Table 1 Distribution of GOLD stages11 and the Fleischner visual scores8 in our data selection in the development set (dev) and the evaluation set (eval).
From: Emphysema subtyping on thoracic computed tomography scans using deep neural networks
(a) GOLD stages | ||
|---|---|---|
GOLD stages | #subjects (dev) | #subjects (eval) |
GOLD0 | 981 | 3178 |
GOLD1 | 182 | 570 |
GOLD2 | 440 | 1371 |
GOLD3 | 305 | 771 |
GOLD4 | 205 | 337 |
Non PFT | 0 | 63 |
Non Smoking | 70 | 36 |
PRISm | 324 | 817 |
Total | 2507 | 7143 |
(b) Centrilobular scores | ||
|---|---|---|
Score | #subjects (dev) | #subjects (eval) |
0 (0–1%) | 782 | 2499 |
1 (1–5%) | 431 | 1322 |
2 (5–10%) | 478 | 1409 |
3 (10–20%) | 430 | 1049 |
4 (20–30%) | 275 | 656 |
5 (30–100%) | 111 | 208 |
Total | 2507 | 7143 |
(c) Paramseptal scores | ||
|---|---|---|
Score | #subjects (dev) | #subjects (eval) |
0 (0–1%) | 1145 | 3857 |
1 (1–5%) | 739 | 1865 |
2 (5–100%) | 623 | 1421 |
Total | 2507 | 7143 |
