How researchers define metastatic cancer can change the apparent improvement in survival over time.
An analysis of 35,262 US breast cancer cases found that carefully aligning staging rules across years was essential to interpreting differences between earlier and later diagnoses.
What researchers studied
Researchers used the US Surveillance, Epidemiology, and End Results programme, known as SEER, to examine breast cancers diagnosed from 2000 through 2023.
They developed year-specific rules prioritising TNM metastatic-stage information to make comparisons across changing coding systems more consistent.
The primary analysis included 35,262 people diagnosed with de novo metastatic, or M1, breast cancer from 2010 to 2023 who had complete data for the model.
The study compared a transition period from 2010 to 2013 with a contemporary period from 2014 to 2023.
Statistical models adjusted for demographic characteristics and metastatic organ sites, while sensitivity analyses tested alternative definitions of stage.
What they found
Patients diagnosed in the contemporary period had a lower adjusted breast-cancer-specific mortality hazard, with a hazard ratio of 0.844 and a 95% confidence interval of 0.819 to 0.870.
A weighted analysis gave a similar estimate of 0.853.
But using an older rule that treated any positive staging indicator as evidence of metastasis produced a nearly null crude comparison, showing how the definition could change the story.
Because the association varied over follow-up, the researchers also reported adjusted hazard ratios of 0.879 in the first 24 months and 0.801 thereafter.
Over 36 months, the restricted mean overall-survival difference was 0.66 months.
What it means
Cancer registries are used to assess progress in diagnosis and care.
If staging conventions change between years, an apparent improvement or deterioration may partly reflect the way patients are classified.
This study offers a methodological warning for clinicians, researchers and health-policy analysts: survival trends should be interpreted only after checking whether comparable populations are being measured.
Important limitation
The analysis was observational and could not isolate the effects of treatment advances, imaging, screening or other changes over time.
Registry data and complete-case selection may introduce residual bias.
The reported associations should not be interpreted as causal treatment benefits.
Source Information
Study: Stage harmonized temporal survival comparisons in de novo metastatic breast cancer using SEER
Authors: Kun Fang, Suxiao Jiang, Meng Luo and Yinling Ma
Journal / Source: Scientific Reports
Published: 4 November 2026
