Table 1. Median abnormality scores assigned by AI-CAD according to clinical and radiological characteristics of the 656 breasts.
Variable | n (%) | Median score (IQR) | P * | |
---|---|---|---|---|
Indication | < 0.001 | |||
Screening | 312 (48) | 26 (15, 51) | ||
Diagnostic | 344 (52) | 52 (21, 95) | ||
Palpability | < 0.001 | |||
Yes | 287 (44) | 61 (27, 97) | ||
No | 369 (56) | 24 (15, 51) | ||
Density | < 0.001 | |||
Category A | 16 (2) | 61 (39, 99) | ||
Category B | 82 (13) | 71 (26, 98) | ||
Category C | 347 (53) | 32 (18, 76) | ||
Category D | 211 (32) | 29 (16, 56) | ||
Density (binary) | < 0.001 | |||
Fatty | 98 (15) | 70 (26, 99) | ||
Dense | 558 (85) | 31 (17, 66) | ||
Finding | < 0.001 | |||
Asymmetry | 115 (18) | 30 (17, 58) | ||
Distortion | 13 (2) | 63 (30, 70) | ||
Mass | 145 (22) | 74 (37, 97) | ||
Mass with microcalcifications | 66 (10) | 98 (61, 100) | ||
Microcalcifications only | 159 (24) | 33 (20, 61) | ||
Negative | 158 (24) | 18 (13, 27) |
*P-values are for comparing the median scores and were calculated using the Mann-Whitney U test for indication, palpability, and density and using the Kruskal-Wallis test for density and finding.
AI-CAD = artificial intelligence-based computer-aided diagnosis, IQR = interquartile range