Breast Imaging in Focus: How Does Digital Breast Tomosynthesis Perform Across Different Age Groups?

There continues to be considerable debate about when women should begin screening mammography and when screening should stop. In a recent article published in the journal Radiology Imaging Cancer, my research team and I examined how performance of screening tomosynthesis varies across age groups from women in their 40s to women aged 80 years and older.

We conducted a retrospective study at a single academic medical center, including more than 300,000 screening tomosynthesis exams in approximately 94,000 women between 2013 and 2019. We evaluated five age groups ranging from 40 to 49 years to 80 years and older. Each exam was classified as true positive, false positive, true negative, or false negative based on one-year follow-up, and we calculated standard screening performance metrics. We also performed multivariable analyzes adjusting for race, breast density, prior screening mammography, and interpreting radiologists.

Our main finding was that screening tomosynthesis performance generally improved with advancing age. Among women in their 40s, the cancer detection rate was 2.9 cancers per 1,000 exams, exceeding the American College of Radiology's minimum benchmark of 2.5 per 1000. The cancer detection rate increased progressively with age, reaching 8.3 per 1000 among women in their 70s and 10.6 per 1000 among women aged 80 and older.

How Does Age Impact Sensitivity, Specificity and Recall Rates with DBT?

Sensitivity generally increased with age. Sensitivity was 72.5 percent among women in their 40s, increasing to 91.7 percent among women in their 70s, and remaining high at 88.2 percent among women aged 80 and older.

Abnormal interpretation rates or recall rates decreased with age. The abnormal interpretation rate was 9.3 percent among women in their 40s and decreased to 5.5 percent among women aged 80 and older.

Specificity also increased from 91 percent to 95.5 percent across these age groups, importantly, higher cancer detection rates among older women were accompanied by fewer false positive exams. These age-related trends persisted after adjustment for race, breast density, prior screening mammography, and interpreting radiologists.

Key Takeaways from the Study on Age and Breast Cancer Screening with DBT

Our findings have important implications for both ends of the screening age spectrum. For women in their 40s, screening tomosynthesis met established performance benchmarks, supporting the value of initiating screening at age 40.

For older women, particularly those age 75 and older, we observed high cancer detection rates, high sensitivity, and relatively low abnormal interpretation rates. These findings suggest that chronological age alone may not be the best criterion for determining when screening should stop.

Our study was conducted at a single academic medical center, which may limit generalizability. In addition, women who continue screening at older ages may be healthier than those who discontinue screening, introducing potential selection bias. Our study evaluated screening performance rather than mortality benefit. We did not assess overdiagnosis or overtreatment, which are important considerations in older women.

In summary, our study of more than 300,000 screening tomosynthesis exams demonstrated that screening performance generally improves with advancing age. Our findings support screening women beginning at age 40 and challenge the use of a fixed age cutoff for discontinuing screening mammography. Instead, decisions about when to stop screening should be individualized based on overall patient health, life expectancy, and patient preferences.

Dr. Bahl is an associate professor at Harvard Medical School and a breast radiologist at Massachusetts General Hospital in Boston.

Reference

1. Lamb LR, Mercaldo S, Destounis S, Bahl M. Screening digital breast tomosynthesis performance across age groups: analysis of 302,700 examinations. Radiol Imaging Cancer. 2026 Nov;8(6):e260066. doi: 10.1148/rycan.260066.