Table 3. Comparison of Lesion Detection Difficulty Metrics and Lesion Characteristics for the Test Set.
Metric | Mean (SD) | |||
---|---|---|---|---|
Lesion diagnosis | Lesion type | |||
Benign | Cancer | Mass | Architectural distortion | |
Total count, No. | 70 | 66 | 121 | 15 |
No. of algorithms that detected lesion within 4 FPs per volume | 7.47 (1.73) | 7.82 (1.76) | 7.63 (1.78) | 7.73 (1.44) |
FPs corresponding to correct prediction considering teams that detected it within 4 FPs per volume | 0.77 (0.49) | 0.55 (0.43) | 0.64 (0.47) | 0.91 (0.47) |
Abbreviation: FP, false positive.