Breast cancer overdiagnoses in U.S. common, inaccurately reported

December 16, 2021

2 min read


Rysar MD, et al. Abstract GS4-06. Presented at: San Antonio Breast Cancer Symposium (virtual meeting); Dec. 7-10, 2021.

Rysar reports no relevant financial disclosures. Please see the abstract for all other researchers’ relevant financial disclosures.

We were unable to process your request. Please try again later. If you continue to have this issue please contact [email protected]

One in seven U.S. women will be overdiagnosed with breast cancer during their lifetimes, according to study results presented at San Antonio Breast Cancer Symposium.

More than one-third of overdiagnosed cases can be attributed to the detection of nonprogressive cancers, researchers concluded.

Study estimates.

Data derived from Rysar MD, et al. Abstract GS4-06. Presented at: San Antonio Breast Cancer Symposium (virtual meeting); Dec. 7-10, 2021.

“Breast cancer overdiagnosis is defined as the mammographic detection of breast cancers that would not have caused symptoms or other harm in a woman’s remaining lifetime,” Marc D. Rysar, PhD, assistant professor in population health sciences at Duke Cancer Institute, said during a presentation. “There are two different sources of overdiagnosis. One is from the detection of nonprogressive or indolent preclinical cancers, and the second is from the detection of progressive cancers that — in the absence of screening — would not have become symptomatic before other causes of death.”

Despite extensive data, there is a lack of consensus about the extent of breast cancer overdiagnosis.

Marc D. Rysar, PhD

Marc D. Rysar

Rysar and colleagues aimed to produce a reliable model-based estimate of breast cancer overdiagnosis that addresses the various limitations of previous research. The modeling study included 35,986 women (median age, 56 years; 55.6% white) who underwent a combined 82,677 breast cancer mammography screens at a Breast Cancer Surveillance Consortium facility between 2000 and 2018.

“Previous modelling studies have been criticized for simplifying assumptions, such as not accommodating nonprogressive cancers, but we accounted for nonprogressive cancers and used an authoritative data source on contemporary mammography practice,” Rysar said.

Mammography screening detected 90% of the 718 breast cancers identified during the study period.

Researchers estimated 3.6% (95% CI, 0.2-13.8%) of all preclinical cancers detected would be nonprogressive. They estimated a mean tumor latency among progressive breast cancers of 6.5 years (95% CI, 4.9-8.6), and they calculated an 81.7% (95% CI, 72.6-89) screening episode sensitivity of mammography.

Based on these estimates, Rysar and colleagues predicted the overall rate of overdiagnosis with biennial screening for individuals aged 50 to 74 years to be 15.3% (95% CI, 9.7-25.3), including 6% (95% CI, 0.2-19) for nonprogressive cancers and 9.3% (95% CI, 5.8-13.6) for progressive cancers.

Researchers attributed approximately 40% of overdiagnoses to the detection of nonprogressive preclinical cancers and about 60% to progressive cancers with longer latency periods.

When they assessed overdiagnosis by screening round, they estimated overdiagnosis rates of 11.5% (95% CI, 3.8-28.3) during first screens at age 50 years and 23.6% (95% CI, 17.7-31.9) during last screens at age 74 years.

“This increase is due to the competition of tumor latency and competing other-cause mortality,” Rysar said. “Because older women are more likely to die [of] a competing risk during tumor latency, they are also more likely to be overdiagnosed.”

It also is important to account for study limitations, Rysar said.

“The model does not distinguish breast cancers by invasive vs. in situ or molecular subtypes,” he said. “The model also assumes the same sensitivity for progressive and nonprogressive cancers and does not yet account for patient features.”