Women made up 99.3 percent of the cohort, with male patients slightly more common among the older group. The investigators excluded carcinoma in situ and patients with recurrent disease, including only those whose initial oncologic management took place at one of the two participating centers. Clinically, the two age groups were more alike than different in some important respects.

The distribution of disease stage at diagnosis was statistically similar: 61.3 percent of younger patients presented with non-metastatic disease, stages I through III, compared with 58.6 percent of older patients, while metastatic stage IV disease accounted for 38.7 percent and 41.4 percent respectively, a difference that did not reach statistical significance. The relatively high proportion of stage IV disease, the authors note, likely reflects the referral patterns of tertiary cancer centers, which receive many patients already diagnosed elsewhere, and should not be read as representative of the broader Brazilian breast cancer population. Two demographic and clinical differences did stand out.

Younger patients had a lower median body mass index, 24.6 versus 26.1 kilograms per square meter, and were significantly more likely to undergo germline genetic testing, at 21.3 percent versus 10.7 percent. Median follow-up was similar across groups, at 28.25 months for the younger cohort and 27.10 months for the older one. During follow-up, 75 patients under 50 and 135 patients 50 or older died.

The researchers assigned each patient an individualized life expectancy drawn from the 2022 Complete Life Tables published by the Brazilian Institute of Geography and Statistics, matched by year of birth, sex, and place of residence. For patients who died, YLL was the difference between projected life expectancy and age at death, with negative values truncated at zero; survivors were assigned a YLL of zero. Among patients who died during follow-up, the median YLL was 42 years for those diagnosed before age 50, with an interquartile range of 37.5 to 46.0 years, compared with 21 years, interquartile range 13.0 to 26.0, for those diagnosed at 50 or older.

The distributions tell the same story visually: YLL values among younger decedents clustered between 40 and 50 years, while those of older decedents clustered between 20 and 30 years. The authors are careful to emphasize that this difference is an inherent arithmetic consequence of death occurring earlier in life, not evidence of more aggressive disease or higher mortality risk in the young. Aggregate burden measures reinforced the pattern.

The total YLL rate per 1,000 patients was 53 percent higher in the younger group, 2,374.6 versus 1,549.1 years, and the combined Disability-Adjusted Life Year rate, which adds Years Lived with Disability to YLL, was 33 percent higher among younger patients, at 3,340.8 versus 2,509.3 per 1,000. By contrast, aggregate YLD rates were nearly identical between groups, 966.3 versus 960.2 per 1,000. YLD was estimated with a simplified stage-based approach, assigning a disability weight of 0.1 to non-metastatic disease and 0.8 to metastatic disease, because the electronic records did not consistently capture transitions between treatment phases, remission, recurrence, or progression.

One of the most striking findings emerged from the regression analysis. After adjustment for age at admission, clinical stage, body mass index, ethnicity, smoking status, and alcohol use, treatment at the participating public institution remained associated with significantly higher odds of experiencing any premature mortality, with an adjusted odds ratio of 2.15, a 95 percent confidence interval of 1.18 to 3.94, and a p-value of 0.013. Stage IV disease carried an adjusted odds ratio of 5.74 for YLL greater than zero compared with stages I through III.

A sensitivity analysis that additionally accounted for germline and somatic testing produced a similar estimate. The observed proportions of patients with YLL greater than zero were 4.9 percent versus 10.2 percent among the under-50 group, and 6.9 percent versus 12.7 percent among the 50-and-older group, at the private and public institutions respectively. The authors stress that because each healthcare sector was represented by a single institution, these results should be interpreted as adjusted institutional associations rather than causal effects of the healthcare system itself.

On the question that has drawn international attention, whether breast cancer is genuinely rising among young women, the study offers a cautious answer. The proportion of diagnoses among patients under 50 fluctuated across the six calendar years, with a numerical increase of roughly 10 percent over the study period in the early-onset group and a 7.1 percent reduction among patients 50 and older, but binomial logistic regression with calendar year as a continuous predictor found no statistically significant temporal trend, p equals 0.30. The investigators suggest that the dip in 2020 followed by a rebound in subsequent years may reflect disruptions and recovery in healthcare utilization, referral patterns, and diagnostic services during the COVID-19 pandemic rather than true changes in disease occurrence.

They also note that as an institutional cohort, the study was not designed to estimate population-based incidence, so it cannot resolve whether the global uptick in early-onset diagnoses reflects a true epidemiological shift, improved detection, or changing risk factor exposures. The study comes with substantial caveats that the authors lay out in detail. It drew on only two institutions in São Paulo state, limiting generalizability to other Brazilian regions where disease burden and risk factors differ.

Calendar year of diagnosis and follow-up duration were not included as covariates in the mortality model, meaning patients diagnosed later had less opportunity for death ascertainment, which may have produced lower YLL estimates, particularly for earlier-stage disease. Body mass index was missing for 37 percent of patients, ethnicity was frequently unreported, histological and molecular subtypes such as triple-negative and HER2-positive disease were not assessed, and patients with hereditary cases were not excluded. Reproductive history, hormone replacement therapy, and treatment modality were also unavailable.

Despite these limitations, the authors argue that the findings underscore the substantial societal burden of premature mortality from early-onset breast cancer even when stage distribution resembles that of older patients, and they call for prospective multicenter studies with longer, standardized follow-up to identify the biological and social determinants shaping outcomes in younger women.

Subject of Research: Clinical outcomes and years of life lost in early-onset breast cancer at two Brazilian tertiary cancer centers

Article Title: Early-onset breast cancer in two Brazilian tertiary cancer centers: a real-world study of years of life lost and clinical outcomes

Article References: Lopes Paim Miranda, D., Dias e Silva, D., Viesser Miyamura, B., de Moura de Souza, M., Teixeira, M. F., Junior, U. D. R., Moura, F., Junior, P. L. S. U., Taranto, P., Pontes Batista, J., Mano, M. S., Aliosha Kaliks Guendelmann, R., Veasey Rodrigues, H., & Rodrigues Beal, J. (2026). Early-onset breast cancer in two Brazilian tertiary cancer centers: a real-world study of years of life lost and clinical outcomes. Frontiers in Oncology, 16, Article 1837330. https://doi.org/10.3389/fonc.2026.1837330

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1837330

Keywords: early-onset breast cancer, years of life lost, DALY, Brazil, cancer epidemiology, young women, germline testing, real-world data, healthcare disparities, tertiary cancer centers, premature mortality, breast cancer burden