The scariest risk of breast cancer screening may have been overstated for decades

A reanalysis of the eight randomized trials that have anchored breast cancer screening policy since the 1960s and 1970s found that mammography’s most feared harm — diagnosing a tumor that would never have caused symptoms — shows up in fewer than 5 percent of screened women, not the 30 percent to 50 percent that has been cited for decades. The finding comes from Sisse Helle Njor of the University of Southern Denmark and Lillebælt Hospital, working with Elsebeth Lynge at the University of Copenhagen and Matejka Rebolj at Queen Mary University of London. Their study revisited the New York Health Insurance Plan trial, the Malmö trial, the Two-County trial, the Edinburgh trial, the Canadian National Breast Screening Study, the Stockholm trial, the Gothenburg trial and the UK Age trial — the full set of randomized studies screening guidelines have leaned on since mammography programs began.

Overdiagnosis is the term for a screening-detected cancer that would never have grown enough to threaten a life, and it has served for years as the strongest argument against routine mammography, cited in guideline debates and by women who skip screening over fear of unnecessary biopsies or treatment. Njor’s team did not collect new patient data. Instead, it reran the eight trials’ results through a single, consistent method for handling the years after screening stopped, rather than letting each trial’s own follow-up cutoff decide how the numbers should be read.

Eight trials measured against Denmark’s own numbers

To judge whether a trial’s pattern of diagnoses pointed to real overdiagnosis or an artifact of timing, the researchers compared each trial against Denmark’s long-running population screening program, which has tracked outcomes for decades outside a trial setting. Danish screening data put overdiagnosis at below 5 percent, and when the eight trials were measured against that same reference pattern, drawn from the Funen region’s screening records, their results lined up with it rather than with the far higher range earlier interpretations had produced.

Matejka Rebolj, a screening epidemiologist at Queen Mary University of London and one of the paper’s authors, put the conclusion in blunt terms: “Randomized trial data are consistent with overdiagnosis of less than five percent, rather than with estimates nearing 50 percent.” Five percent is the number the reanalysis was built to defend against decades of trial data that, examined without a common yardstick, had been read as evidence of a much larger problem.

How early trial cutoffs inflated the count

The reanalysis, titled “Breast cancer overdiagnosis in mammography trials — separating signal from noise,” examined 52 separate timepoints across the trials and traced the inflated estimates to when each study stopped counting cancers. Randomized screening trials run for a fixed period, then track both the screened and unscreened groups afterward to see whether the unscreened group’s cancer count eventually catches up, since a screened woman’s cancer is typically found years before it would otherwise have surfaced. If a trial ends before that catch-up period has fully played out, the early gap between the two groups gets counted as permanent overdiagnosis rather than an artifact of screened women simply being diagnosed sooner.

Elsebeth Lynge, at the University of Copenhagen’s Department of Public Health, described the same mechanism more plainly: researchers who do not account for the timing can mistake an early rise in diagnoses for overdiagnosis that later evidence shows was never there. Several of the eight trials, including the Canadian National Breast Screening Study and the UK Age trial, ended follow-up well before that decline had time to register.

A figure that has shaped mammography guidance for years

The distinction reaches beyond the journal pages where it is debated. The American Cancer Society’s own patient guidance lists overdiagnosis among the specific harms it discloses to women weighing whether to start or continue screening, alongside false positives and unnecessary biopsies, and it has done so using estimate ranges drawn from the same trial literature this reanalysis revisits.

Estimates near the high end of that range have circulated in screening debates for close to two decades. The National Cancer Institute has tracked overdiagnosis studies that put the figure anywhere from roughly one in seven screen-detected cancers to close to half, a spread wide enough that clinicians counseling patients on screening decisions have had little firm ground to stand on. A five-percent estimate, if it holds up under further scrutiny, narrows that spread dramatically.

Robert A. Smith, the paper’s fifth author and a longtime screening researcher, joined Njor, Lynge, Rebolj and Casper Urth Pedersen of the University of Southern Denmark on the reanalysis, which frames its finding as a matter of applying one consistent method rather than discovering new trial data. Njor herself has cautioned that the numbers cited against screening for years rested on an interpretation that was not as straightforward as it looked, given how differently the eight original trials handled the years after screening ended.

No new trial has been run to confirm the below-5-percent figure outside Denmark’s own screening population, and the paper’s authors have not said whether other countries’ screening programs would show the same pattern once measured the same way. That leaves the number resting, for now, on a single national reference point and a reanalysis of data that is, in some cases, more than 50 years old.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview