Table 3.
Comparison study with mammography and Trx1 level
| Variable | BI-RADS | Trx1 (95% CI) | No. | % |
| Breast cancer patientsa | 0 | 23.22 (19.89 to 26.55) | 29 | 28.2 |
| (n = 103) | 1 | 22.02 (NA) | 1 | 1 |
| 2 | 10.26 (NA) | 1 | 1 | |
| 3 | 22.38 (NA) | 1 | 1 | |
| 4 | 21.84 (20.14 to 23.53) | 52 | 50.5 | |
| 5 | 20.43 (17.87 to 22.99) | 19 | 18.4 | |
| 6 | 0(NA) | 0 | 0 | |
| Women without cancerb | 0 | 2.53(NA) | 1 | 2.4 |
| (n = 42) | 1 | 6.16(4.49 to 7.82) | 32 | 76.2 |
| 2 | 6.10(2.34 to 9.87) | 8 | 19 | |
| 3 | 3.61(NA) | 1 | 2.4 | |
| Variable | Confirmed BC patients | Women without cancer | ||
| Mammography | BI-RADS 4-5 | 71 | 0 | |
| BI-RADS 0-3 | 32 | 42 | ||
| Total | 103 | 42 | ||
| Trx1 | ≥ 11.4 ng/ml | 100 | 5 | |
| < 11.4 ng/ml | 3 | 37 | ||
| Total | 103 | 42 | ||
| Mammography + Trx1c | Positive | 101 | 0 | |
| Negative | 2 | 42 | ||
| Total | 103 | 42 | ||
| Variable | Sensitivity(%) | Specificity(%)d | ||
| Mammography | 68.93 | 100.00 | ||
| Trx1 | 97.09 | 88.10 | ||
| Mammography + Trx1 | 98.06 | 100.00 | ||
a BC patients were selected first by their biopsy results, and their matching mammograms were retrieved afterward for analysis. There was 103 out of the 106 participating BC patients
bThe women without cancer was classified according to BI-RADS score. Among the candidates (n = 114), only those whose prior mammograms were archived at CNUH were selected (n = 42)
c In combined tests with mammography and Trx1 level, it was regarded as positive when the test results showed the BI-RADS category over 4 or the Trx1 level was higher than the cut-off value (11.4 ng/ml). Similarly, it was regarded as negative when either result was in BI-RADS categories 1-3 or lower Trx1 level than the cut-off value
d It is impossible to get 100% specificity in the real world. However, the test set in the present study showed mathematically perfect results due to the relatively small number of women without cancer, which indicates the potential of the combined tests