Impacto dos subtipos moleculares e variáveis clínico-patológicas nos desfechos do câncer de mama: um estudo entre grandes bases de dados

Abstract

Breast cancer represents a major global health challenge, and the use of publicly available databases can contribute to a more comprehensive understanding of disease patterns. Given the potential of these data sources to support complementary analyses, this study aimed to explore different publicly available databases to investigate demographic, tumor, and therapeutic characteristics of breast cancer. An observational, retrospective study was conducted using hospital-based, population-based, and molecular records from four databases: the São Paulo Hospital-Based Cancer Registry (HBCR), the National HBCR, the Surveillance, Epidemiology, and End Results (SEER) Program, and the genomic and clinical data from the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC). Together, these four databases comprised 1,128,269 breast cancer records. Diagnoses were made between 2000 and 2015, except for the METABRIC database. Fine-Gray competing risks modeling was employed to evaluate tumor recurrence over a 120-month follow-up period, stratified into two time intervals (0–60 and 61–120 months). Cox proportional hazards models were applied to estimate 5-year overall and cause-specific survival. Road travel distances to cancer treatment facilities in Brazil were estimated using the geobr and osrm statistical packages. All analyses were performed in the RStudio computing environment, version 2026.01.1+403. Independent predictors of the average risk of tumor recurrence included tumor size, the number of positive lymph nodes, histological grade 3, and intrinsic molecular subtypes defined by the PAM50 gene expression signature. Temporal dynamics analysis demonstrated that the risk of tumor recurrence was not constant over time. Regarding 5-year survival, age and tumor stage were independent predictors of the average risk of overall and cause-specific mortality. Year of diagnosis was inversely associated with the risk of cause-specific mortality, serving as a temporal marker of changes occurring throughout the study period. An average reduction of 2.81% in the Sâo Paulo HBCR and 2.96% in the population-based registry was observed in the risk of cause-specific mortality for each more recent year of diagnosis. In the National HBCR, 52.09% of the sample received treatment outside their municipality of residence. Median road travel distances by treatment modality ranged from 76.6 km for hormone therapy to 114.0 km for immunotherapy. Geographic mapping revealed inequalities in the distribution of oncology services, demonstrating that treatment centers were highly concentrated in the Southeast, South, and Northeast regions of Brazil. The study concludes that different databases provide complementary information that expands the analytical capabilities for investigating breast cancer outcomes, enabling analyses that would not be possible using a single data source. These findings highlight the value of combining information from multiple data sources to support researchers and healthcare decision-makers in monitoring disease patterns and planning oncology care strategies.