Figures
- Figure 1–
Computed tomography images demonstrating a) normal appearance and various visual manifestations of diffuse lung disease: b) ground glass, c) reticular changes (arrows), d) honeycombing (arrows) and e) emphysema (arrow)). In training datasets, the consensus of four thoracic radiologists was used to identify multiple volumes of interest of 15×15×15 pixels corresponding to normal, ground-glass density, reticular abnormalities, honeycombing and emphysema.
- Figure 2–
Representative results from the quantitative analysis of CALIPER. The two rows (a–c) and (d–f) correspond to the two time-points from one of the patients. The individual voxels of the lung regions in the original sections (a and d) are classified and colour-coded into one of the classes of visible abnormalities (b and e). c and f) Three-dimensional maximum feature projections. CALIPER quantification and radiologist reviews were both indicative of an overall increase in ground-glass opacities, reticular abnormalities and honeycombing.
Tables
- Table 1– Demographic data#
Age years 72.4±6.9 Females/males 28 (51)/27 (49) Height cm 168.0±9.4 Smoking history Active 1 (2) Previous 34 (62) Never 20 (36) Transplanted 2 (4) Surgical lung biopsy confirmation 31 (56) Mean interval between HRCT scans days 289.2±109.7 (55–540) Median follow-up years (range) 2.4 (0.1–8.5) Deaths 37 (67) Scan type 110 HRCT scans GE 83 75 Siemens 27 25 Data are presented as mean±sd, n (%), mean±sd (range) or n, unless otherwise stated. HRCT: high-resolution computed tomography. #: total subjects n=55.
- Table 2– Raw and % predicted pulmonary function test values corresponding with computed tomography scans at time-points 1 and 2
Parameter Time-point 1 Time-point 2 Difference±sd TLC % pred 67.3±9.6 65.9±9.7 -3.4±5.6 TLC L 3.9±0.8 3.9±0.8 -0.2±0.3 FVC % pred 67.5±14.1 64.1±16.8 -3.4±8.2 FVC L 2.4±0.8 2.3±0.8 -0.1±0.3 FEV1 % pred 72.0±14.7 68.9±16.9 -3.1±8.5 FEV1 L 2.0±0.6 1.9±0.6 -0.1±0.2 DLCO % pred 50.6±14.5 45.7±12.4 -7.0±11.6 DLCO mL·min−1·mmHg−1 11.7±4.0 10.7±3.0 -1.5±2.6 Data are presented as mean±sd. TLC: total lung capacity; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 s; DLCO: diffusing capacity of the lung for carbon monoxide.
- Table 3– Mean volumetric measurements and changes for computed tomography findings at time-points 1 and 2
Finding Time-point 1 Time-point 2 Difference±sd Total normal L 2.5±1.0 2.2±1.1 -0.3±0.6 Total emphysema L 0.02±0.04 0.01±0.03 -0.002±0.03 Total emphysema % 0.4±1.2 0.4±1.0 0.0±0.9 Total GGO L 0.3±0.3 0.4±0.4 0.08±0.3 Total GGO# % 10.2±10.6 13.3±12.8 3.2±9.8 Total reticular L 0.5±0.3 0.6±0.4 0.06±0.2 Total reticular# % 15.1±9.9 18.1±12.4 3.0±6.4 Total honeycombing L 0.2±0.3 0.2±0.3 -0.01±0.2 Total honeycombing# % 6.0±6.9 6.0±8.8 -0.1±6.3 Total ILD L 1.1±0.6 1.2±0.6 0.1±0.4 Total ILD# % 31.3±16.8 37.4±22.1 6.1±13.0 Total lung volume L 3.6±0.8 3.5±0.9 -0.2±0.4 Data are presented as mean±sd, unless otherwise stated. GGO: ground-glass opacities; ILD: interstitial lung disease (ground glass + reticular + honeycombing). #: % of total lung volume.
- Table 4– Radiologists' assessment of computed tomography findings at time-points 1 and 2
Finding Time-point 1 Time-point 2 Difference±sd Radiologist 1 Total emphysema % 1.3±4.0 1.3±3.9 0.0±0.7 Total GGO % 14.9±17.5 17.5±20.7 2.6±8.5 Total reticular % 17.6±9.6 19.9±10.9 2.3±4.6 Total honeycombing % 3.5±7.9 3.7±8.4 0.2±1.0 Radiologist 2 Total emphysema % 1.3±3.9 1.4±3.9 0.1±0.5 Total GGO % 22.2±20.3 24.8±24.4 2.6±11.0 Total reticular % 18.2±8.2 18.6±8.5 0.4±1.9 Total honeycombing % 1.7±6.2 1.7±6.3 0.1±0.5 Data are presented as mean±sd, unless otherwise stated. GGO: ground-glass opacities.
- Table 5– Survival analysis of pulmonary function tests and CALIPER measurement
Difference Univariable Adjusted HR 95% CI p-value HR 95% CI p-value TLC % pred# 1.19 0.67–2.12 0.55 4.17 1.42–12.0 0.009 FVC % pred# 1.10 0.76–1.62 0.60 1.54 0.95–2.51 0.081 FEV1 % pred# 1.21 0.82–1.79 0.33 1.46 0.93–2.29 0.099 DLCO % pred# 1.42 0.93–2.16 0.11 2.14 1.27–3.60 0.004 Total emphysema¶ 1.19 0.87–1.65 0.28 1.29 0.92–1.79 0.14 Total emphysema¶ % 1.10 0.85–1.41 0.48 1.16 0.87–1.54 0.32 Total GG¶ 1.40 1.05–1.85 0.022 1.30 0.88–1.90 0.19 Total GG¶ % 1.29 0.98–1.70 0.072 1.21 0.85–1.73 0.28 Total reticular¶ 1.35 0.91–1.99 0.14 1.91 1.21–3.0 0.006 Total reticular¶ % 1.38 1.03–1.84 0.032 1.93 1.30–2.89 0.001 Total honeycombing¶ 0.95 0.74–1.21 0.68 1.03 0.80–1.31 0.84 Total honeycombing¶ % 0.94 0.74–1.20 0.62 1.02 0.80–1.29 0.90 Total ILD¶ 1.47 1.08–2.0 0.016 1.70 1.19–2.43 0.003 Total ILD¶ % 1.30 0.99–1.71 0.062 1.52 1.08–2.15 0.017 Emphysema radiologist 1¶ 1.32 0.89–1.97 0.17 1.32 0.88–1.98 0.18 Emphysema radiologist 2¶ 0.82 0.52–1.29 0.38 1.48 0.79–2.77 0.22 GG radiologist 1¶ 1.16 0.86–1.57 0.32 1.40 0.97–2.02 0.070 GG radiologist 2¶ 1.02 0.77–1.35 0.88 1.24 0.92–1.67 0.16 Honeycombing radiologist 1¶ 1.07 0.74–1.54 0.73 1.12 0.72–1.75 0.61 Honeycombing radiologist 2¶ 1.10 0.79–1.54 0.56 1.11 0.75–1.65 0.60 Reticular radiologist 1¶ 1.11 0.84–1.47 0.46 1.20 0.87–1.64 0.27 Reticular radiologist 2¶ 1.14 0.70–1.84 0.60 1.03 0.63–1.68 0.92 Change radiologist 1 Improved/stable Ref. Ref. Worse 2.1 1.09–4.04 0.026 2.22 1.08–4.56 0.029 Change radiologist 2 Improved/stable Ref. Ref. Worse 1.86 0.96–3.61 0.066 2.50 1.12–5.54 0.025 CALIPER: Computer-Aided Lung Informatics for Pathology, Evaluation and Ratings; HR: hazard ratio; TLC: total lung capacity; % pred; % predicted; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; DLCO: diffusing capacity of the lung for carbon monoxide; ILD: interstitial lung disease (ground glass (GG) + reticular + honeycombing); Ref: reference.#: HRs are for a 15-unit change in DLCO % pred and a 10-unit change in FEV1 % pred, FVC % pred and TLC % pred; ¶: HRs are for a sd change.
Supplementary material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Files in this Data Supplement:
Supplementary material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Files in this Data Supplement:
- Supplementary material -
1. Texture Feature extraction
1.1 Training set generation
1.2 Visualization of VOI similarities
1.3 Assessing separability quality of CVM similarity metric
1.4 Identification of exemplar VOIs
2. Data Processing
2.1 Lung Segmentation
2.2 Airway extraction
2.3 Lung Separation
2.4 Vessel Extraction
2.5 Parenchymal Classification
3. Variations in Parenchymal Classification due to slice thickness
- Supplementary material -
Disclosures
Files in this Data Supplement:
