Table 1.
Characteristics of the Tongji dataset used for algorithm development and internal validation
| Tongji dataset (n=3086) | Development set (n=2447) | Internal validation set (n=639) | ||
|---|---|---|---|---|
| Sex | ||||
| Male | 1544 (50%) | 1235 (50%) | 309 (48%) | |
| Female | 1542 (50%) | 1212 (50%) | 330 (52%) | |
| Age, years | 55 (39–67) | 55 (39–67) | 55 (38–66) | |
| COVID-19 positive CT scans | 2086 (68%) | 1647 (67%) | 439 (69%) | |
| COVID-19 negative CT scans | 1000 (32%) | 800 (33%) | 200 (31%) | |
| CT manufacturers | ||||
| GE Medical System (Chicago, IL, USA) | 140 (5%) | 122 (5%) | 18 (3%) | |
| MinFound Medical Systems (Shaoxing, China) | 6 (<1%) | 6 (<1%) | 0 | |
| Siemens (Munich, Germany) | 2180 (71%) | 1723 (70%) | 457 (72%) | |
| Toshiba (Tokyo, Japan) | 74 (2%) | 51 (2%) | 23 (4%) | |
| United Imaging Healthcare (Shanghai, China) | 686 (22%) | 545 (22%) | 141 (22%) | |
Data are n (%), or median (IQR).