Skip to main content
. 2018 Sep 22;38(2):172–180. doi: 10.14366/usg.18012

Table 6.

Interobserver agreement according to the radiologist’s experience in the tests performed in the first year of the MUST-BE trial

Final assessment κ-valuea)
Initially modified categorization
Re-modified categorization
Junior (<10 yr) Senior (>10 yr) Junior (<10 yr) Senior (>10 yr)
Category 2 0.451 (0.288-0.616) 0.457 (0.311-0.599) 0.781 (0.675-0.878) 0.807 (0.717-0.890)
Category 3 0.401 (0.263-0.540) 0.410 (0.298-0.519) 0.557 (0.403-0.722) 0.503 (0.327-0.644)
Category 4 0.584 (0.464-0.716) 0.528 (0.439-0.627) 0.593 (0.476-0.717) 0.620 (0.524-0.708)
Category 5 0.790 (0.564-0.928) 0.612 (0.416-0.779) 0.586 (0.229-0.839) 0.542 (0.334-0.716)
Overall 0.528 (0.431-0.641) 0.490 (0.413-0.577) 0.643 (0.553-0.737) 0.648 (0.568-0.726)
Interpretation after dichotomization as benigna) and suspiciousb) lesions 0.680 (0.571-0.794) 0.659 (0.571-0.742) 0.680 (0.579-0.800) 0.756 (0.669-0.841)
Subcategorization of category 4
 Category 4a 0.601 (0.342-0.880) 0.437 (0.244-0.681) 0.542 (0.349-0.730) 0.478 (0.335-0.625)
 Category 4b 0.340 (0.088-0.648) 0.044 (0.040-0.167) 0.184 (0.030-0.374) 0.251 (0.129-0.388)
 Category 4c 0.485 (0.164-0.762) 0.313 (0.118-0.531) 0.397 (0.105-0.650) 0.593 (0.392-0.770)
 Overall subcategorization 0.477 (0.276-0.714) 0.270 (0.136-0.444) 0.375 (0.227-0.553) 0.423 (0.307-0.537)

Numbers in parentheses are the 95% confidence intervals.

MUST-BE, Mammography and Ultrasonography Study for Breast Cancer Screening Effectiveness.

a)

Benign lesions were category 2 or 3.

b)

Suspicious lesions were category 4 or 5.