Table 2.
Unadjusted performance, biopsy recommendation, breast cancer detection rates, and breast cancer prognostic characteristics by computer-aided detection (CAD) or year of mammogram*
| Characteristic | Never implemented CAD |
Implemented CAD |
||||
| 1998–2002 | 2003–2006 | P | Before CAD | After CAD | P | |
| Screening mammograms, No. | ||||||
| Total | 635 083 | 328 287 | — | 404 310 | 253 426 | — |
| Without breast cancer | 631 790 | 326 708 | — | 402 473 | 252 413 | — |
| With breast cancer† | 3293 | 1579 | — | 1837 | 1013 | — |
| Invasive breast cancers | 2671 | 1276 | — | 1469 | 761 | — |
| Ductal carcinoma in situ | 622 | 303 | — | 368 | 252 | — |
| Performance measure, % (95% CI) | ||||||
| Specificity | 91.1 (91.0 to 91.2) | 91.3 (91.2 to 91.4) | .008 | 91.9 (91.8 to 92.0) | 91.4 (91.3 to 91.5) | <.001 |
| Recall rate | 9.3 (9.2 to 9.3) | 9.1 (9.0 to 9.2) | .005 | 8.4 (8.3 to 8.5) | 8.9 (8.8 to 9.0) | <.001 |
| Sensitivity | 80.7 (79.3 to 82.0) | 84.0 (82.1 to 85.8) | .005 | 79.7 (77.8 to 81.5) | 81.1 (78.6 to 83.5) | .35 |
| Positive predictive value | 4.5 (4.3 to 4.7) | 4.4 (4.2 to 4.7) | .63 | 4.3 (4.1 to 4.5) | 3.6 (3.4 to 3.9) | <.001 |
| Breast biopsy recommendation‡ | ||||||
| Biopsies recommended, No. | 9068 | 4271 | — | 5116 | 2647 | — |
| Biopsies recommended per 1000 screens (95% CI) | 14.3 (14.0 to 14.6) | 13.0 (12.6 to 13.4) | <.001 | 12.7 (12.3 to 13.0) | 10.4 (10.1 to 10.8) | <.001 |
| Cancer detection rate per 1000 mammograms (95% CI) | ||||||
| All breast cancers | 4.2 (4.0 to 4.3) | 4.0 (3.8 to 4.3) | .30 | 3.6 (3.4 to 3.8) | 3.2 (3.0 to 3.5) | .01 |
| Invasive breast cancers | 3.3 (3.2 to 3.5) | 3.2 (3.0 to 3.4) | .27 | 2.8 (2.7 to 3.0) | 2.3 (2.1 to 2.5) | <.001 |
| Ductal carcinoma in situ | 0.9 (0.8 to 0.9) | 0.9 (0.8 to 1.0) | .91 | 0.8 (0.7 to 0.9) | 0.9 (0.8 to 1.0) | .13 |
| Ductal carcinoma in situ, % | 18.9 | 19.2 | .80 | 20.0 | 24.9 | .003 |
| Invasive cancer stage (n = 5504), No. (%) | ||||||
| 1 | 1483 (63.2) | 737 (62.8) | .19 | 773 (60.1) | 405 (58.2) | .036 |
| 2 | 738 (31.4) | 353 (30.1) | — | 420 (32.6) | 216 (31.0) | — |
| 3 | 103 (4.4) | 67 (5.7) | — | 75 (5.8) | 65 (9.3) | — |
| 4 | 23 (1.0) | 17 (1.5) | — | 19 (1.5) | 10 (1.4) | — |
| Invasive tumor size (n = 5613) | ||||||
| Mean (SD), mm | 18.1 (14.4) | 18.0 (15.4) | .85 | 18.6 (14.1) | 19.4 (15.0) | .25 |
| Small (<15 mm), No. (%) | 1460 (60.2) | 724 (61.8) | .36 | 765 (58.1) | 397 (56.8) | .56 |
| Lymph node involvement of invasive cancers (n = 5845) | ||||||
| Nodes negative, No. (%) | 1893 (74.4) | 904 (74.6) | .92 | 993 (72.5) | 531 (73.7) | .59 |
| Advanced cancers (n = 7521), No. (%) | 1247 (38.8) | 570 (37.3) | .31 | 712 (40.0) | 369 (37.1) | .14 |
*
CI = confidence interval; SD = standard deviation. P values are from two-sided χ2 or paired t-tests for categorical or continuous outcomes respectively.
†
Includes invasive breast cancer and ductal carcinomas in situ.
‡
Radiologist recommended biopsy, surgical consultation, or fine needle aspiration within 90 days of the screening mammogram.