Breast cancer–specific mortality in invasive lobular versus ductal carcinoma over 30 years.

636 Background: Invasive lobular carcinoma (ILC) differs biologically from invasive ductal carcinoma (IDC) and has been associated with delayed recurrence. However, long-term patterns of breast cancer-specific mortality (BCSM) following ILC remain incompletely described. We aimed to characterize temporal patterns of BCSM associated with ILC vs IDC over extended follow-up. Methods: We conducted a population-based study using the SEER program. Women diagnosed with stage I–III IDC or ILC from 1990 to 2012 were included. BCSM was the primary endpoint, with death from other causes treated as a competing event. Unadjusted cumulative incidence functions estimated BCSM by histology. Fine-Gray competing risks models were used to compare histologies overall and within hormone receptor (HR)-defined subgroups, with all models adjusted for age at diagnosis, year (y) of diagnosis, HR status (where applicable), tumor size, nodal status, tumor grade, surgery, radiation, and chemotherapy. Non-proportional effects were identified visually and modeled using time-varying coefficients. Landmark competing risks analyses assessed late BCSM among long-term survivors. Results: Among 423,479 women included, 41,027 (9.7%) had ILC and 382,452 (90.3%) had IDC; median follow-up was 16.7 y (range 0.1–33 y). During follow-up, 77,926 breast cancer deaths and 111,814 deaths from other causes occurred. In unadjusted analyses of the overall population, BCSM was lower for ILC than IDC at 5 and 10 years and higher for ILC at 20 and 30 years (Table). In adjusted Fine-Gray models, lobular histology demonstrated a significant time-varying association with BCSM (p<0.001). Associations between histology and BCSM differed by HR status (p for interaction=0.0029). In time-varying models restricted to HR+ disease, the adjusted subdistribution hazard ratio (sHR) for ILC versus IDC increased over time, reaching 1.19 (95% CI 1.15–1.23) at 10 y, 1.34 (1.28–1.41) at 20 y, and 1.44 (1.36–1.54) at 30 y. In landmark analyses among HR+ survivors, ILC remained associated with higher late BCSM beyond 10 y (sHR 1.36, 95% CI 1.29–1.43) and beyond 15 y (sHR 1.32, 95% CI 1.20–1.45). Conclusions: ILC is associated with a distinct temporal pattern of BCSM compared with IDC. Although unadjusted early BCSM is lower for ILC, adjusted risk increases over time and exceeds that of IDC in the late post-diagnosis period, particularly among HR+ disease. These findings highlight the importance of long-term risk assessment and survivorship strategies tailored to lobular histology. Unadjusted cumulative incidence of BCSM by histology (%; 95% CI). Overall population HR+ Y since diagnosis IDC ILC IDC ILC 5 9.8 (9.7–9.9) 7.3 (7.0–7.5) 6.5 (6.4–6.6) 6.4 (6.1–6.6) 10 15.0 (14.9–15.2) 14.3 (13.9–14.6) 11.9 (11.8–12.0) 13.4 (13.1–13.8) 20 19.6 (19.5–19.8) 21.6 (21.2–22.1) 17.2 (17.0–17.4) 21.0 (20.5–21.4) 30 22.0 (21.8–22.2) 24.5 (23.9–25.2) 19.8 (19.5–20.0) 23.9 (23.1–24.6)