Sixty-three patients with IPF were recruited from Shinshu University Hospital in a period from April 2016 to October 2019 for the current study. The IPF diagnosis was based on the ATS/ERS/JRS/ALAT guidelines.19,20 Twenty patients were excluded from the current study due to idiopathic interstitial pneumonias other than IPF, chronic hypersensitivity pneumonitis, ILD due to autoimmune disease, drug-induced ILD, sarcoidosis, pneumoconiosis, lung cancer, cardiovascular diseases, or late sequelae of pulmonary tuberculosis. Three patients were excluded because of respiratory tract infection and IPF exacerbation before 6 months of the study.

At the patients' first visits to the outpatient clinic, chief complains, symptoms, smoking history, clinical course, and complications were obtained from the patients by attending physicians. In the meantime, physical examination, laboratory examination, and chest x-ray were performed for primary diagnoses of COPD or IPF. In their second medical visits, pulmonary function tests and chest CT scanning were performed. The chest CT images were initially reviewed by 2 experts of pulmonologists (Y.K. and K.F., with 20 and 38 years of working experience, respectively) who were blinded to the patients' clinical data. Then, all clinical data were reviewed to determine the diagnoses of COPD or IPF based on the diagnostic criteria. In occasional case, the diagnosis of IPF was made by a multidisciplinary team discussion in our department. As a result, this prospective cohort study recruited 40 consecutive patients with COPD and 40 consecutive patients with IPF. Eleven individuals that were healthy nonsmokers were included as control subjects. These subjects had no respiratory or cardiovascular disease, but they underwent spirometry and a chest CT for this study.

We investigated correlations between the quantitative assessments performed by the software program and the visual assessments performed as described previously21–26 in patients with COPD or IPF. We determined the sensitivity and specificity of the LAV and HAV parameters for detecting emphysema and pulmonary fibrosis, respectively, with the visual assessments as reference. We also investigated correlations between pulmonary function parameters and the quantitative CT assessment parameters.

The study protocol was approved by the institutional review board of Shinshu University School of Medicine (Matsumoto, Japan). Each subject provided written informed consent to participate in the study (approval number: 3294, date of approval: December 9, 2015).

Quantitative CT Assessment Compared With Visual CT Assessment

All patients underwent chest CTs during an inspiratory breath-hold in the supine position with a 64-row multidetector CT scanner (LightSpeed VCT; GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom). The CT scanner settings were 120 kV tube voltage, variable tube current, 64 × 0.625 mm collimation, and a 0.4-second rotation time. Image reconstruction was performed with the standard algorithm and a slice thickness of 0.625 mm. After scanning, we obtained the Digital Imaging and Communications in Medicine–formatted volumetric image data for the whole lungs. The CT images were automatically analyzed on our Windows computer with an image-analyzing software program (INTAGE Station LungVision version 3.0; Cybernet, Inc, Tokyo, Japan). With automatic procedures, we could isolate the lung parenchyma from the mediastinum and thoracic wall, and then determine the percentages of LAV (LAV%) and HAV (HAV%) within a few minutes. The LAV% was defined as the percent of lung tissue with densities of −950 HU or less, based on previous findings.2,9–12 The HAV% was defined as the percent of lung tissue with densities of −750 HU or greater, based on previous findings.6 The percentage of DLV% was defined as LAV% + HAV%.

We investigated whether the LAV% and HAV% could reflect the extent of emphysema and pulmonary fibrosis, respectively. We assessed the validity of the quantitative assessment produced by the software program compared with a visual assessment performed with an established, previously described method.21–26 Emphysema was scored visually in the bilateral upper, middle, and lower lung fields, according to the methods of Goddard et al.27 The visual scores for low attenuation areas (LAAs) were calculated as the sum of the scores for the 6 lung fields (score range, 0–24). The visual detection of pulmonary fibrosis on chest CTs was performed as previously described.19,20 To grade severity, the extent of IPF was scored visually, as previously described.28 The chest CT images were reviewed by 2 expert pulmonologists (Y.K. and K.F., with 20 and 38 years of experience, respectively) in a model blind to clinical information of patients. They separately scored the extent of emphysema (the LAA score) and graded the severity on extent of interstitial change. With the cases of disagreement, discussion and reevaluation were performed to reach the agreement (Figs. 1, 2).

F1
FIGURE 1:
A–C, The figures were CT images of a 67-year-old man with COPD. The visual score in LAAs was 18. D–F, These CT images were corresponding to the figures of A, B, and C accordingly. They were obtained by using INTAGE Station LungVision version 3.0; Cybernet, Inc, Tokyo, Japan. The green color is for trachea and bronchi; the dark brown for normal attenuated lung area (between −950 and −750 HU); yellow for LAAs (lower than −950 HU); red for HAAs (higher than −750 HU). The LAV% was 20.7%; HAV% was 10.4%; and DLV% was 31.1%.
F2
FIGURE 2:
A–C, The figures were CT images of a 55-year-old man with IPF. The visual score of LAAs was zero. The severity of pulmonary fibrosis on chest CT was moderate. D–F, These CT images were corresponding to the figures of A, B, and C accordingly. They were obtained by using INTAGE Station LungVision version 3.0; Cybernet, Inc, Tokyo, Japan. The green color is for trachea and bronchi; the dark brown for normal attenuated lung area (between −950 and −750 HU); yellow for LAAs (lower than −950 HU); red for HAAs (higher than −750 HU). The LAV% was 0.3%; HAV% was 32.9%; and DLV% was 33.2%.

Pulmonary Function Tests

All patients underwent spirometry and measurements of the lung DLCO, the DLCO corrected for alveolar volume (DLCO/VA), a global measure of ventilation heterogeneity (the slope of phase III of the single breath nitrogen washout test [delta N2]), the functional residual capacity (FRC), the total lung capacity (TLC), and the residual volume (RV). Measurements were performed with a pulmonary function testing system (Chestac-8900; CHEST Co, Ltd, Tokyo, Japan), as previously described.21–26 We also evaluated the CPI, which was calculated based on pulmonary function parameters, as follows: CPI = 91.0 – (0.65 × percent predicted DLCO) – (0.53 × percent predicted forced vital capacity [FVC]) + (0.34 × percent predicted forced expiratory volume in 1 second [FEV1]).17

Statistical Analyses

Statistical analyses were conducted with StatFlex version 6.0 (Artech, Osaka, Japan). Simple correlations were examined by calculating Pearson correlation coefficients. A receiver operating characteristic (ROC) curve analysis was performed to determine the sensitivity and specificity of the quantitative parameters, LAV% and HAV%, for detecting the presence of emphysema and pulmonary fibrosis, respectively, with the visual assessment as reference. In the ROC analysis, we also determined cutoff values for the LAV% and HAV% for detecting emphysema and pulmonary fibrosis, respectively, because these parameters also included normal anatomical structures, such as pulmonary arteries (in the HAV%). Therefore, cutoff values were determined to exclude normal structures and more accurately identify diseased areas. Comparisons of continuous variables between 2 groups were performed with the Student t test. Comparisons of continuous variables among 3 groups were performed with a 1-way analysis of variance, followed by the Tukey-Kramer multiple comparisons correction. P values less than 0.05 were considered significant in all statistical analyses.

RESULTS

Table 1 shows the clinical characteristics and chest CT findings in the healthy control, COPD, and IPF groups. Table 2 shows the details of pulmonary function in these groups. The FEV1/FVC was significantly lower, and the VC, FVC, FRC, RV, and TLC were significantly higher in the COPD group than in the IPF group. The DLCO was significantly lower in the IPF group than in the COPD group.

TABLE 1 - Clinical Characteristics and Chest CT Findings in the Healthy Control, COPD, and IPF Groups
Healthy Control (n = 11) COPD (n = 40) IPF (n = 40)
Age, y 45.00 ± 3.84 75.08 ± 1.38* 70.70 ± 1.53*
Sex
 Male/female, n/n 11/0 34/6 34/6
Body mass index, kg/m2 23.95 ± 0.65 22.18 ± 0.49 24.09 ± 0.36
Smoking index, pack-years 0.00 ± 0.00 47.59 ± 5.64* 31.00 ± 3.86*†
Chest CT findings
 Quantitative assessment performed by the software program
  Total lung volume, cm3 5625 ± 250 5734 ± 129 4182 ± 200*‡
  LAV%, % 0.45 ± 0.10 13.10 ± 2.26* 4.06 ± 0.99‡
  HAV%, % 9.85 ± 0.35 9.86 ± 0.23 22.79 ± 1.82*‡
  DLV%, % 10.27 ± 0.29 22.96 ± 2.15* 26.83 ± 1.90‡
 Visual assessment based on the previous reports
  Emphysema +/−, n/n 0/11 34/6 18/22
  Visual score of LAA 0.00 ± 0.00 10.45 ± 1.38* 5.10 ± 0.1.02†
  Severity of pulmonary fibrosis
   Mild, n 0 0 20
   Moderate, n 0 0 10
   Severe, n 0 0 10
Values are presented as the means ± standard error of the mean.
*P < 0.01 versus normal group.
P < 0.05 versus COPD group.
P < 0.01 versus COPD group.

TABLE 2 - Pulmonary Function and CPI in the Healthy Control, COPD and IPF Groups
Healthy Control (n = 11) COPD (n = 40) IPF (n = 40)
Pulmonary function
 VC, % predicted 114.28 ± 5.77 105.28 ± 3.09 87.03 ± 3.75*†
 FVC, % predicted 112.05 ± 5.76 103.64 ± 3.43 86.95 ± 3.95*†
 FEV1, % predicted 106.62 ± 5.49 69.98 ± 3.26* 79.08 ± 2.93*
 FEV1/FVC, % 84.46 ± 3.82 53.35 ± 2.13* 75.27 ± 2.14†
 FRC, % predicted NA 105.82 ± 4.05 82.28 ± 3.50†
 RV, % predicted NA 141.04 ± 5.55 93.82 ± 4.08†
 TLC, % predicted NA 119.98 ± 2.73 90.80 ± 3.28†
 DLCO, % predicted NA 68.52 ± 3.75 51.62 ± 2.78†
 DLCO/VA,% predicted NA 82.41 ± 4.30 85.00 ± 4.22
 Delta N2, %N2/L NA 3.46 ± 0.42 2.90 ± 0.55
CPI NA NA 38.25 ± 2.51
Values are presented as the means ± standard error of the mean.
*P < 0.01 versus healthy control group.
P < 0.01 versus COPD group.
NA indicates not applicable.

Figure 3A shows the correlation between the quantitative LAV% and the visual LAA score (r = 0.865, P < 0.001). The ROC analysis revealed that the LAV% could detect emphysema with 86.8% sensitivity and 84.2% specificity, with a visual assessment as reference, with 1.5% as the cutoff LAV% value (Fig. 3B). Figure 4A shows the correlation between the quantitative HAV% and the visual assessment of pulmonary fibrosis severity (r = 0.840, P < 0.001). The ROC analysis revealed that the HAV% could detect pulmonary fibrosis with 87.5% sensitivity and 96.1% specificity, with the visual assessment as reference, with 12% as the cutoff HAV% value (Fig. 4B).

F3
FIGURE 3:
A, The correlation between the LAV% and the visual score of LAAs in all subjects (n = 91). B, The ROC curve of LAV%. The ROC analysis revealed 86.8% sensitivity and 84.2% specificity for LAV% to detect emphysema on a visual assessment, using 1.5% as the cutoff value (area under the curve = 0.926).
F4
FIGURE 4:
A, The correlation between the HAV% and the severity of pulmonary fibrosis on chest CT in all subjects (n = 91). B, The ROC curve of HAV%. The ROC analysis revealed 87.5% sensitivity and 96.1% specificity of HAV% to detect pulmonary fibrosis on a visual assessment, using 12% as the cutoff value (area under the curve = 0.962).

We determined the correlations between the pulmonary function parameters and the quantitative assessment parameters for the COPD (Table 3) and IPF (Table 4) groups. We found significant correlations between the LAV% and the diffusion capacity of the lung (DLCO, DLCO/VA), between the LAV% and the delta N2, between the DLV% and the diffusion capacity of the lung (DLCO, DLCO/VA), and between the DLV% and the delta N2 in the COPD group. We also found significant correlations between the DLV% and DLCO, between the LAV% and DLCO/VA, and between the DLV% and the CPI in the IPF group. A correlation between the DLV% and DLCO was observed in patients with COPD (Fig. 5A; r = −0.7922, P < 0.0001) and patients with IPF (Fig. 5B; r = −0.7257, P < 0.0001). Figure 5C shows the correlation between the DLV% and DLCO in all patients with COPD and IPF (r = −0.7570, P < 0.0001). Figure 6 shows the correlation between the DLV% and the CPI in patients with IPF (r = 0.8254, P < 0.0001).

TABLE 3 - Correlations Between Pulmonary Function Parameters and Parameters of the Quantitative Assessment Performed by the Software Program (the Total Lung Volume, LAV%, HAV%, and DLV%) in the COPD Group (n = 40)
COPD
Total Lung Volume LAV% HAV% DLV%
r P r P r P r P
VC (% predicted) −0.0306 0.8514 −0.3743 0.0173 −0.1029 0.5276 −0.4037 0.0098
FVC (% predicted) −0.1445 0.3736 −0.4948 0.0012 0.0169 0.9176 −0.5176 0.0006
FEV1 (% predicted) −0.4092 0.0087 −0.6142 <0.0001 0.3365 0.0338 −0.6092 <0.0001
FEV1/FVC (%) −0.3554 0.0244 −0.4056 0.0094 0.4494 0.0036 −0.3783 0.0161
FRC (% predicted) 0.1050 0.5248 −0.2490 0.1264 0.0075 0.9640 −0.2615 0.1079
RV (% predicted) 0.5409 0.0004 0.5256 0.0006 −0.2967 0.0667 0.5213 0.0007
TLC (% predicted) 0.2698 0.0967 −0.1731 0.2919 −0.2234 0.1715 −0.2058 0.2087
DLCO (% predicted) −0.2372 0.1459 −0.7684 <0.0001 0.1472 0.3710 −0.7922 <0.0001
DLCO/VA (% predicted) −0.4196 0.0078 −0.7923 <0.0001 0.2618 0.1074 −0.8052 <0.0001
Delta N2, %N2/L 0.2641 0.1042 0.8256 <0.0001 −0.3958 0.0126 0.8260 <0.0001

TABLE 4 - Correlations Between Pulmonary Function Parameters and Parameters of the Quantitative Assessment Performed by the Software Program (the Total Lung Volume, LAV%, HAV%, and DLV%) in the IPF Group (n = 40)
IPF
Total Lung Volume LAV% HAV% DLV%
r P r P r P r P
VC (% predicted) 0.7305 <0.0001 0.2296 0.1542 −0.7079 <0.0001 −0.5594 0.0002
FVC (% predicted) 0.6410 <0.0001 0.2165 0.1796 −0.6245 <0.0001 −0.4859 0.0015
FEV1 (% predicted) 0.2806 0.0794 0.0090 0.9559 −0.4015 0.0102 −0.3810 0.0153
FEV1/FVC (%) −0.5874 0.0001 −0.3332 0.0356 0.5246 0.0005 0.3285 0.0385
FRC (% predicted) 0.5691 0.0001 −0.0630 0.6992 −0.5799 0.0001 −0.5902 0.0001
RV (% predicted) 0.6507 <0.0001 0.1804 0.2652 −0.5637 0.0002 −0.4468 0.0038
TLC (% predicted) 0.8451 <0.0001 0.2361 0.1424 −0.7609 <0.0001 −0.6068 <0.0001
DLCO (% predicted) 0.3476 0.0280 −0.5655 0.0001 −0.4469 0.0038 −0.7257 <0.0001
DLCO/VA (% predicted) −0.2981 0.0617 −0.7460 <0.0001 0.1068 0.5118 −0.2885 0.0710
Delta N2, %N2/L −0.4731 0.0024 0.0918 0.5784 0.5667 0.0002 0.5885 0.0001
CPI −0.6522 <0.0001 0.2445 0.1337 0.7244 <0.0001 0.8254 <0.0001

F5
FIGURE 5:
A, The correlation between the DLV% and the DLCO in patients with COPD (n = 40). B, The correlation between the DLV% and the DLCO in patients with IPF (n = 40). C, The correlation between the DLV% and the DLCO in all patients with COPD and IPF (n = 80).
F6
FIGURE 6:
The correlation between the DLV% (DLV%) and the CPI in patients with IPF (n = 40).

Figures 7A to D show the data for the total lung volume, LAV%, HAV%, and DLV% in the healthy control, COPD, and IPF groups. The total lung volume was significantly lower in the IPF group than in the other groups. The LAV% was significantly higher in the COPD group than in the other groups. The HAV% was significantly higher in the IPF group than in the other groups. The DLV% was significantly lower in the healthy control group than in the other groups. There was no significant difference in the DLV% between the COPD and IPF groups.

F7
FIGURE 7:
A, The box-and-whisker diagrams of the total lung volume in the healthy control (n = 11), COPD (n = 40), and IPF (n = 40) groups. B, The box-and-whisker diagrams of the low attenuation volume (LAV%) in the healthy control (n = 11), COPD (n = 40), and IPF (n = 40) groups. C, The box-and-whisker diagrams of the HAV% in the healthy control (n = 11), COPD (n = 40), and IPF (n = 40) groups. D, The box-and-whisker diagrams of the diseased lung volume (DLV%) in the healthy control (n = 11), COPD (n = 40), and IPF (n = 40) groups.

DISCUSSION

In this study, we applied the density mask technique with 2 thresholds to assess the LAV and HAV simultaneously for detecting emphysema and pulmonary fibrosis. The ROC analysis revealed that LAV% and HAV% had high sensitivity and high specificity for detecting emphysema and pulmonary fibrosis, respectively. Moreover, the DLV% was significantly correlated with the lung diffusion capacity in patients with COPD or IPF. The DLV% was also significantly correlated with the DLCO in all patients.

Imaging software programs can provide highly accurate, reproducible measurements.29 Some evidence has suggested that quantitative emphysema parameters were better correlated with morphological measures of emphysema than visual scoring.30,31 Although fully automated quantitative CT eliminates interobserver and intraobserver variation, it is important to note that high-quality quantitative CT analyses require verification by an expert analyst, particularly in clinical applications, where quantitative CT might be used to make a diagnosis or the findings might influence treatment decisions.29 Thus, the present study confirmed previous findings that showed that the results of a quantitative assessment performed with the software program were well correlated with the visual assessment based on an established method.21–26

A few methods, including the density histogram, density mask technique, and texture classification method, were reported to be potentially useful in clinical practice, due to their convenience, despite scarce evidence in large cohorts.6 The texture classification method can be used to differentiate the parenchymal pathology associated with emphysema and that associated with ILD.32 This method appeared to be more successful than the density mask analysis.6 However, the texture classification method is significantly more computationally expensive,6 and it cannot be widely used in clinical practice, because the software programs are not commercially available, to our knowledge. The density histogram can evaluate the extent of emphysema or pulmonary fibrosis with kurtosis and skewness scores. However, these parameters are not “user friendly” in routine practice.13 In contrast, the density mask technique is more often used, because it is convenient and universally applicable.6,33 Our findings suggested that simultaneously assessing the LAV and HAV based on volumetric image data of whole lungs might improve the utility of the density mask technique. In addition, we propose that the density mask technique could be used, not only for potentially quantitatively evaluating lung diffusion capacity in patients with COPD and IPF, but also as a possible tool to screen for loss of lung diffusion capacity in diseases such as pneumonia, lung cancer, and pulmonary hypertension. DLV% values may also be helpful in predicting disease severity in patient with COPD and IPF. The increased DLV% values along with clinical courses of COPD and IPF may predict the disease severity and prognosis of these comorbidities. Thus, this software-based quantitative CT analysis is highly expected for benefit to patient management. Further study is necessary to verify this issue.

In a previous study, a semiautomatic software program was used for simultaneously assessing the %LAA and %HAA in 4 CT slices selected from CT images of patients with CPFE.15 That study revealed that the %HAA and the percentage of attenuation area (%AA), defined as %LAA + %HAA, were significantly correlated with DLCO. Our results were consistent with those results, although the patient populations were different. We assessed the LAV% and HAV% simultaneously, based on volumetric data from whole lungs in patients with COPD and IPF. We found that the DLV% was significantly correlated with the DLCO. This finding was equivalent in patients with COPD, patients with IPF, and all patients with COPD and IPF. In this study, 18 of 40 patients with IPF had emphysema in addition to pulmonary fibrosis, so-called CPFE. The patients with CPFE showed obstructive damage due to emphysema and restrictive damage due to pulmonary fibrosis in pulmonary function, resulting in confounded airflow pattern in CPFE. The software-based quantitative CT analysis is expected to be valuable to evaluate the correlation of the lung parenchyma with pulmonary function in CPFE.

We found that, in patients with COPD, the DLV% was slightly better correlated with the DLCO and DLCO/VA (which reflected lung diffusion capacity) compared with the LAV%. The LAV% was previously shown to be negatively correlated with the DLCO and DLCO/VA in patients with COPD.34 In addition, the HAAs, defined as regions with an attenuation between −600 and −250 HU, were shown to be associated with cigarette smoking.35 The HAA was also shown to be associated with biomarkers of inflammation, extracellular matrix remodeling, reduced lung function, and an increased risk of death among community-dwelling adults.36 These observations might explain why the DLV% (defined as LAV% + HAV%) was better correlated with lung diffusion capacity than the LAV%, because in the present study, all patients with COPD had a smoking history of >10 pack-years, and the pathological changes observed in COPD included chronic inflammation.18

Our finding that the DLV% was strongly correlated with lung diffusion capacity parameters (Figs. 5A–C) suggested that the DLV%, which is easily obtained from a chest CT, might be useful for predicting lung diffusion capacity in patients with COPD or IPF that cannot undergo pulmonary function tests. In addition, the DLV% was strongly correlated with the CPI in patients with IPF. The CPI was developed to improve the previous prognostic measures in IPF by adjusting for emphysema on chest CT and pulmonary function test. It provided a better fit to IPF-related mortality data due to a correction for the confounding effects of emphysema.17 Consequently, the mortality rate associated with IPF was predicted more accurately with the CPI than with any other pulmonary function parameter.17 Thus, the DLV% might be useful for predicting mortality in patients with IPF. However, additional studies are required to confirm this hypothesis.

We found no significant difference in the HAV% between the healthy control and COPD groups. This finding was inconsistent with the reduction in alveolar and capillary surface areas associated with emphysema, based on the chest CTs of patients with COPD. However, in the present study, some patients with COPD had very little emphysema combined with bronchial wall thickness, which might have been detected as an increase in the HAV%. Consequently, we speculated that the heterogeneity of COPD, with and without emphysema and/or bronchial wall thickness, led to our finding that the average HAV% in the COPD group was equivalent to that observed in the healthy control group.

The present study had several limitations that warrant mention. First, the study was performed in a single academic institution with a single type of CT scanner, so the sample sizes were not big enough to reach powerful statistics. In addition, the data were heavily skewed to males in all volunteers (n = 11) and 34 of 40 patients of both COPD and IPF groups due to originally dominances of COPD and IPF in males. Nevertheless, the sex ratio was matched in the 3 groups of health controls, COPD, and IPF. Additional prospective studies with larger samples are required to proof the current results. Second, some patients in the IPF group did not have pathological confirmation of IPF. However, the role of CT has been expanded to include making a diagnosis of IPF without a surgical lung biopsy. Although the diagnosis of IPF was based on the ATS/ERS/JRS/ALAT guidelines,19,20 IPF cases were identified with 1 of 2 different definitions, depending on the time of diagnosis. These different definitions might have affected the results. Third, the image acquisition and analysis protocols might have differed between this study and previous studies that used quantitative CT. This difference was due to the lack of an international consensus concerning the methodology of quantitative CT. In future, a consensus on the methodology should be conducted to standardize quantitative CT.

CONCLUSIONS

We demonstrated that a quantitative CT analysis performed with a commercially available software program might be useful in clinical practice, and it provided the advantage of simple, automatic procedures. The present results suggest quantitative CT analysis may improve the precision of the assessment of DLV%, which itself could be a useful tool in predicting lung diffusion capacity in patients with the clinical diagnosis of COPD or IPF. The strong correlation between the DLV% and the CPI are worthy of additional studies to confirm that the DLV% can predict mortality in patients with IPF.

REFERENCES

1. Chen A, Karwoski RA, Gierada DS, et al. Quantitative CT analysis of diffuse lung disease. Radiographics. 2020;40:28–43.
2. Müller NL, Staples CA, Miller RR, et al. "density mask". An objective method to quantitate emphysema using computed tomography. Chest. 1988;94:782–787.
3. Grydeland TB, Dirksen A, Coxson HO, et al. Quantitative computed tomography: emphysema and airway wall thickness by sex, age and smoking. Eur Respir J. 2009;34:858–865.
4. Nakano Y, Wong JC, de Jong PA, et al. The prediction of small airway dimensions using computed tomography. Am J Respir Crit Care Med. 2005;171:142–146.
5. Fernandes L, Fernandes Y, Mesquita AM. Quantitative computed tomography imaging in chronic obstructive pulmonary disease. Lung India. 2016;33:646–652.
6. Ohkubo H, Nakagawa H, Niimi A. Computer-based quantitative computed tomography image analysis in idiopathic pulmonary fibrosis: a mini review. Respir Investig. 2018;56:5–13.
7. Mejía M, Carrillo G, Rojas-Serrano J, et al. Idiopathic pulmonary fibrosis and emphysema: decreased survival associated with severe pulmonary arterial hypertension. Chest. 2009;136:10–15.
8. Tasaka S, Mizoguchi K, Funatsu Y, et al. Cytokine profile of bronchoalveolar lavage fluid in patients with combined pulmonary fibrosis and emphysema. Respirology. 2012;17:814–820.
9. Wang Z, Gu S, Leader JK, et al. Optimal threshold in CT quantification of emphysema. Eur Radiol. 2013;23:975–984.
10. Johannessen A, Skorge TD, Bottai M, et al. Mortality by level of emphysema and airway wall thickness. Am J Respir Crit Care Med. 2013;187:602–608.
11. Tho NV, Trang le TH, Murakami Y, et al. Airway wall area derived from 3-dimensional computed tomography analysis differs among lung lobes in male smokers. PLoS One. 2014;9:e98335.
12. Higami Y, Ogawa E, Ryujin Y, et al. Increased epicardial adipose tissue is associated with the airway dominant phenotype of chronic obstructive pulmonary disease. PLoS One. 2016;11:e0148794.
13. Shin KE, Chung MJ, Jung MP, et al. Quantitative computed tomographic indexes in diffuse interstitial lung disease: correlation with physiologic tests and computed tomography visual scores. J Comput Assist Tomogr. 2011;35:266–271.
14. Ohkubo H, Kanemitsu Y, Uemura T, et al. Normal lung quantification in usual interstitial pneumonia pattern: the impact of threshold-based volumetric CT analysis for the staging of idiopathic pulmonary fibrosis. PLoS One. 2016;11:e0152505.
15. Matsuoka S, Yamashiro T, Matsushita S, et al. Quantitative CT evaluation in patients with combined pulmonary fibrosis and emphysema: correlation with pulmonary function. Acad Radiol. 2015;22:626–631.
16. Matsuoka S, Yamashiro T, Matsushita S, et al. Morphological disease progression of combined pulmonary fibrosis and emphysema: comparison with emphysema alone and pulmonary fibrosis alone. J Comput Assist Tomogr. 2015;39:153–159.
17. Wells AU, Desai SR, Rubens MB, et al. Idiopathic pulmonary fibrosis: a composite physiologic index derived from disease extent observed by computed tomography. Am J Respir Crit Care Med. 2003;167:962–969.
18. Global Initiative for Chronic Obstructive Lung Disease [homepage on the Internet]. Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Pulmonary Disease. NHLBI/WHO Workshop Report. Bethesda, MD: National Heart, Lung and Blood Institute; 2001 [Updated 2020]. Available at: http://www.goldcopd.com. Accessed November 1, 2020.
19. Raghu G, Collard HR, Egan JJ, et al. An official ATS/ERS/JRS/ALAT statement: idiopathic pulmonary fibrosis: evidence-based guidelines for diagnosis and management. Am J Respir Crit Care Med. 2011;183:788–824.
20. Raghu G, Remy-Jardin M, Myers JL, et al. Diagnosis of idiopathic pulmonary fibrosis. An official ATS/ERS/JRS/ALAT clinical practice guideline. Am J Respir Crit Care Med. 2018;198:e44–e68.
21. Kitaguchi Y, Fujimoto K, Hanaoka M, et al. Pulmonary function impairment in patients with combined pulmonary fibrosis and emphysema with and without airflow obstruction. Int J Chron Obstruct Pulmon Dis. 2014;9:805–811.
22. Ueno F, Kitaguchi Y, Shiina T, et al. The preoperative composite physiologic index may predict mortality in lung cancer patients with combined pulmonary fibrosis and emphysema. Respiration. 2017;94:198–206.
23. Kitaguchi Y, Fujimoto K, Hayashi R, et al. Annual changes in pulmonary function in combined pulmonary fibrosis and emphysema: over a 5-year follow-up. Respir Med. 2013;107:1986–1992.
24. Kitaguchi Y, Fujimoto K, Hanaoka M, et al. Clinical characteristics of combined pulmonary fibrosis and emphysema. Respirology. 2010;15:265–271.
25. Kitaguchi Y, Fujimoto K, Kubo K, et al. Characteristics of COPD phenotypes classified according to the findings of HRCT. Respir Med. 2006;100:1742–1752.
26. Fujimoto K, Kitaguchi Y, Kubo K, et al. Clinical analysis of chronic obstructive pulmonary disease phenotypes classified using high-resolution computed tomography. Respirology. 2006;11:731–740.
27. Goddard PR, Nicholson EM, Laszlo G, et al. Computed tomography in pulmonary emphysema. Clin Radiol. 1982;33:379–387.
28. Kurashima K, Takayanagi N, Tsuchiya N, et al. The effect of emphysema on lung function and survival in patients with idiopathic pulmonary fibrosis. Respirology. 2010;15:843–848.
29. Herth FJF, Kirby M, Sieren J, et al. The modern art of reading computed tomography images of the lungs: quantitative CT. Respiration. 2018;95:8–17.
30. Bankier AA, De Maertelaer V, Keyzer C, et al. Pulmonary emphysema: subjective visual grading versus objective quantification with macroscopic morphometry and thin-section CT densitometry. Radiology. 1999;211:851–858.
31. Cavigli E, Camiciottoli G, Diciotti S, et al. Whole-lung densitometry versus visual assessment of emphysema. Eur Radiol. 2009;19:1686–1692.
32. Xu Y, van Beek EJ, Hwanjo Y, et al. Computer-aided classification of interstitial lung diseases via MDCT: 3D adaptive multiple feature method (3D AMFM). Acad Radiol. 2006;13:969–978.
33. Van Tho N, Wada H, Ogawa E, et al. Recent findings in chronic obstructive pulmonary disease by using quantitative computed tomography. Respir Investig. 2012;50:78–87.
34. Crossley D, Renton M, Khan M, et al. CT densitometry in emphysema: a systematic review of its clinical utility. Int J Chron Obstruct Pulmon Dis. 2018;13:547–563.
35. Lederer DJ, Enright PL, Kawut SM, et al. Cigarette smoking is associated with subclinical parenchymal lung disease: the multi-ethnic study of atherosclerosis (MESA)-lung study. Am J Respir Crit Care Med. 2009;180:407–414.
36. Podolanczuk AJ, Oelsner EC, Barr RG, et al. High attenuation areas on chest computed tomography in community-dwelling adults: the MESA study. Eur Respir J. 2016;48:1442–1452.
Keywords:

quantitative CT; COPD; IPF; CPFE; LungVision

Copyright © 2021 The Author(s). Published by Wolters Kluwer Health, Inc.