Kefah Mokbel, Chair of Breast Cancer Surgery at London Breast Institute, shared on LinkedIn:
“Is breast cancer overdiagnosis largely a methodological artefact? A new JNCI meta-analysis says yes.
Njor and colleagues (JNCI, 14 September 2026) have reanalysed the randomised mammography trials within a single unified framework and concluded that overdiagnosis is under 5 percent – an order of magnitude below the figures that dominate screening debates.
The approach
Cumulative incidence by trial arm was expressed as relative proportions and compared against a reference pattern derived from the population-based service screening programme in Funen, Denmark.
Control arms that received an exit screen were reanalysed as screened, correcting a long-standing source of inflated excess incidence.
Across 52 timepoints, 73 observed relative proportions aligned closely with the Funen-derived expectation, for invasive disease with and without DCIS. Deviations clustered in early trial phases, where reporting was incomplete.
Why this matters
The UK Independent Panel put overdiagnosis at roughly 19 percent of cancers diagnosed during the screening period in invited women. The Canadian NBSS produced higher figures still. Those estimates underpin much of the informed consent literature and the decision aids we hand to women.
Three caveats before this changes practice
This is a compatibility analysis, not a pooled quantitative estimate. Agreement with an expected pattern is weaker evidence than a confidence interval around a number. Everything rests on the Funen reference being the correct counterfactual. Choose a different reference population and the conclusion may move.
Reclassifying exit-screened control arms is methodologically defensible but does much of the analytical work here. That decision deserves independent scrutiny. Trial-era film mammography also tells us little about DCIS detection in the DBT era.
Balanced counselling should reflect genuine uncertainty in both directions – overstating overdiagnosis deters attendance, understating it erodes informed consent.”
Title: Breast cancer overdiagnosis in mammography trials-separating signal from noise: a meta-analysis
Authors: Sisse Helle Njor, Casper Urth Pedersen, Elsebeth Lynge, Robert A Smith, Matejka Rebolj
Read the full article.
You may also find this interesting:
