Table 1 Details of training data
From: Development and performance evaluation of a deep learning lung nodule detection system
Characteristic | # |
|---|---|
Data information | |
 No. of chest CTs | 1997 |
 No. of patients | 1769 |
  No. of men* | 688 |
  No. of women* | 397 |
 Mean age* (y) | 67.7 ± 10.9 |
 Slice thickness (mm) | |
  ≤ 1.0 | 475 |
   ≤ 2.0 | 547 |
   ≤ 3.0 | 308 |
   ≤ 4.0 | 1 |
   ≤ 5.0 | 437 |
 Contrast | |
  Contrast-enhanced | 386 |
  Non-contrast-enhanced | 1611 |
 Kernel type | |
  Lung kernel | 1268 |
  Abdomen kernel | 533 |
  Bone kernel | 196 |
 Manufacturer | |
  Toshiba Medical Systems Corp | 1001 |
  GE Healthcare | 511 |
  Siemens Healthineers AG | 423 |
  Philips | 62 |
Nodule information | |
 No. of nodules | 10,467 (5.2/scan) |
 Mean nodule size (mm) | 8.3 ± 7.9 |
 No. of nodules according to size (mm) | |
   ≤ 5.0 | 4150 |
   ≤ 10.0 | 4406 |
   ≤ 20.0 | 1071 |
   ≤ 30.0 | 538 |
   ≤ 40.0 | 156 |
   ≤ 50.0 | 91 |
   > 50.0 | 55 |
 Lobular distribution | |
  Right upper | 2852 |
  Right middle | 980 |
  Right lower | 2177 |
  Left upper | 2601 |
  Left lower | 1857 |