Pan · The New England journal of medicine 2017 · meta-analysis of clinical trials · n=62923

20-Year Risks of Breast-Cancer Recurrence after Stopping Endocrine Therapy at 5 Years.

Cited 1866 times in the scientific literature.

Level 1 - systematic review of randomized trials

Meta-analysis of individual patient data from 88 clinical trials

PubMed 29117498 · doi:10.1056/NEJMoa1701830 · record verified 2026-08-30

What was done

A meta-analysis evaluated data from 88 clinical trials involving 62,923 women with ER-positive breast cancer who remained disease-free after completing 5 years of scheduled endocrine therapy. Kaplan-Meier and Cox regression analyses stratified by trial and treatment were used to determine recurrence and mortality risks from year 5 to year 20 according to tumor diameter, nodal status (TN), tumor grade, Ki-67, progesterone-receptor status, and HER2 status.

What was found

Distant recurrences continued at a steady rate throughout years 5 to 20. Distant recurrence risk correlated strongly with original TN status: 13% for T1N0, 20% for T1N1-3, 34% for T1N4-9, 19% for T2N0, 26% for T2N1-3, and 41% for T2N4-9. Breast cancer death showed a similar relationship with TN status, whereas contralateral breast cancer did not. Tumor grade and Ki-67 status added moderate independent predictive value, while progesterone-receptor and HER2 status (in trials without trastuzumab) were not predictive. In T1N0 disease, 5-to-20-year distant recurrence risk was 10% for low-grade, 13% for moderate-grade, and 17% for high-grade disease (risks of any recurrence or contralateral breast cancer were 17%, 22%, and 26%, respectively).

Why it matters

This study establishes that the risk of distant recurrence in ER-positive early breast cancer persists steadily for at least 20 years. These long-term absolute risk estimates provide the quantitative baseline needed to evaluate the trade-offs of extending endocrine therapy beyond 5 years.

Limits

Key predictive markers were not measured across the entire cohort (tumor grade was missing in 19,333 patients, Ki-67 was available in only 7,692 patients, and HER2 in 15,418 patients from non-trastuzumab trials). The abstract provides observational follow-up from heterogeneous trial protocols and does not evaluate newer genomic risk scores or modern extended therapy regimens.

Cited by