Breast cancer is a heterogeneous disease with substantial variability in relapse rates across clinical and molecular features of the tumor. Clinicopathologic prognostic models (e.g.,Adjuvant!) are widely used to guide therapy. Prognostic gene expression signatures are then used to help clarify treatment options for clinically intermediate risk patients. PAM50 is a validated "open source" gene expression signature which has recently been endorsed for informing treatment decisions. However, for 40% of breast cancer patients estimated risk, and hence treatment, remains uncertain. Improvement in prognostic models is an active area of research. Models that integrate genomic and clinicopathologic factors show large improvements in prognostic ability. Newer statistical methods (non-linear and time-varying effects, competing risks for comorbid conditions) have improved risk estimates among subgroups and increased accuracy for late relapse (>5 or 10 years after diagnosis). Combination of orthogonal gene expression signatures with microRNA's (miRNA's) promises to increase the proportion of subjects assigned to low and high-risk groups. However, clinical utility of newer models is limited because additional validation in large cohorts with long-term follow-up is needed. Team: We have assembled unique resources to address these issues, including 1,723 well-annotated primary tumor samples from the WHEL cohort of community-dwelling early stage breast cancer survivors. WHEL participants have been actively followed for 15+ years, with regular assessment of comorbidities, body mass index, and physical activity. The team is comprised of leaders in breast cancer treatment and assay development, in genomics, in behavioral research, and in mathematical statistics and prognostic modeling.
We aim to: 1. Assay PAM50, hypoxia gene signatures, and an 18 miRNA panel. Assays will be done at the Genome Institute, Washington Univ., St. Louis. We expect high-quality assays on ~1300 WHEL subjects. 2. Validate the PAM50/hypoxia clinco-genomic signatures, assess potential clinical benefit and identify weaknesses. We will compute risk estimates from published clinicopathologic models (Adjuvant), and genomic (PAM50, hypoxia, miRNA) signatures. We will validate models by comparing predicted and observed survival. Impact: The expected independent validation of the PAM50/hypoxia clinico-genomic signature will help translate its use into the clinic. Improvements for clinically intermediate-risk patientsare anticipated. 3, Develop a new prognostic model accounting for competing comorbidities, nonlinear and time- varying effects, and integrating novel genomic, and clinical factors. New models will integrate clinical and genomic features (i.e., PAM50, hypoxia signatures, miRNAs, comorbidities, obesity) and incorporate statistical refinements such as competing risk from non-cancer mortality and time-varying effects. Impact: We expect to further improve risk estimates for clinically intermediate risk and late relapse groups, and to quantify the expected clinical benefit which might result from these improved models.

Public Health Relevance

Breast cancer is a complex disease, and prognostic models which use clinical data or genomic signatures assayed from the tumor are used by clinicians to predict risk of relapse and aid in treatment decisions. Modifiable lifestyle factors such as physical activity or obesity are known to impact prognosis, but these factors are rarely incorporated into breast cancer survival prediction models. In this project, using a unique cohort of breast cancer patients with archived primary tumor samples, longitudinally collected lifestyle data, and long term follow-up, we propose to validate and compare several existing prognostic tools across a wide spectrum of breast cancer subtypes and to develop an improved prognostic model which integrates clinical, genomic, and lifestyle / psychosocial risk factors such as depression, obesity and physical activity levels.

National Institute of Health (NIH)
National Cancer Institute (NCI)
Research Project (R01)
Project #
Application #
Study Section
Epidemiology of Cancer Study Section (EPIC)
Program Officer
Divi, Rao L
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of California San Diego
Family Medicine
Schools of Medicine
La Jolla
United States
Zip Code
Pierce, John P; Patterson, Ruth E; Senger, Carolyn M et al. (2014) Lifetime cigarette smoking and breast cancer prognosis in the After Breast Cancer Pooling Project. J Natl Cancer Inst 106:djt359
Villasenor, Adriana; Flatt, Shirley W; Marinac, Catherine et al. (2014) Postdiagnosis C-reactive protein and breast cancer survivorship: findings from the WHEL study. Cancer Epidemiol Biomarkers Prev 23:189-99
Bao, Lei; Pu, Minya; Messer, Karen (2014) AbsCN-seq: a statistical method to estimate tumor purity, ploidy and absolute copy numbers from next-generation sequencing data. Bioinformatics :
Emond, Jennifer A; Pierce, John P; Natarajan, Loki et al. (2014) Risk of breast cancer recurrence associated with carbohydrate intake and tissue expression of IGFI receptor. Cancer Epidemiol Biomarkers Prev 23:1273-9