Background: PAM50 intrinsic subtypes have been shown to affect breast cancer prognosis.
Methods: A British Columbia cohort of 718 postmenopausal women treated with tamoxifen, without chemotherapy, had tumors intrinsically subtyped (luminal A, luminal B, basal, HER2) and centrally reviewed by immunohistochemistry (IHC) for estrogen and progesterone receptor (ER and PgR). We tested whether intrinsic subtype and other patient and tumor characteristics were associated with type of death.
Results: At median 11.7 years of follow-up, 429 (60%) of 718 women died: 30% of deaths were breast cancer-specific; 30% were other type. In 425 women <70 years, 32% died of breast cancer and 19% of other type. In 293 women ≥ 70, 27% died of breast cancer and 45% of other type. Intrinsic subtype was associated with breast cancer (P = .001); and older age, with other type (P < .001). Additionally, stepwise cause-specific models indicated larger tumor size (P < .001), more positive lymph nodes (P < .001), and less PgR stain (P = .03) were associated with worse breast cancer survival; more positive lymph nodes (P = .002) and lymphovascular invasion (P = .02) were associated with worse other type. Adjusted breast cancer and other type survival is provided by factors at 5, 10, and 15 years.
Conclusions: Intrinsic subtype was associated with breast cancer death, whereas age was associated with other type; most deaths in women ≥ 70 were from other type.
Keywords: Breast cancer–specific death; Other cause–specific death; Postmenopausal; Survival; Tamoxifen.
Copyright © 2017 Elsevier Inc. All rights reserved.