Mark Alan Helvie MD, Presenter: Institutional grant, General Electric Company Consultant, General Electric Company

R. Edward Hendrick PhD, Abstract Co-Author: Consultant, General Electric Comp Advisory Board, Bracco Group Medical Advisory Board, Koning Corporation

To estimate the number of women needed to screen to prevent one breast cancer death (NNS) with annual or biennial screening mammography. Confusion exists between number needed to invite to screening in a randomized control trial (RCT) to prevent one breast cancer death (NNI) and NNS.

Cancer Intervention and Surveillance Modeling Network (CISNET) modeling used by the United States Preventative Service Task Force (USPSTF) to evaluate mortality reduction of screening mammography is used to estimate NNS and NNS per life-year gained (NNS/LYG) with annual or biennial screening mammography, without further assumptions. We used mean values of six CISNET models for mortality reduction and LYG per 1000 women screened. 2009 USPSTF NNI values based on historic RCT data are compared to NNS. These NNI values reflect the nature of RCTs, which include non-compliance, crossover, and variable numbers of screening rounds and intervals of screening.

For women ages 40-49 years undergoing annual screening mammography, CISNET modeling estimates a NNS of 746, 2.6 times smaller than the NNI of 1904 estimated by the USPSTF for this age group. NNS is 351 for women 50-59 (3.8 times smaller than NNI), 233 for women 60-69 (1.6 times smaller than NNI), and 377 for women 70-79. For annually screened women, NNS per life-year gained is 28 for age 40-49, 17 for age 50-59, 16 for age 60-69, and 40 for age 70-79. For biennially screened women, NNS is 1034 for age 40-49, 455 for age 50-59, 303 for age 60-69, and 426 for age 70-79; NNS/LYG is 37 for age 40-49, 22 for age 50-59, 20 for age 60-69, and 45 for age 70-79. Annual screening of women ages 40-84 results in a NNS of 84 and a NNS/LYG of 5.3. Biennial screening of women age 50-74 results in a NNS of 144 and a NNS/LYG of 9.1.

NNS based on CISNET modeling is far smaller than USPSTF estimates of NNI based on RCT data. 84 women would need to be screened annually starting at age 40 to save one life from breast cancer and 5.3 women screened annually to gain one life-year. Our results support screening mammography starting at age 40 years as a reasonable strategy to save lives from breast cancer.

NNS to save one life from breast cancer based on CISNET  modeling is far smaller than USPSTF estimates of NNI based on historic experimental RCT

Helvie, M, Hendrick, R, Mammography Screening: A New Estimate of Number Needed to Screen to Prevent One Breast Cancer Death.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11003555.html