Two-Stage Model for Time to Breast Cancer Mortality Among a Cohort of Initially Disease-Free Women

• Activity Number: 211 - Disease Prediction
• Type: Contributed
• Date/Time: Tuesday, August 10, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Bernard Rosner* and Robert Glynn and Heather Eliassen and Susan Hankinson and Rulla Tamimi and Wendy C Chen and Michelle Holmes and Yi Mu and Graham A Colditz and Walter C Willett and Shelley S Tworoger
• Companies: Harvard Medical School/School of Public Health and Harvard Medical School and Harvard Medical School/School of Public Health and Harvard Medical School/School of Public Health/UMass-Amherst and Harvard Medical School/School of Public Health/Weill Cornell Medical School and Harvard Medical School/Brigham and Women's Hospital/Dana Farber Cancer Institute and Harvard Medical School/School of Public Health and Harvard Medical School/Brigham and Women's Hospital and Washington University School of Medicine and Harvard Medical School/School of Public Health and Moffitt Cancer Center
• Keywords: survival analysis; lethal cancer; multiple event analysis; multi-state model

Abstract

We propose a two-stage model for evaluating risk of lethal breast cancer based on separate models for breast cancer incidence among disease-free individuals and breast-cancer-specific mortality among cases, and integrate them into a cumulative incidence model for lethal breast cancer among disease-free women. The methods are applied to Nurses’ Health Study data, where we evaluated cumulative incidence of lethal breast cancer over 25 years among cancer-free women in 1990 using fixed baseline risk factors. Between 1/1/90-12/31/14 4257 incident breast cancers occurred resulting in 553 deaths due to breast cancer. Some risk factors (e.g., current estrogen & progesterone use) were positively associated with incidence, but inversely associated with mortality, yielding a RR for lethal breast cancer = 1.02 (95% CI = 0.77,1.34). However, other risk factors (e.g., weight gain > 30 kg vs ±5 kg since age 18) were positively associated with both incidence and mortality yielding a RR for lethal breast cancer = 1.84 (95% CI = 1.31, 2.59). This two-stage model may be a useful tool for identifying pre-diagnosis factors that lead to more aggressive and ultimately lethal disease.