A Decades-Old Fear About Mammograms May Have Been Wrong Almost the Whole Time

Keypoints:

• A new meta-analysis reexamined all eight major randomized mammography screening trials
• Earlier estimates suggested 30% to 50% of screen-detected breast cancers represented overdiagnosis
• The reanalysis found the trial data is consistent with overdiagnosis under 5%
• Researchers compared trial patterns against decades of Danish routine screening data
• Findings were published September 14 in JNCI: Journal of the National Cancer Institute

For decades, one number has haunted the case for breast cancer screening - the claim that up to half of all cancers caught by mammograms wouldn't have caused harm if left undetected, a phenomenon called overdiagnosis. A new meta-analysis suggests that number was built on a methodological blind spot, and the true rate may sit below 5%.

Led by Sisse Helle Njor, a professor at the University of Southern Denmark and Lillebælt Hospital, the team went back to all eight major randomized trials of mammography screening ever conducted and reanalyzed them together, rather than treating each trial's overdiagnosis estimate in isolation. 

"Randomized trials have often been cited as evidence that overdiagnosis is a substantial problem," Njor said, describing the goal as bringing the full body of evidence together for a clearer picture.

The comparison point came from an unusual natural experiment. Denmark rolled out organized breast cancer screening in some regions roughly 17 years before others, letting researchers track exactly how diagnosis rates shifted both right after screening began and over the following decades. 

When the team lined up matching timepoints between the historical trials and Denmark's long-term screening data, the excess cancer cases detected in the trials closely tracked what Denmark's own programs would predict, patterns consistent with an overdiagnosis rate under 5%, not the 30% to 50% range some earlier estimates suggested.

The authors are careful about what this reanalysis can and can't settle. It reinterprets older trial data rather than running a new randomized experiment, and it doesn't resolve every open question in modern screening debates. Still, for a fear that's shaped screening guidelines and personal decisions for years, a reanalysis pulling every major trial together toward one consistent, much lower number is the kind of finding that tends to ripple through clinical guidance over time.