Table 4.
Interobserver agreement in three quality control tests performed in the first year of the MUST-BE trial
| Final assessment | κ-valuea) |
||
|---|---|---|---|
| BI-RADS categorization | Initially modified categorization | Re-modified categorization | |
| Category 2 | 0.416 (0.325-0.522) | 0.485 (0.359—0.612) | 0.781 (0.694-0.859) |
| Category 3 | 0.393 (0.257-0.555) | 0.435 (0.338-0.537) | 0.511 (0.381-0.632) |
| Category 4 | 0.441 (0.250-0.604) | 0.556 (0.484-0.643) | 0.592 (0.513-0.668) |
| Category 5 | 0.455 (0.261—0.609) | 0.674 (0.477-0.820) | 0.511 (0.327-0.649) |
| Overall | 0.495 (0.318—0.姑1) | 0.521 (0.451-0.600) | 0.626 (0.560-0.688) |
| Interpretation after dichotomization as benignb) and suspiciousc) lesions | 0.512 (0.450-0.589) | 0.676 (0.599-0.751) | 0.725 (0.648-0.797) |
The numbers in parentheses are 95% confidence intervals.
MUST-BE, Mammography and Ultrasonography Study for Breast Cancer Screening Effectiveness; BI-RADS, Breast Imaging Reporting and Data System.
a)
Interobserver agreement was evaluated according to the BI-RADS categorization, the initially modified categorization, and the re-modified categorization including the size and anterior-posterior ratio criteria.
b)
Benign lesions were category 2 or 3.
c)
Suspicious lesions were category 4 or 5.