Abstract
Background
Sarcoidosis staging primarily has relied on the Scadding chest radiographic system, although chest CT imaging is finding increased clinical use.
Research Question
Whether standardized chest CT scan assessment provides additional understanding of lung function beyond Scadding stage and demographics is unknown and the focus of this study.
Study Design and Methods
We used National Heart, Lung, and Blood Institute study Genomics Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) cases of sarcoidosis (n = 351) with Scadding stage and chest CT scans obtained in a standardized manner. One chest radiologist scored all CT scans with a visual scoring system, with a subset read by another chest radiologist. We compared demographic features, Scadding stage and CT scan findings, and the correlation between these measures. Associations between spirometry and diffusing capacity of the lungs for carbon monoxide (Dlco) results and CT scan findings and Scadding stage were determined using regression analysis (n = 318). Agreement between readers was evaluated using Cohen’s κ value.
Results
CT scan features were inconsistent with Scadding stage in approximately 40% of cases. Most CT scan features assessed on visual scoring were associated negatively with lung function. Associations persisted for FEV1 and Dlco when adjusting for Scadding stage, although some CT scan feature associations with FVC became insignificant. Scadding stage was associated primarily with FEV1, and inclusion of CT scan features reduced significance in association between Scadding stage and lung function. Multivariable regression modeling to identify radiologic measures explaining lung function included Scadding stage for FEV1 and FEV1 to FVC ratio (P < .05) and marginally for Dlco (P < .15). Combinations of CT scan measures accounted for Scadding stage for FVC. Correlations among Scadding stage and CT scan features were noted. Agreement between readers was poor to moderate for presence or absence of CT scan features and poor for degree and location of abnormality.
Interpretation
In this study, CT scan features explained additional variability in lung function beyond Scadding stage, with some CT scan features obviating the associations between lung function and Scadding stage. Whether CT scan features, phenotypes, or endotypes could be useful for treating patients with sarcoidosis needs more study.
Key Words: chest CT imaging, chest radiography, Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS), lung function, phenotypes, sarcoidosis, Scadding stage
Take-home Points.
Study Question: Does standardized CT scan assessment provide additional understanding of lung function beyond Scadding stage and demographics?
Results: In this study, associations between lung function and CT scan feature persisted for FEV1 and diffusing capacity of the lungs for carbon monoxide when adjusting for Scadding stage and demographics, with some CT scan features obviating the associations between lung function and Scadding stage.
Interpretation: These data support that additional information regarding lung function in patients with sarcoidosis is gained from CT imaging over Scadding stage and demographics alone, suggesting a potential role for CT imaging to guide sarcoidosis clinical management.
Sarcoidosis is a multisystemic disease characterized by granulomatous inflammation.1 Up to 90% of patients show lung involvement, demonstrated by abnormalities on chest radiography (CXR), pulmonary function testing (PFT), or both.1, 2, 3 Scadding stage4 has long been used to assess CXR abnormalities with relative ease of imaging, low cost, and low risk.5,6 Regardless, Scadding staging has several limitations.5, 6, 7
Unlike traditional clinical staging systems in which patients progress from one stage to the next (eg, in cancer), patients with sarcoidosis do not demonstrate temporal or linear progression from one stage to the next, or progress through all stages.5 Although stage IV disease portends a worse prognosis with more severe disease unlikely to resolve and patients with stage 0 and I disease are more likely to achieve disease resolution,4 overall, the prognostic capabilities of Scadding staging are limited.8 Individuals may demonstrate varying extents or types of CXR abnormalities and may be classified in the same Scadding stage.4, 5, 6 Individuals with stage IV disease with fibrosis due to architectural or airway distortion vs honeycombing with traction bronchiectasis are not differentiated, even though they may have distinct PFT abnormalities.9 Furthermore, Scadding stage does not correlate reliably with lung function abnormalities or physiologic impairment,10, 11, 12, 13 and changes in Scadding stage may not correspond with PFT changes.14 These myriad challenges make assessment of disease status difficult. Other imaging modalities may allow more accurate assessment of pulmonary sarcoidosis and may provide a relationship with physiologic impairment.
High-resolution CT (HRCT) imaging of the chest is being used more frequently in sarcoidosis, with potential increased sensitivity to detect parenchymal abnormalities (PAs) and better estimation of disease status and lung impairment.14 Chest HRCT imaging detects and better quantifies micronodules, ground-glass, and reticular abnormalities that may not be apparent on CXR.7,9,15 Furthermore, chest CT imaging may provide prognostic findings with demonstration of nodules and ground-glass opacities associated with disease resolution, whereas reticular abnormalities are less likely to be associated with resolution.16, 17, 18 Standardized objective assessment of CT findings also has correlated with lung function, further supporting a role for chest HRCT imaging in sarcoidosis assessment.18, 19, 20
Although different scoring systems have been proposed to quantify CT scan findings, including by Oberstein et al21 and Remy-Jardin et al,19 they are limited by small studies and lack of standardized imaging algorithms. As part of the extensive clinical phenotyping obtained through the multicenter Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) study, PFT results, CXR, and standardized chest HRCT imaging were obtained.22,23 We sought to evaluate if chest HRCT imaging findings would provide information beyond conventional CXR Scadding stage alone in regards to lung function impairment in patients with pulmonary sarcoidosis.
Study Design and Methods
Study Enrollment, Design, and Data
Patients were enrolled as described previously22,23 at multiple centers throughout the United States based on the GRADS protocol (N = 368). Patients gave written informed consent according to the site’s institutional review board review of the overall protocol. Patient diagnosis was confirmed according to accepted criteria,2 including consistent clinical features, exclusion of alternative diagnoses, and biopsy showing nonnecrotizing granulomas. Patients underwent self-administered questionnaires, including self-reported race, research CXR, and chest HRCT imaging,24 as well as PFT.25,26 See e-Appendix 1 for additional details.
CT Imaging Scoring Methods
A single visual assessment score (VAS) was derived for each chest HRCT scan, including Oberstein score21 (e-Table 1). The scoring system was developed in conjunction with a chest radiologist (C. F.) to evaluate the most common abnormalities noted in sarcoidosis and α1-antitrypsin deficiency according to the original design of the GRADS study.22 A single chest radiologist (C. F.) anonymized to disease diagnosis reviewed all chest HRCT scans using this system; a second chest radiologist (I. C.) interpreted the first 168 CT scans using an abbreviated VAS, limited to sarcoidosis variables without quantification of severity and Oberstein scoring (see e-Appendix 1).
Statistical Analysis
Baseline demographic features, clinical measures, and chest HRCT imaging VAS measures were compared by Scadding stage (n = 351). For baseline demographic features and clinical measures, χ2 tests and one-way ANOVA tests were used for discrete and continuous variables, respectively. For CT scan measures, Fisher exact tests were used. Summary tables for the complete data sets (n = 318) are in e-Tables 2 and 3.
We developed definitions of inconsistency between CXR and CT scan findings based on Scadding stage and the VAS (n = 351) (e-Table 4). The observed distribution of inconsistencies stratified by Scadding stage was computed, along with percentage of inconsistent findings.
Associations between PFT and CT scan findings were evaluated using linear regression for those with complete PFT data, age, height, BMI, sex, race/ethnicity, Scadding stage, and GRADS study CT imaging (n = 318). When more than five individuals demonstrated a sarcoidosis CT scan feature of interest, it was included in the analysis. All models were adjusted for age, sex, height, race/ethnicity, and BMI. We used P = .01 for statistical significance. We performed regression modeling, with separate models fitted for PFT outcomes: FVC, FEV1, and FEV1 to FVC ratio measured before and after bronchodilator administration and diffusing capacity of the lungs for carbon monoxide (Dlco) measured after bronchodilator administration. For each PFT measure, two base models were fitted: one included only age, height, BMI, sex, and race/ethnicity and a second included these variables and Scadding stage. One CT scan measure was added to each base model to assess whether the CT scan measure was independently associated with PFT findings. Multivariable regression modeling was performed including all CT scan measures and Scadding stage and applying backward selection, using P values from a partial F test for significance of single-measure inclusion with cutoff of P = .15 as stopping criterion for the backward selection. Age, sex, height, race/ethnicity, and BMI were forced in all models. Two CT scan measures, presence of micronodules and overall Oberstein score, were not included in multivariable modelling because we used micronodule distribution and the six Oberstein component scores for selection. In total, five demographic measures, 23 CT scan measures, and Scadding stage were considered in multivariable modelling, totaling 29 features, or 48 accounting for representation of nominal variables.
To quantify associations among measures descriptively, a heatmap of pairwise associations between CT scan VAS measures and/or Scadding stage was created using Cramér’s V with a bias correction. In accordance with Cohen,27 we conservatively considered cutoffs of 0.1, 0.3, and 0.5 for small, moderate, and large associations, respectively.
Cohen’s κ value was used to compute agreement between VAS results from two readers using both dichotomized assessment (presence or absence) and assessment of location and extent of abnormality, when available. A CT scan variable was not included if fewer than five individuals demonstrated the condition by one reviewer.
Results
Three hundred fifty-one patients with sarcoidosis enrolled in the GRADS study had complete CT scan VAS and Scadding stage data (Table 1), with Scadding stage reflecting the recruitment strategy of the GRADS study.22 The distributions of sex, education status, and income status were comparable across stages. At enrollment, modest differences in Scadding stage were noted by race/ethnicity (P = .09), with most White patients having stage I or II disease and most Black patients having stage II or IV disease. Patients with stage II or IV disease had significantly lower BMI than those with other stages of disease. Patients with stage IV disease were slightly older, had obstructive lung disease, and had decreased Dlco compared with patients with all other stages of disease.
Table 1.
Demographics and Pulmonary Function Test Results by Scadding Stage
| Variable | No. Missing | Scadding Stage |
Total (n = 351) | P Value | ||||
|---|---|---|---|---|---|---|---|---|
| 0 (n = 44 [12.5%]) | I (n = 75 [21.4%]) | II (n = 104 [29.6%]) | III (n = 46 [13.1%]) | IV (n = 82 [23.4%]) | ||||
| Demographics | ||||||||
| Sex | 4 | .15 | ||||||
| Female | 28 (63.6) | 44 (60.3) | 46 (44.7) | 23 (50.0) | 45 (55.6) | 186 (53.6) | ||
| Male | 16 (36.4) | 29 (39.7) | 57 (55.3) | 23 (50.0) | 36 (44.4) | 161 (46.4) | ||
| Race/ethnicity | 5 | .09 | ||||||
| Asian, American Indian, Alaska Native, or not identifying a single primary race | 1 (2.3) | 0 (0.0) | 6 (5.8) | 1 (2.2) | 2 (2.5) | 10 (2.9) | ||
| Black | 6 (13.6) | 12 (16.4) | 25 (24.3) | 12 (26.7) | 29 (35.8) | 84 (24.3) | ||
| Hispanic | 3 (6.8) | 4 (5.5) | 4 (3.9) | 1 (2.2) | 5 (6.2) | 17 (4.9) | ||
| White | 34 (77.3) | 57 (78.1) | 68 (66.0) | 31 (68.9) | 45 (55.6) | 235 (67.9) | ||
| Education status | 7 | .15 | ||||||
| High school | 8 (18.6) | 14 (19.2) | 24 (23.3) | 4 (8.7) | 17 (21.5) | 67 (19.5) | ||
| Some college | 10 (23.3) | 33 (45.2) | 29 (28.2) | 20 (43.5) | 27 (34.2) | 119 (34.6) | ||
| College diploma | 14 (32.6) | 11 (15.1) | 32 (31.1) | 11 (23.9) | 22 (27.8) | 90 (26.2) | ||
| Graduate or professional degree | 11 (25.6) | 15 (20.5) | 18 (17.5) | 11 (23.9) | 13 (16.5) | 68 (19.8) | ||
| Income, $ | 12 | .74 | ||||||
| < 50,000 | 14 (33.3) | 22 (31.0) | 27 (26.5) | 12 (26.1) | 31 (39.7) | 106 (31.3) | ||
| 50,000-99,999 | 12 (28.6) | 17 (23.9) | 30 (29.4) | 16 (34.8) | 23 (29.5) | 98 (28.9) | ||
| 100,000-149,999 | 7 (16.7) | 14 (19.7) | 20 (19.6) | 10 (21.7) | 14 (17.9) | 65 (19.2) | ||
| ≥ 150,000 | 9 (21.4) | 18 (25.4) | 25 (24.5) | 8 (17.4) | 10 (12.8) | 70 (20.6) | ||
| Age, y | 4 | 52.95 ± 9.42 | 50.61 ± 11.53 | 52.90 ± 9.40 | 50.81 ± 10.84 | 55.49 ± 9.01 | 52.75 ± 10.10 | .03 |
| Height, inches | 1 | 66.61 ± 4.05 | 67.01 ± 3.41 | 67.60 ± 4.26 | 66.74 ± 4.51 | 66.45 ± 4.07 | 66.97 ± 4.06 | .37 |
| BMI, kg/m2 | 1 | 33.36 ± 6.39 | 32.58 ± 7.68 | 29.46 ± 5.30 | 32.43 ± 7.28 | 28.14 ± 5.54 | 30.69 ± 6.61 | < .01 |
| Pulmonary function test parameter | ||||||||
| FVC, L | ||||||||
| Before bronchodilator administration | 12 | 3.74 ± 1.02 | 3.78 ± 1.07 | 3.79 ± 1.12 | 3.56 ± 1.14 | 2.98 ± 1.15 | 3.57 ± 1.15 | < .01 |
| After bronchodilator administration | 18 | 3.69 ± 1.04 | 3.77 ± 1.07 | 3.84 ± 1.10 | 3.58 ± 1.14 | 3.08 ± 1.15 | 3.60 ± 1.14 | < .01 |
| FEV1, L | ||||||||
| Before bronchodilator administration | 12 | 2.90 ± 0.76 | 2.89 ± 0.88 | 2.78 ± 0.87 | 2.63 ± 0.97 | 1.95 ± 0.87 | 2.61 ± 0.94 | < .01 |
| After bronchodilator administration | 18 | 2.92 ± 0.82 | 2.97 ± 0.91 | 2.91 ± 0.88 | 2.72 ± 0.99 | 2.08 ± 0.90 | 2.70 ± 0.96 | < .01 |
| FEV1 to FVC ratio, % | ||||||||
| Before bronchodilator administration | 12 | 78.00 ± 5.35 | 76.23 ± 8.35 | 73.49 ± 8.01 | 74.10 ± 11.93 | 66.03 ± 13.59 | 73.03 ± 10.68 | < .01 |
| After bronchodilator administration | 18 | 79.77 ± 9.32 | 78.42 ± 8.71 | 75.83 ± 8.16 | 76.14 ± 12.69 | 68.14 ± 13.93 | 75.08 ± 11.35 | < .01 |
| Dlco after bronchodilator administration, mL/min/mm Hg | 23 | 24.38 ± 7.44 | 24.18 ± 7.60 | 23.57 ± 7.21 | 21.54 ± 8.06 | 17.46 ± 7.12 | 22.11 ± 7.85 | < .01 |
Data are presented as No. (%) or mean ± SD unless otherwise indicated. Dlco = diffusing capacity of the lungs for carbon monoxide.
Classification of CT Scan Findings by CXR
The distribution of CT scan VAS measures tended to differ significantly by Scadding stage (Table 2). CT scan features that did not differ by Scadding stage were rare, with ≤ 10 patients having the CT scan feature (eg, emphysema, Oberstein component score pleural thickening). A number of individuals with Scadding stage 0 or I had indications of PA on CT imaging, including micronodules, ground-glass opacities, and mosaic attenuation (Table 3). Thirty-nine percent of patients showed CT scan findings inconsistent with CXR-based Scadding stage. For Scadding stages 0-III, at most 62% of patients with the given stage had consistent CT scan findings. Patients with stage 0 or I disease showed inconsistent findings most of the time. Specifically, 20.2% and 26.0% of patients with Scadding stage II or III disease showed PA on CXR but not on CT imaging, whereas 31.9% and 48.0% of patients with Scadding stage 0 or I disease showed PA on CT imaging but not CXR. The inconsistencies between CT imaging and CXR findings were accounted for by PA and lymphadenopathy, with similar rates for each separately and a smaller percentage for both together.
Table 2.
Distribution of Each CT Scan Visual Assessment Score Measure by Scadding Stage
| Variable | Scadding Stage |
Total (N = 351) | P Value | ||||
|---|---|---|---|---|---|---|---|
| 0 (n = 44 [12.5%]) | I (n = 75 [21.4%]) | II (n = 104, [29.6%]) | III (n = 46 [13.1%]) | IV (n = 82 [23.4%]) | |||
| Lymphadenopathy | |||||||
| Mediastinal lymphadenopathy | 15 (34.1) | 45 (60.0) | 74 (71.2) | 13 (28.3) | 49 (59.8) | 196 (55.8) | < .01 |
| Hilar lymphadenopathy | 10 (22.7) | 37 (49.3) | 63 (60.6) | 8 (17.4) | 39 (47.6) | 157 (44.7) | < .01 |
| Nodules | |||||||
| Micronodules | 9 (20.5) | 24 (32.0) | 66 (63.5) | 23 (50.0) | 42 (51.2) | 164 (46.7) | < .01 |
| Micronodule distributiona | < .01 | ||||||
| None | 36 (81.8) | 51 (68.0) | 40 (38.5) | 23 (50.0) | 40 (48.8) | 190 (54.1) | |
| Perilymphatic | 4 (9.1) | 6 (8.0) | 4 (3.8) | 2 (4.3) | 6 (7.3) | 22 (6.3) | |
| Peribronchovascular | 3 (6.8) | 10 (13.3) | 17 (16.3) | 3 (6.5) | 11 (13.4) | 44 (12.5) | |
| Both | 1 (2.3) | 8 (10.7) | 42 (40.4) | 17 (37.0) | 24 (29.3) | 92 (26.2) | |
| Random | 0 (0.0) | 0 (0.0) | 1 (1.0) | 1 (2.2) | 1 (1.2) | 3 (0.9) | |
| Conglomerate mass | 0 (0.0) | 0 (0.0) | 8 (7.7) | 2 (4.3) | 1 (1.2) | 11 (3.1) | .02 |
| Parenchymal opacity and airway and vascular distortiona | |||||||
| Ground-glass opacities | 4 (9.1) | 14 (18.7) | 36 (34.6) | 15 (32.6) | 55 (67.1) | 124 (35.3) | < .01 |
| Honeycombing | 1 (2.3) | 3 (4.0) | 2 (1.9) | 2 (4.3) | 19 (23.2) | 27 (7.7) | < .01 |
| Reticular abnormality | 5 (11.4) | 4 (5.3) | 25 (24.0) | 14 (30.4) | 48 (58.5) | 96 (27.4) | < .01 |
| Consolidation | 0 (0.0) | 7 (9.3) | 25 (24.0) | 7 (15.2) | 46 (56.1) | 85 (24.2) | < .01 |
| Mosaic attenuation | 8 (18.2) | 17 (22.7) | 19 (18.3) | 9 (19.6) | 44 (53.7) | 97 (27.6) | < .01 |
| Interlobular septal thickening | 3 (6.8) | 5 (6.7) | 19 (18.3) | 6 (13.0) | 27 (32.9) | 60 (17.1) | < .01 |
| Traction bronchiectasis | 1 (2.3) | 10 (13.3) | 37 (35.6) | 19 (41.3) | 74 (90.2) | 141 (40.2) | < .01 |
| Cystic changes | 3 (6.8) | 1 (1.3) | 2 (1.9) | 0 (0.0) | 2 (2.4) | 8 (2.3) | .31 |
| BVB distortion | 5 (11.4) | 29 (38.7) | 79 (76.0) | 33 (71.7) | 78 (95.1) | 224 (63.8) | < .01 |
| Distribution and pattern of distortion | |||||||
| Distribution cranial caudala | < .01 | ||||||
| None | 42 (95.5) | 63 (84.0) | 54 (51.9) | 22 (47.8) | 32 (39.0) | 213 (60.7) | |
| Upper | 1 (2.3) | 12 (16.0) | 48 (46.2) | 23 (50.0) | 49 (59.8) | 133 (37.9) | |
| Lower | 1 (2.3) | 0 (0.0) | 2 (1.9) | 1 (2.2) | 1 (1.2) | 5 (1.4) | |
| Distribution axiala | < .01 | ||||||
| None | 42 (95.5) | 63 (84.0) | 65 (62.5) | 24 (52.2) | 45 (54.9) | 239 (68.1) | |
| Central | 1 (2.3) | 12 (16.0) | 37 (35.6) | 20 (43.5) | 33 (40.2) | 103 (29.3) | |
| Peripheral | 1 (2.3) | 0 (0.0) | 2 (1.9) | 2 (4.3) | 4 (4.9) | 9 (2.6) | |
| Oberstein components | |||||||
| Oberstein bronchovascular bundle component (BVB) | < .01 | ||||||
| 0 | 39 (88.6) | 41 (54.7) | 19 (18.3) | 11 (23.9) | 2 (2.4) | 112 (31.9) | |
| 1 | 3 (6.8) | 25 (33.3) | 32 (30.8) | 16 (34.8) | 13 (15.9) | 89 (25.4) | |
| 2 | 0 (0.0) | 9 (12.0) | 27 (26.0) | 16 (34.8) | 27 (32.9) | 79 (22.5) | |
| 3 | 2 (4.5) | 0 (0.0) | 26 (25.0) | 3 (6.5) | 40 (48.8) | 71 (20.2) | |
| Oberstein parenchymal consolidation component (PC) | < .01 | ||||||
| 0 | 37 (84.1) | 58 (77.3) | 40 (38.5) | 29 (63.0) | 8 (9.8) | 172 (49.0) | |
| 1 | 3 (6.8) | 12 (16.0) | 28 (26.9) | 9 (19.6) | 20 (24.4) | 72 (20.5) | |
| 2 | 3 (6.8) | 4 (5.3) | 23 (22.1) | 8 (17.4) | 31 (37.8) | 69 (19.7) | |
| 3 | 1 (2.3) | 1 (1.3) | 13 (12.5) | 0 (0.0) | 23 (28.0) | 38 (10.8) | |
| Oberstein intraparenchymal nodules component (ND) | < .01 | ||||||
| 0 | 35 (79.5) | 46 (61.3) | 31 (29.8) | 22 (47.8) | 27 (32.9) | 161 (45.9) | |
| 1 | 8 (18.2) | 29 (38.7) | 38 (36.5) | 14 (30.4) | 37 (45.1) | 126 (35.9) | |
| 2 | 0 (0.0) | 0 (0.0) | 21 (20.2) | 5 (10.9) | 13 (15.9) | 39 (11.1) | |
| 3 | 1 (2.3) | 0 (0.0) | 14 (13.5) | 5 (10.9) | 5 (6.1) | 25 (7.1) | |
| Oberstein septal and nonseptal lines component (LS) | < .01 | ||||||
| 0 | 40 (90.9) | 68 (90.7) | 74 (71.2) | 34 (73.9) | 41 (50.0) | 257 (73.2) | |
| 1 | 3 (6.8) | 6 (8.0) | 30 (28.8) | 12 (26.1) | 40 (48.8) | 91 (25.9) | |
| 2 | 1 (2.3) | 1 (1.3) | 0 (0.0) | 0 (0.0) | 1 (1.2) | 3 (0.9) | |
| Oberstein pleural thickening component (PLT) | .69 | ||||||
| 0 | 44 (100.0) | 74 (98.7) | 102 (98.1) | 46 (100.0) | 78 (95.1) | 344 (98.0) | |
| 1 | 0 (0.0) | 1 (1.3) | 2 (1.9) | 0 (0.0) | 2 (2.4) | 5 (1.4) | |
| 2 | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 2 (2.4) | 2 (0.6) | |
| Oberstein lymph node enlargement component (LN) | < .01 | ||||||
| 0 | 29 (65.9) | 25 (33.3) | 27 (26.0) | 32 (69.6) | 31 (37.8) | 144 (41.0) | |
| 1 | 12 (27.3) | 26 (34.7) | 28 (26.9) | 9 (19.6) | 23 (28.0) | 98 (27.9) | |
| 2 | 3 (6.8) | 23 (30.7) | 47 (45.2) | 5 (10.9) | 27 (32.9) | 105 (29.9) | |
| 3 | 0 (0.0) | 1 (1.3) | 2 (1.9) | 0 (0.0) | 1 (1.2) | 4 (1.1) | |
| Other findings | |||||||
| Emphysema | 0 (0.0) | 1 (1.3) | 2 (1.9) | 2 (4.3) | 4 (4.9) | 9 (2.6) | .44 |
| Pulmonary artery enlargement | 0 (0.0) | 0 (0.0) | 2 (1.9) | 2 (4.3) | 6 (7.3) | 10 (2.8) | .04 |
| Prior thoracic surgery | 1 (2.3) | 2 (2.7) | 1 (1.0) | 1 (2.2) | 0 (0.0) | 5 (1.4) | .44 |
Data are presented as No. (%) or mean ± SD unless otherwise indicated. No patients were reported as having tree-in-bud or cavitary lesions. BVB = bronchovascular bundle.
For assessments of distribution, a marking of none is consistent both with a presence of abnormalities that lacked one of the specific distributional patterns listed and with an absence of abnormalities.
Table 3.
Distribution of Chest Radiography and CT Scan Inconsistencies by Scadding Stage
| Variable | Scadding Stage |
Total (n = 351) | ||||
|---|---|---|---|---|---|---|
| 0 (n = 44) | I (n = 75) | II (n = 104) | III (n = 46) | IV (n = 82) | ||
| Consistent | 45.5 | 30.7 | 61.5 | 54.3 | 100.0 | 61.0 |
| Inconsistent lymphadenopathy and PA | 11.4 | 13.3 | 8.7 | 4.3 | 0.0 | 7.4 |
| Inconsistent lymphadenopathy only | 22.7 | 21.3 | 18.3 | 19.6 | 0.0 | 15.4 |
| Inconsistent PA only | 20.5 | 34.7 | 11.5 | 21.7 | 0.0 | 16.2 |
Data are presented as percentage. Details of CT scan-based determination of inconsistencies are provided in e-Table 4. PA = parenchymal abnormality.
CT Scan Findings Inform Lung Function
Because results for PFT before and after bronchodilator administration were similar when comparing CT imaging with Scadding stage, only analyses for PFT obtained after bronchodilator administration are reported for brevity. Nearly all individual CT scan measures were associated negatively with PFT measures (Fig 1, red circles), indicating that presence of a CT scan abnormality was associated with worse lung function. Adjustment for Scadding stage attenuated associations such that some CT scan measures were no longer associated with lung function. For Dlco (Fig 1, first column), Oberstein parenchymal consolidation (PC) and presence of mosaic attenuation, honeycombing, ground-glass opacities, and emphysema each remained associated with decrease in Dlco after adjustment for Scadding stage (red triangles). For FEV1 (Fig 1, second column), traction bronchiectasis, bronchovascular bundle (BVB) distortion, consolidation, mosaic attenuation, and Oberstein subscales for BVB and PC remained associated with CT scan findings after adjustment for Scadding stage. Unlike Dlco, honeycombing and ground-glass opacities no longer were associated with FEV1 when including Scadding stage. After adjustment for Scadding stage (Fig 2, third column), many associations between CT scan measures and FVC became insignificant. Interestingly, the association between FEV1 to FVC ratio and traction bronchiectasis, emphysema, consolidation, BVB distortion, and axial distribution of distortion remained significant after adjustment for Scadding stage (Fig 1, fourth column). Adjustment for Scadding stage reduced the effect of Obserstein PC by 8% to 25% for Dlco and FVC and attenuated FEV1 by 35%. Scadding stage showed a smaller confounding effect on emphysema and Dlco and FEV1, reducing the effect by 10% to 13% (e-Tables 5-8).
Figure 1.
Graphs showing estimated associations between a single CT scan measure and each pulmonary function test variable. Coefficients and 95% CIs are presented. The dashed black line demarks a coefficient estimate of 0. Circles are estimates from models with the base factors plus the CT scan measure. Triangles are models that also include Scadding stage. Red is a P value of < .01 and blue is a P value of ≥ .01. Only models with a significant effect in at least one model across different pulmonary function tests are presented. Overall Oberstein score is excluded because it was the only CT scan measure not treated as categorical. BVB = bronchovascular bundle; Dlco = diffusing capacity of the lungs for carbon monoxide; PC = parenchymal consolidation.
Figure 2.
Graphs showing partial F-test P values assessing significance of Scadding stage before (red line) and after (blue circles) inclusion of an individual CT scan measure in a linear model predicting a single PFT measure. All models included age, race/ethnicity, sex, height, and BMI. P values are plotted on a negative log base 10 scale. The dashed line indicates the threshold for significance, P = .01, with points to the right considered significant with respect to that threshold. BVB = bronchovascular bundle; Dlco = diffusing capacity of the lungs for carbon monoxide; LA = lymphadenopathy; LS = septal and nonseptal lines; ND = nodule; PC = parenchymal consolidation. See e-Table 1 for expansions of other Oberstein component definitions.
Next, we evaluated Scadding stage and PFT results and then included specific CT scan features (Fig 2). Scadding stage was significantly associated with FEV1 to a greater degree than with Dlco, FVC, and FEV1 to FVC ratio. Although no single CT scan measure equated to Scadding stage, reductions in significance between Scadding stage and each lung function measure were noted for many CT scan features. Scadding was no longer associated with FVC when including traction bronchiectasis and Oberstein PC and with FEV1 to FVC ratio when including traction bronchiectasis and Oberstein BVB. When adjusting for many CT scan PAs, such as reticular abnormalities, mosaic attenuation, ground-glass opacities, consolidation, and Oberstein PC, the association between Scadding stage and lung function was attenuated. Notable exceptions included nodules, lymphadenopathy, interseptal thickening, and some Oberstein components.
Associations Among CT Scan Measures
Next, we assessed associations among various CT scan measures and Scadding stage (Fig 3). The strongest associations among CT scan VAS and Scadding stage were found between Oberstein components and VAS measures that directly assess the same feature. PC, BVB distortion, traction bronchiectasis, and Scadding stage formed a block of largely associated CT scan measures (Fig 3, upper right corner). More moderate associations were found among these measures, micronodules, and PAs. Smaller associations were observed between measures of lymphadenopathy and micronodules and other associated measures. Ground-glass opacities demonstrated a strong association with Oberstein PC and mosaic association, but only a small association with direct assessment of consolidation. Pulmonary artery enlargement and emphysema were associated moderately with each other and, with pleural thickening and conglomerate masses, mostly were unassociated with other measures.
Figure 3.
Heatmap showing strength of associations among CT scan measures and Scadding stage as measured by Cramér’s V with bias correction. BVB = bronchovascular bundle; LA = lymphadenopathy; LS = septal and nonseptal lines; ND = nodule; PC = parenchymal consolidation. See e-Table 1 for expansions of other Oberstein component definitions.
Using multivariable regression modeling including all CT scan VAS measures with backward selection to identify CT scan measures explaining PFT, Scadding stage remained in final models for FEV1 and FEV1 to FVC ratio (P < .05) and marginally for Dlco (P < .15) (Fig 4, bottom panel, e-Table 9). This finding was notable only for stage 3 or 4 disease compared with stage 0 disease. A collection of CT scan measures (differing for each PFT outcome) mostly accounted for Scadding stage for Dlco and FVC. Only mosaic attenuation and Oberstein PC were associated with all PFT measures, whereas honeycombing, cystic changes, distribution axial, and emphysema were selected for three of four PFT measures. Six measures were not included in any model: mediastinal and hilar lymphadenopathy, ground-glass opacities, and Oberstein BVB, nodule, and lymphadenopathy. Interestingly, five CT scan measures were unassociated with any PFT measure in the individual analysis, and yet were associated in final multivariable models: micronodule distribution and conglomerate nodules, cystic changes, Oberstein pleural thickening, and pulmonary artery enlargement.
Figure 4.
Graphs showing estimated associations in backward-selected models between each PFT measure and included CT scan measures. Coefficients and 95% CIs are presented. The dashed black line demarks a coefficient estimate of 0. Red indicates significance at the level P < .01, blue indicates significance at the level P < .05, and light gray indicates significance at the level P < .15. BVB = bronchovascular bundle; Dlco = diffusing capacity of the lungs for carbon monoxide; LS = septal and nonseptal lines; PC = parenchymal consolidation.
CT Scan Reader Concordance
The above analyses were conducted using one reader. When comparing reader-to-reader concordance, mixed results were noted (Table 4, e-Table 10), with κ values ranging from 0.15 (poor agreement) to 0.64 (moderate agreement). Better agreement was noted for the presence or absence of features vs agreement in location or Oberstein severity (eg, lymphadenopathy, micronodule distribution, and Oberstein components). Honeycombing, cystic changes, emphysema and subtypes, and evidence of thoracic surgery were found to be rare by both readers, although they disagreed on the presence of a condition in an individual (e-Table 10).
Table 4.
Concordance Estimates and 95% CIs of Duplicate VAS Reads for 168 Individuals
| Variable | Cohen’s κ Value (95% CI) |
|---|---|
| Mediastinal lymphadenopathy | |
| Location | 0.58 (0.47-0.7) |
| Presence | 0.62 (0.5-0.74) |
| Hilar lymphadenopathy | |
| Location | 0.54 (0.41-0.66) |
| Presence | 0.55 (0.42-0.68) |
| Micronodule distribution | |
| Location | 0.15 (0.04-0.26) |
| Presence | 0.25 (0.1-0.39) |
| BVB distortion | 0.46 (0.32-0.59) |
| Traction bronchiectasis | 0.4 (0.25-0.54) |
| Ground-glass opacities | 0.48 (0.34-0.62) |
| Reticular abnormality | 0.28 (0.13-0.43) |
| Consolidation | 0.33 (0.13-0.53) |
| Mosaic attenuation | 0.27 (0.11-0.43) |
| Oberstein BVB | |
| Severity | 0.2 (0.1-0.3) |
| Presence | 0.56 (0.42-0.69) |
| Oberstein PC | |
| Severity | 0.26 (0.14-0.39) |
| Presence | 0.42 (0.28-0.56) |
| Oberstein nodule | |
| Severity | 0.3 (0.19-0.42) |
| Presence | 0.33 (0.18-0.47) |
| Oberstein LS | |
| Severity | 0.19 (0.05-0.34) |
| Presence | 0.24 (0.09-0.38) |
| Oberstein lymphadenopathy | |
| Severity | 0.5 (0.39-0.61) |
| Presence | 0.64 (0.52-0.76) |
Estimates are for concordance of presence-or-absence assessments unless otherwise indicated. BVB = bronchovascular bundle distortion; LS = septal and nonseptal lines; PC = parenchymal consolidation; VAS = visual assessment score.
Discussion
Leveraging the results of the multicenter GRADS study, we were able to compare CXR-based Scadding stage with chest HRCT imaging features visually scored by thoracic radiologists with lung function. We found CT scan scoring to be inconsistent with Scadding stage designation in almost 40% of patients. We found CXR-based Scadding stage to be associated with spirometry and Dlco data, although chest HRCT imaging was able to explain additional variability in PFT data even when Scadding stage was included, suggesting that additional information is derived from some CT scan features (eg, presence of mosaic attenuation or degree of Oberstein PC). Some chest HRCT imaging features were associated with PFT measures even in the presence of Scadding stage (eg, Dlco with ground-glass opacities and honeycombing and FEV1 with BVB distortion and consolidation). Chest HRCT imaging features were associated with each other and with specific Scadding stages, suggesting potential "CT scan phenotypes." This is supported further by the finding of associations between CT features and Oberstein scores. Interreader variability in CT scans likely limits its use, as at best moderate agreement between presence or absence of CT scan features and at worst poor agreement in reference to location or severity was noted. Regardless, we present new information regarding the usefulness of chest CT imaging, revealing associations between PFT abnormalities and CT scan features in pulmonary sarcoidosis. Overall, these results suggest that CT imaging may provide additional characterization of lung abnormalities with implications for PFT to a greater degree than Scadding staging.
Scadding stage has been commonly used to categorize patients with pulmonary sarcoidosis. In contrast to other interstitial lung diseases that rely on chest HRCT imaging, CXR has been used to assess sarcoidosis disease status.7 Over decades, experience with Scadding staging led to its use in prognosis, although with many caveats. However, it is widely accepted that chest HRCT imaging provides additional detail lacking on review of CXR, as supported by prior smaller studies in which chest CT imaging better detected lymphadenopathy and PA.28,29 These studies did not seek to compare findings between Scadding stage and chest HRCT imaging directly, as we did.
Evaluating specific CT scan features, we found significant variability across all stages (Table 2), including the presence of nodules, PA, BVB distortion, and traction bronchiectasis, with many of these present in Scadding stage 0 and I disease. Using chest HRCT imaging to define Scadding-like score in our study, we identified many patients who were categorized inconsistently relative to Scadding stage (Table 3). These inconsistencies were seen widely across disease stages 0 through III, with inconsistencies based on lymphadenopathy, PA, or both, with PA being the most discrepant. Inconsistencies related to PA seen on CXR but not seen on CT imaging were surprisingly frequent. A smaller recent study compared chest HRCT imaging features with Scadding stage and demonstrated similar results to ours, with discordance between CT scan and Scadding stage in 50% of patients, primarily because of differences in PA.7,30 Our study was unable to address inconsistencies related to fibrosis and may have overrepresented consistencies when comparing CXR with CT imaging; the recent study reported high concordance with stage IV disease and chest HRCT imaging.
Previous studies have suggested that Scadding stage is a poor predictor of lung function,10, 11, 12, 13 whereas chest HRCT imaging PAs have correlated with PFT results.20 Similarly, we found CT scan features to be widely associated with worse lung function (Fig 1), with adjustment for Scadding stage attenuating some associations (Fig 1, blue triangles). However, Scadding stage was associated with spirometry and Dlco values (Fig 2, red line), and significance decreased when various CT scan measures were included (Fig 2, blue circle). Multivariable regression modeling did not include Scadding stage in the final model for FVC and included Scadding stage only marginally in the final model for Dlco, with a collection of CT scan measures providing approximately the same information as Scadding stage for Dlco and FVC (Fig 4). To our knowledge, no other study has evaluated the association between CXR and CT imaging results independently and in combination as we have, supporting that Scadding staging demonstrates some association with lung function, although a more pronounced relationship is noted with CT scan findings. A limitation of our study is that it is cross-sectional. A prior study demonstrated that serial CT scan changes moderately agreed with spirometry changes and that only fair agreement was noted with CXR changes.30 In contrast to our study, CXR-based findings did not strengthen the association between PFT results and CT scan-based findings, although sample size was relatively small (n = 73) and prognosis was the primary focus. These data support that additional information regarding lung function in patients with sarcoidosis is gained from CT imaging over Scadding stage, suggesting a potential role for chest HRCT imaging to guide clinical management. It will be interesting to determine if similar disease activity and response to treatment information can be gained from initial and serial chest HRCT imaging as in the case of PET studies, which are used to evaluate the intensity and extent of inflammatory activity of sarcoidosis throughout the body and have correlated well with both chest HRCT imaging and PFT findings.31, 32, 33, 34, 35, 36, 37 Future studies will be needed to clarify these relationships further, especially longitudinally.
Efforts to group radiographic parenchymal findings into categories that impart physiologic information have been attempted, and scoring systems exist that analyze numerous chest HRCT imaging components used in our study, including Oberstein score.19,21 These approaches rely on quantitative categories and qualitative assessments of degree of an abnormality, thus making agreement between scorers challenging.20 Although initial studies were limited by size (N = 2121 and N = 9519), correlation between CT scan findings and degree of disease inflammation was seen,21 with limited prognostic assessment.19 We found associations with Oberstein score components and PFT supporting this approach. We also noted similar PFT results associations across numerous CT scan features (Figure 1, Figure 2) and groups of CT scan features highly associated with each other (Fig 3). The Oberstein BVB score in our study was associated with CT scan BVB distortion, consolidation and traction bronchiectasis, and Scadding stage, supporting that it may indicate a sarcoidosis phenotype (Fig 3). In addition, BVB distortion and consolidation were associated with ground-glass opacities and reticulation CT scan features, suggesting fibrotic features.
It is possible that CT scan phenotypes may be defined by a limited number of key CT imaging features,32 with correlation between features. To date, limited CT imaging phenotypes have been identified and determined to be functionally relevant besides the fibrotic phenotype and its subtypes. In fibrotic sarcoidosis, three CT scan patterns have been proposed with associated lung function abnormalities: predominant BVB distortion associated with obstruction, honeycombing with restriction and low Dlco, and linear PA with relatively minor impairment.14,33 Similarly, we found BVB distortion to be associated with low FEV1 and obstruction and interlobular septal thickening to be unassociated with PFT findings. Extent of lung fibrosis on CT imaging is associated with mortality34 and prognosis.32 The GRADS study aimed to define gene expression phenotypes/endotypes and found genes associated with progressive Scadding stage and CT scan features, with overlap with PFT findings.23 Reticular abnormalities are associated with > 750 genes and with genes associated with traction bronchiectasis, ground-glass opacities, progressive Scadding stage, Dlco, and spirometry to varying degrees. CT scan features were associated with three of four gene modules and Scadding stage was associated with only one gene module where CT scan adenopathy also was associated with the module. These data support that overlap between CT scan features, Scadding stage, and PFT results occurs and raise a question regarding whether sarcoidosis phenotypes/endotypes using multiple data sources, both clinical and omics, might provide clinically relevant information for evaluation or treatment of patients with sarcoidosis.
Our study is limited by interobserver variability, as is any similar study using VAS. The GRADS study was set up with a scoring system to address two different diseases using one radiologist evaluating all CT scans,22,23 as have other studies.32 A subset of CT scans were reviewed with a limited VAS. Based on these data, we found mixed concordance across different CT scan measures, although agreement was optimum for the presence or absence of a feature vs location or severity, similar to other studies.20 Even with CXR, only fair agreement is noted among chest radiologists (eg, fibrosis and lymphadenopathy presence)38 and when evaluating change in Scadding stage over time.39 This issue confounds radiographic studies using VAS, and thus efforts to improve reader agreement or other ways to assess radiographic abnormalities are needed.40 Our study was secondary to the primary outcomes for the GRADS study,22 aiming to investigate lung microbiome and genome in sarcoidosis and α1-antitrypsin deficiency. Strengths include the size of the population, larger than most if not all compiled to date; the presence of standardized data and CT scans and VAS from almost all participants; as well as the multicenter nature of our study.
Interpretation
We evaluated a large population of patients with sarcoidosis and demonstrated that CT imaging and Scadding stage provide different and overlapping information related to lung function. We found that while Scadding stage is limited in its assessment, it associates with PFT results. This study suggests that CT imaging provides additional information, and in FVC and to a lesser degree Dlco, CT imaging alone obviates information provided by Scadding stage. Correlation between CT scan features raises a question about CT scan phenotypes and endotypes. Investigation into pulmonary phenotypes or endotypes that include clinical and omics data in larger populations and investigating changes over time and prognosis are next steps to be considered to advance the care of patients with sarcoidosis.
Funding/Support
This work was supported by the National Institutes of Health [Grants U01 HL112707, U01 HL112694, U01 HL112695, U01 HL112696, U01 HL112702, U01 HL112708, U01 HL112711, U01 HL112712, UL1 RR029882, UL1 RR025780, R01HL110883, R01 HL114587, R01 HL127349, U01 HL137159, CTSI U54 grant 9 UL1 TR000005, CDC NMVB 5U24 OH009077, CTSA UL1 TR002535, R01HL142049, and 5T32HL007085].
Financial/Nonfinancial Disclosures
The authors have reported to CHEST the following: L. A. M. has received grants from the National Institutes of Health [Grants R01HL140357, R01HL142049, R01HL136681], Ann Theodore Foundation, the Foundation for Sarcoidosis Research, Mallinckrodt Pharmaceuticals, and the University of Cincinnati [Mallinckrodt Pharmaceuticals Foundation Grant] and serves on the Scientific Advisory Board for the Foundation for Sarcoidosis Research and Boehringer Ingelheim. B. S. B. has received a grant from Mallinckrodt Pharmaceuticals. E. S. C. has received grants from the National Institutes of Health, the Foundation for Sarcoidosis Research, American Thoracic Society, serves on the Scientific Advisory Board for the Foundation for Sarcoidosis Research, and has participated in clinical trials sponsored by aTyr Pharma and LabCorp Drug. L. L. K. has received grants from the National Institutes of Health [Grant R01HL157533], An Theodore Foundation, and Mallinckrodt Pharmaceuticals. D. A. L. was coinvestigator on a grant from the National Institutes of Health [Grant R01HL142049] and served on a Scientific Advisory Board for Boehringer Ingelheim, Inc. N. K. is a scientific founder at Thyron; served as a consultant to Boehringer Ingelheim, Pliant, Astra Zeneca, RohBar, Veracyte, Augmanity, CSL Behring, Splisense, Galapagos, Fibrogen, GSK, Merck, and Thyron over the last 3 years; reports equity in Pliant and Thyron; reports grants from Veracyte, Boehringer Ingelheim, and BMS; and reports nonfinancial support from Astra Zeneca. None declared (W. L. L., I. C., G. K. B., E. M. B., W. P. D., E. H., K. G., S. R. W., C. F.).
Acknowledgments
Author contributions: L. A. M., W. L. L., N. E. C., L. L. K., W. P. D., and E. H. contributed to the conception and design of the study. B. S. B., I. C., E. M. B., W. P. D., E. H., K. G., E. S. C., L. L. K., C. F., D. A. L., N. K., S. R. W., and L. A. M. contributed to the acquisition of the data. W. L. L., G. K. B., and N. E. C. ensured accuracy and completeness of the statistical analysis. B. S. B., W. L. L., N. E. C., and L. A. M. drafted and revised the manuscript. All authors reviewed, revised, and approved the manuscript before submission. Dr Fuhrman is deceased.
Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.
Additional information: The e-Appendix and e-Tables are available online under “Supplementary Data.”
Footnotes
Drs Benn and Lippitt contributed equally to this manuscript as first authors.
Drs Carlson and Maier contributed equally to this manuscript as senior authors.
Supplementary Data
e-Figure.
References
- 1.Drent M., Crouser E.D., Grunewald J. Challenges of Sarcoidosis and Its Management. N Engl J Med. 2021;385(11):1018–1032. doi: 10.1056/NEJMra2101555. [DOI] [PubMed] [Google Scholar]
- 2.Crouser E.D., Maier L.A., Wilson K.C., et al. Diagnosis and detection of sarcoidosis. An official American Thoracic Society clinical practice guideline. Am J Respir Crit Care Med. 2020;201(8):e26–e51. doi: 10.1164/rccm.202002-0251ST. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Baughman R.P., Teirstein A.S., Judson M.A., et al. Clinical characteristics of patients in a case control study of sarcoidosis. Am J Respir Crit Care Med. 2001;164(10 pt 1):1885–1889. doi: 10.1164/ajrccm.164.10.2104046. [DOI] [PubMed] [Google Scholar]
- 4.Scadding J.G. Prognosis of intrathoracic sarcoidosis in England. A review of 136 cases after five years’ observation. Br Med J. 1961;2(5261):1165–1172. doi: 10.1136/bmj.2.5261.1165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Levy A., Hamzeh N., Maier L.A. Is it time to scrap Scadding and adopt computed tomography for initial evaluation of sarcoidosis? F1000Research. 2018;7 doi: 10.12688/f1000research.11068.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Culver D.A., Baughman R.P. It’s time to evolve from Scadding: phenotyping sarcoidosis. Eur Respir J. 2018;51(1):1800050. doi: 10.1183/13993003.00050-2018. [DOI] [PubMed] [Google Scholar]
- 7.Zhang Y., Du S.S., Zhao M.M., et al. Chest high-resolution computed tomography can make higher accurate stages for thoracic sarcoidosis than X-ray. BMC Pulm Med. 2022;22(1):146. doi: 10.1186/s12890-022-01942-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lin N.W., Arbet J., Mroz M.M., et al. Clinical phenotyping in sarcoidosis using cluster analysis. Respir Res. 2022;23(1):88. doi: 10.1186/s12931-022-01993-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bonella F., Culver D.A., Israël-Biet D. Vol. 96. European Respiratory Society; 2022. Sarcoidosis. [Google Scholar]
- 10.Harrison B.D., Shaylor J.M., Stokes T.C., Wilkes A.R. Airflow limitation in sarcoidosis—a study of pulmonary function in 107 patients with newly diagnosed disease. Respir Med. 1991;85(1):59–64. doi: 10.1016/s0954-6111(06)80211-8. [DOI] [PubMed] [Google Scholar]
- 11.Karetzky M., McDonough M. Exercise and resting pulmonary function in sarcoidosis. Sarcoidosis Vasc Diffuse Lung Dis. 1996;13(1):43–49. [PubMed] [Google Scholar]
- 12.Barros W.G., Neder J.A., Pereira C.A., Nery L.E. Clinical, radiographic and functional predictors of pulmonary gas exchange impairment at moderate exercise in patients with sarcoidosis. Respiration. 2004;71(4):367–373. doi: 10.1159/000079641. [DOI] [PubMed] [Google Scholar]
- 13.Yeager H., Rossman M.D., Baughman R.P., et al. Pulmonary and psychosocial findings at enrollment in the ACCESS study. Sarcoidosis Vasc Diffuse Lung Dis. 2005;22(2):147–153. [PubMed] [Google Scholar]
- 14.Nunes H., Uzunhan Y., Gille T., Lamberto C., Valeyre D., Brillet P.Y. Imaging of sarcoidosis of the airways and lung parenchyma and correlation with lung function. Eur Respir J. 2012;40(3):750–765. doi: 10.1183/09031936.00025212. [DOI] [PubMed] [Google Scholar]
- 15.Daniloff E.M., Lynch D.A., Bartelson B.B., Newell J.D., Jr., Bernstein S.M., Newman L.S. Observer variation and relationship of computed tomography to severity of beryllium disease. Am J Respir Crit Care Med. 1997;155(6):2047–2056. doi: 10.1164/ajrccm.155.6.9196114. [DOI] [PubMed] [Google Scholar]
- 16.Murdoch J., Müller N.L. Pulmonary sarcoidosis: changes on follow-up CT examination. AJR Am J Roentgenol. 1992;159(3):473–477. doi: 10.2214/ajr.159.3.1503008. [DOI] [PubMed] [Google Scholar]
- 17.Brauner M.W., Lenoir S., Grenier P., Cluzel P., Battesti J.P., Valeyre D. Pulmonary sarcoidosis: CT assessment of lesion reversibility. Radiology. 1992;182(2):349–354. doi: 10.1148/radiology.182.2.1732948. [DOI] [PubMed] [Google Scholar]
- 18.Akira M., Kozuka T., Inoue Y., Sakatani M. Long-term follow-up CT scan evaluation in patients with pulmonary sarcoidosis. Chest. 2005;127(1):185–191. doi: 10.1378/chest.127.1.185. [DOI] [PubMed] [Google Scholar]
- 19.Remy-Jardin M., Giraud F., Remy J., Wattinne L., Wallaert B., Duhamel A. Pulmonary sarcoidosis: role of CT in the evaluation of disease activity and functional impairment and in prognosis assessment. Radiology. 1994;191(3):675–680. doi: 10.1148/radiology.191.3.8184045. [DOI] [PubMed] [Google Scholar]
- 20.Drent M., De Vries J., Lenters M., et al. Sarcoidosis: assessment of disease severity using HRCT. Eur Radiol. 2003;13(11):2462–2471. doi: 10.1007/s00330-003-1965-x. [DOI] [PubMed] [Google Scholar]
- 21.Oberstein A., von Zitzewitz H., Schweden F., Müller-Quernheim J. Non invasive evaluation of the inflammatory activity in sarcoidosis with high-resolution computed tomography. Sarcoidosis Vasc Diffuse Lung Dis. 1997;14(1):65–72. [PubMed] [Google Scholar]
- 22.Moller D.R., Koth L.L., Maier L.A., et al. Rationale and design of the Genomic Research in Alpha-1 Antitrypsin Deficiency and Sarcoidosis (GRADS) Study. Sarcoidosis protocol. Ann Am Thorac Soc. 2015;12(10):1561–1571. doi: 10.1513/AnnalsATS.201503-172OT. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Vukmirovic M., Yan X., Gibson K.F., et al. Transcriptomics of bronchoalveolar lavage cells identifies new molecular endotypes of sarcoidosis. Eur Respir J. 2021;58(6):2002950. doi: 10.1183/13993003.02950-2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Regan E.A., Hokanson J.E., Murphy J.R., et al. Genetic epidemiology of COPD (COPDGene) study design. COPD. 2010;7(1):32–43. doi: 10.3109/15412550903499522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Miller M.R., Crapo R., Hankinson J., et al. General considerations for lung function testing. Eur Respir J. 2005;26(1):153–161. doi: 10.1183/09031936.05.00034505. [DOI] [PubMed] [Google Scholar]
- 26.Graham B.L., Brusasco V., Burgos F., et al. 2017 ERS/ATS standards for single-breath carbon monoxide uptake in the lung. Eur Respir J. 2017;49(1):1600016. doi: 10.1183/13993003.00016-2016. [DOI] [PubMed] [Google Scholar]
- 27.Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Lawrence Erlbaum Associates, Inc.; 1988. Set correlation and multivariate methods; pp. 467–530. [Google Scholar]
- 28.Aleksonienė R., Zeleckienė I., Matačiūnas M., et al. Relationship between radiologic patterns, pulmonary function values and bronchoalveolar lavage fluid cells in newly diagnosed sarcoidosis. J Thorac Dis. 2017;9(1):88–95. doi: 10.21037/jtd.2017.01.17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ors F., Gumus S., Aydogan M., Sari S., Verim S., Deniz O. HRCT findings of pulmonary sarcoidosis; relation to pulmonary function tests. Multidiscip Respir Med. 2013;8(1):8. doi: 10.1186/2049-6958-8-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zappala C.J., Desai S.R., Copley S.J., et al. Accuracy of individual variables in the monitoring of long-term change in pulmonary sarcoidosis as judged by serial high-resolution CT scan data. Chest. 2014;145(1):101–107. doi: 10.1378/chest.12-2479. [DOI] [PubMed] [Google Scholar]
- 31.Lewis P.J., Salama A. Uptake of fluorine-18-fluorodeoxyglucose in sarcoidosis. J Nucl Med. 1994;35(10):1647–1649. [PubMed] [Google Scholar]
- 32.Walsh S.L., Wells A.U., Sverzellati N., et al. An integrated clinicoradiological staging system for pulmonary sarcoidosis: a case-cohort study. Lancet Respir Med. 2014;2(2):123–130. doi: 10.1016/S2213-2600(13)70276-5. [DOI] [PubMed] [Google Scholar]
- 33.Abehsera M., Valeyre D., Grenier P., Jaillet H., Battesti J.P., Brauner M.W. Sarcoidosis with pulmonary fibrosis: CT patterns and correlation with pulmonary function. AJR Am J Roentgenol. 2000;174(6):1751–1757. doi: 10.2214/ajr.174.6.1741751. [DOI] [PubMed] [Google Scholar]
- 34.Jeny F., Uzunhan Y., Lacroix M., et al. Predictors of mortality in fibrosing pulmonary sarcoidosis. Respir Med. 2020;169 doi: 10.1016/j.rmed.2020.105997. [DOI] [PubMed] [Google Scholar]
- 35.Mostard R.L., Verschakelen J.A., van Kroonenburgh M.J., et al. Severity of pulmonary involvement and (18)F-FDG PET activity in sarcoidosis. Respir Med. 2013;107(3):439–447. doi: 10.1016/j.rmed.2012.11.011. [DOI] [PubMed] [Google Scholar]
- 36.Cremers J.P., Van Kroonenburgh M.J., Mostard R.L., et al. Extent of disease activity assessed by 18F-FDG PET/CT in a Dutch sarcoidosis population. Sarcoidosis Vasc Diffuse Lung Dis. 2014;31(1):37–45. [PubMed] [Google Scholar]
- 37.Vender R.J., Aldahham H., Gupta R. The role of PET in the management of sarcoidosis. Curr Opin Pulm Med. 2022;28(5):485–491. doi: 10.1097/MCP.0000000000000892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Baughman R.P., Shipley R., Desai S., et al. Changes in chest roentgenogram of sarcoidosis patients during a clinical trial of infliximab therapy: comparison of different methods of evaluation. Chest. 2009;136(2):526–535. doi: 10.1378/chest.08-1876. [DOI] [PubMed] [Google Scholar]
- 39.Zappala C.J., Desai S.R., Copley S.J., et al. Optimal scoring of serial change on chest radiography in sarcoidosis. Sarcoidosis Vasc Diffuse Lung Dis. 2011;28(2):130–138. [PubMed] [Google Scholar]
- 40.Ryan S.M., Fingerlin T.E., Mroz M., et al. Radiomic measures from chest high-resolution computed tomography associated with lung function in sarcoidosis. Eur Respir J. 2019;54(2):1900371. doi: 10.1183/13993003.00371-2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.