Importance:

The U.S. Preventive Services Task Force (USPSTF) is updating its 2016 guidelines for screening mammography for breast cancer.

Objective:

To provide the USPSTF with updated model-based estimates of the benefits and harms of breast cancer screening strategies that vary by the ages to begin and end screening, screening modality, and screening interval. Models estimated outcomes for the overall average-risk population of U.S. female persons and for groups of female persons based on Black race, breast density, elevated relative risk of breast cancer, and level of comorbidity.

Design:

Comparative modeling using six microsimulation and analytic models that produce outcomes with and without breast cancer screening in a hypothetical cohort of average-risk U.S. 40-year-old female persons (all races) born in 1980 with no previous breast cancer diagnosis. Analyses were repeated for groups of female persons by Black race, breast density category, elevated risk, and comorbidity level.

Exposures:

Screening from ages 40, 45, or 50 years until ages 74 or 79 years with digital mammography (DM) or digital breast tomosynthesis (DBT) annually or biennially or a hybrid combination of the two intervals. Screening strategies using DBT were evaluated in strata according to breast density categories and, separately, for modestly elevated risk levels of breast cancer (relative risk 1.5 and 2.0). Screening strategies with additional stopping ages (69 and 84) were evaluated for female persons older than 65 years according to four levels of comorbidity (none, low, moderate, severe). Full adherence with all screening was assumed, and all cases received immediate treatment regardless of the method of detection according to current treatment dissemination patterns in the United States.

Main Outcome and Measures:

Estimated lifetime benefits (breast cancer deaths averted, percent reduction in breast cancer mortality, life-years gained [LYG], quality-adjusted life-years [QALYs] gained), harms (false-positive recalls, benign biopsies, overdiagnosis with overtreatment), number of screening tests, and the stage distribution of breast cancers for a cohort of 1,000 40-year-old female persons screened. Trade-offs of harm and benefit were evaluated through efficiency frontier plots and by calculating harm-to-benefit and benefit-to-harm ratios. Efficient (and near-efficient) strategies were those that required fewer mammograms (or similar) per LYG and per breast cancer mortality reduction relative to other strategies.

Results:

Modeling identified five efficient screening strategies resulting in the highest breast cancer mortality reduction and LYG. Efficient strategies involved DBT and biennial screening (ages 50–74, 40–79, or 45–79), annual screening (ages 40–79), and a hybrid combination of intervals (annual at ages 40–49 with biennial at ages 50–79). Across all models for a cohort of 1,000 average-risk 40-year-old female persons including all races, estimated median breast cancer mortality reduction across these five DBT efficient screening strategies compared to no screening ranged from 25.4% to 41.7%, LYG ranged from 120.8 to 229.7, deaths averted ranged from 6.7 to 11.5, lifetime number of mammograms ranged from 11,208 to 34,441, median false-positive recalls ranged from 873 to 2,224, and the number of overdiagnosed cases ranged from 12 to 25.

Four models of breast cancer in Black female persons identified three efficient DBT screening strategies, two with biennial (ages 40–79 or 45–79) and one with annual (ages 40–79) screening. Across the four models for a cohort of 1,000 average-risk Black female persons, estimated median breast cancer mortality reduction across these three efficient screening strategies compared to no screening ranged from 31.2% to 39.6%, LYG ranged from 219.4 to 309.0, deaths averted ranged from 11.7 to 15.5, lifetime number of mammograms ranged from 14,755 to 33,577, false-positive recalls ranged from 1,107 to 2,074, and the number of overdiagnosed cases ranged from 20 to 25. Breast cancer mortality disparities for Black female persons persisted if all female persons obtained mammography with the same screening strategy. More intensive screening for Black female persons (e.g., biennial ages 40 or 45 to 79 with female persons overall screened at ages 50–74) could reduce the elevated disparity in breast cancer mortality rates from 42% to 30%.

Compared with DM, DBT resulted in fewer false-positive recalls, with minimal or modest improvements in mortality for female persons overall and for Black female persons. No DM strategies were efficient or near-efficient in most models for female persons overall or for Black female persons.

When models estimated screening outcomes for female persons with greater breast cancer risk, due to either more dense breast tissue or other risk factors such as a first-degree family history of breast cancer, trade-offs in the benefits and harms of screening improved. Trade-offs were also superior for female persons with a lower comorbidity burden.

Limitations:

To isolate the benefits of screening, all modeled scenarios assumed 100% screening adherence and prompt evaluation of abnormal screening results, which may overestimate the benefit of screening compared to real world implementation. Relative performance of compared strategies might change if adherence or evaluation patterns differ by age, race, or screening frequency. We did not consider imaging modalities besides mammography, individuals at high risk of breast cancer due to genetic susceptibility, or potential risk of breast cancer due to screening-related radiation. Model projections were based on a 1980 U.S. birth cohort with current screening performance and treatment effectiveness assumed for breast cancer diagnosed in the future.

Conclusions:

This collaborative modeling analysis suggests that several mammography screening strategies reduce breast cancer mortality and increase life expectancy in average-risk female persons. Strategies with biennial screening, start ages at 40 or 45, and cessation age at 79 resulted in greater incremental gains in mortality reduction per mammogram compared with most strategies involving annual screening, start age at 50, and/or cessation age at 74. For some groups of female persons with higher risk of breast cancer and breast cancer death, more intensive screening can maintain similar benefit-to-harm trade-offs and reduce mortality disparities.