Age is not just a number when it comes to prognosis for invasive breast cancer. According to data presented during the 2020 Society of Surgical Oncology (SSO) International Conference on Surgical Cancer Care, age at diagnosis of breast cancer is a highly prognostic clinical variable that warrants inclusion in future breast cancer staging models.1
Findings from a retrospective analysis of adult women diagnosed with invasive breast cancer showed that age at diagnosis significantly modified the effect of clinical prognostic stage on the risk of mortality (P < .0001). For patients with stage IA breast cancer, for example, the predicted 36-month overall survival decreased dramatically with age. Differences in age within each stage group also became more pronounced as stages increased, authors of the study reported.
Helen Johnson, MD
“The addition of age to the American Joint Committee on Cancer (AJCC) 8th edition clinical prognostic staging system has the potential to improve prognostic value and risk stratification in breast cancer,” said lead study author, Helen Johnson, MD, a general surgery resident at East Carolina University/Vidant Medical Center, Greenville, North Carolina. “As our understanding of the biology of breast cancer evolves, we expect we will not be able to capture the complexities of staging in a simple table, such as the one we use currently. We will increasingly rely on calculator tools like the one we have developed.”
Background and Methods
As Dr. Johnson reported, mounting evidence supports age at diagnosis as an independent predictor of breast cancer prognosis. Previous work has demonstrated a complex quadratic relationship between these two variables.2
For this retrospective population analysis, Dr. Johnson and colleagues collected data from the Surveillance, Epidemiology, and End Results (SEER) database, looking at adult women diagnosed with breast cancer between 2010 and 2015. The researchers extracted all variables necessary to deduce AJCC 8th edition clinical prognostic stage. The primary endpoint was overall survival, as this was used to derive and validate the current staging system.
Dr. Johnson and colleagues then constructed hierarchic statistical models to evaluate the additive and multiplicative effects of age at diagnosis on prognosis. Model One reflected the current clinical prognostic staging system. Model Two incorporated age as a quadratic variable to accurately reflect its complex relationship with survival. Finally, Model Three included age and an interaction term between clinical prognostic stage and age that was found to be significant.
The researchers used the Akaike information criterion (an estimator of in-sample prediction error), time-dependent receiver operator characteristic curves, and the category-free net reclassification improvement index to compare models.
Risk of Breast Cancer–Specific Mortality Increases With Age
Of the 86,637 women included in the study, more than half were classified as Clinical Prognostic Stage 1, and the mean age at diagnosis was 60 years. Mean follow-up was 36 months, and breast cancer–associated mortality represented 63% of the total 6,415 (7.4%) deaths.
As Dr. Johnson reported, age significantly modified the effect of Clinical Prognostic Stage on risk of mortality (P < .0001). “Overall survival curves and breast cancer–specific mortality curves both demonstrated differential prognoses based not only on Clinical Prognostic Stage, but also age at diagnosis,” said Dr. Johnson.
According to Dr. Johnson, predicted 36-month overall survival for stage IA breast cancer was 97.1% based on Clinical Prognostic Stage alone. However, when the model incorporating age at diagnosis was used, 36-month overall survival fell from 99.3% at age 40 to 76.6% at age 90. The differences in age within each stage group also became more dramatic as stage increased, Dr. Johnson reported.
“Overall survival curves at 12, 16, 36, and 48 months all indicate that the relationship between Clinical Prognostic Stage and survival is complex and is dependent on age at diagnosis,” said Dr. Johnson.
Based on this model, Dr. Johnson and colleagues developed an online calculator that is publicly available (https://tinyurl.com/y3qeh8cl). For a given patient, providers can select Clinical Prognostic Stage and age and view individual survival predictions.
DISCLOSURE: Dr. Johnson reported no conflicts of interest.
1. Johnson HM, Irish W, Vohra N, et al: Putting the age in stage. 2020 Society of Surgical Oncology International Conference on Surgical Cancer Care. Abstract 11.
2. Johnson HM, Irish W, Muzaffar M, et al: Quantifying the relationship between age at diagnosis and breast cancer-specific mortality. Breast Cancer Res Treat 177:713-722, 2019.