Principal investigator Dr. Etta Pisano from the University of North Carolina School of Medicine in Chapel Hill and colleagues evaluated data from the Digital Mammographic Imaging Screening Trial (DMIST) to further compare the accuracy of FFDM to FSM in various subgroups of women.
DMIST, sponsored by the American College of Radiology Imaging Network (ACRIN), involved nearly 50,000 asymptomatic women who presented for screening mammography at 33 sites in the U.S. and Canada. Five different FFDM systems were used:
- SenoScan (Fischer Imaging, Denver)
- FCRm (Fujifilm Medical Systems USA, Stamford, CT)
- Senographe 2000D (GE Healthcare, Chalfont St. Giles, U.K.)
- Selenia Full Field Digital Mammography System (Hologic, Bedford, MA)
- Lorad/Trex Digital Mammography System (Hologic), subsequently replaced by Lorad/Hologic Selenia Full Field Digital Mammography System (Hologic)
Breaking out data
Pisano and colleagues retrospectively broke out DMIST data according to the following 10 subgroups:
- Pre- or perimenopausal women under 50 with dense breasts
- Pre- or perimenopausal women under 50 with fatty breasts
- Postmenopausal women younger than 50 with dense breasts
- Postmenopausal women younger than 50 with fatty breasts
- Pre- or perimenopausal women between 50 and 64 with dense breasts
- Pre- or perimenopausal women between 50 and 64 with fatty breasts
- Postmenopausal women between 50 and 64 with dense breasts
- Postmenopausal women between 50 and 64 with fatty breasts
- Women age 65 or older with dense breasts
- Women age 65 or older with fatty breasts
They evaluated all cancers according to certain characteristics, such as mammographic method: digital versus film versus both versus neither; lesion type: mass, calcifications, or other; digital machine type; mammographic and pathologic size and diagnosis; existence of prior mammographic study at the time of interpretation; months since prior mammographic study; and compressed breast thickness.
The researchers estimated receiver operating characteristic curves for digital and film-screen mammography from combined DMIST data using a seven-point malignancy score assigned to each woman at the time of screening mammography and before further action was taken. The areas under the receiver operating characteristic curves (AUCs) were compared with a bivariate binormal model.
Pisano and her colleagues also estimated sensitivity, specificity, and positive predictive value (PPV) of digital and film-screen mammography on the basis of the BI-RADS scores assigned at initial screening interpretation.
| FFDM |
Subgroup
(see above) |
AUC |
Sensitivity |
Specificity |
PPV |
| 1 |
0.791 |
0.591 |
0.896 |
0.033 |
| 2 |
0.949 |
0.857 |
0.905 |
0.027 |
| 3 |
0.836 |
0.714 |
0.902 |
0.044 |
| 4 |
0.922 |
0.857 |
0.923 |
0.066 |
| 5 |
0.756 |
0.609 |
0.911 |
0.075 |
| 6 |
0.777 |
0.667 |
0.932 |
0.087 |
| 7 |
0.766 |
0.554 |
0.916 |
0.053 |
| 8 |
0.731 |
0.441 |
0.933 |
0.046 |
| 9 |
0.820 |
0.515 |
0.913 |
0.073 |
| 10 |
0.705 |
0.532 |
0.938 |
0.092 |
|
| All data courtesy of Radiology, February 2008, Vol. 246: 2, p. 380.
|
| FSM |
Subgroup
(see above) |
AUC |
Sensitivity |
Specificity |
PPV |
| 1 |
0.544 |
0.273 |
0.889 |
0.015 |
| 2 |
0.850 |
0.714 |
0.912 |
0.024 |
| 3 |
0.951 |
0.857 |
0.905 |
0.054 |
| 4 |
0.800 |
0.571 |
0.926 |
0.047 |
| 5 |
0.743 |
0.522 |
0.906 |
0.062 |
| 6 |
0.677 |
0.500 |
0.922 |
0.058 |
| 7 |
0.668 |
0.446 |
0.912 |
0.041 |
| 8 |
0.744 |
0.529 |
0.935 |
0.056 |
| 9 |
0.705 |
0.500 |
0.908 |
0.070 |
| 10 |
0.877 |
0.694 |
0.945 |
0.127 |
|
FFDM: Who benefits most?
The subgroup of pre- or perimenopausal women younger than 50 years old with dense breasts was the only one for which FFDM performed significantly better than FSM: The AUC for digital mammography was 0.79 versus 0.54 for film-screen mammography. Sixteen cancers in this group were found with digital and missed with film-screen mammography, while only two were found with film-screen and missed with digital.
The results also showed a trend toward improved accuracy of FSM over FFDM in women age 65 or over with fatty breasts, but it was not significant, according to the researchers. In this group, 15 cancers were found with film-screen mammography and missed with digital, and four were found with digital and missed with film-screen mammography.
Although the study authors explored possible reasons for these two trends, they could not discover definitive causes based on the original DMIST data.
By Kate Madden Yee
AuntMinnie.com staff writer
January 29, 2008
Related Reading
All-digital mammography for breast cancer screening not cost effective, December 31, 2007
Making dollars and sense out of Medicaid and FFDM reimbursement, December 27, 2007
More DMIST analysis supports FFDM in younger women, dense breasts, November 26, 2006
Copyright © 2008 AuntMinnie.com