Survival of de novo stage IV breast cancer patients over three decades

Abstract

Background: De novo stage IV breast cancer patients (BC_IV) depict a clinical picture not influenced by adjuvant therapy. Therefore, the time-dependent impact of changes in diagnostics and treatments on progression and survival can best be evaluated in this subgroup.

Methods: BC_IV patients from 1978 to 2013 registered in the Munich Cancer Registry were divided into four periods, and the trends were analysed. Survival was estimated by Kaplan-Meier methods, and prognostic factors were fitted with Cox proportional hazard models.

Results: Between 1978 and 2013, 88,759 patients were diagnosed with 92,807 cases of invasive and non-invasive BC. Of these patients, 4756 patients had distant metastases (MET) at diagnosis. The 5-year survival rate improved from 17.4 to 24.7%, while the pattern of metastases did not change. Improved staging diagnostics, a screening programme and primary systemic therapy changed the composition of prognostic strata. Patients with a similar composition as the 1978-1987 cohort exhibited a median survival difference of 13 months; however, neither univariate nor multivariate analysis showed a survival effect for the four periods as a surrogate indicator for changing treatments. HER2+ patients have with 27.6 months a slightly longer survival than all other BC_IV patients.

Conclusions: Survival of de novo BC_IV has only modestly improved since the late 1970s, partially masked by changing distributions of prognostic factors due to changes in diagnostics.

Keywords: Breast cancer; Metastasis organs; Stage IV; Survival; Trends.

Fig. 1

Kaplan–Meier estimates for overall survival depending of prognostic factors for stage IV BC. a All patients (n = 4756), b hormone receptor [HR+ (84.2%) HR− (15.8%) (n = 3774)], c first treatment: surgical versus systemic with or without subsequent surgery (n = 4756), d T-status with distinct subgroups pT-category, unknown pT and systemic therapy first (n = 4756), e age as prognostic factor (n = 4756) and f reference period 1978–1987 versus samples of patients from the following 3 periods with identical distribution of the 15 strata with age and T-status (n = 1416, log-rank test: p = 0.006). All survival curves, which do not reach the end of x-axis, include up to 10 patients at risk

Fig. 2

Kaplan–Meier estimates for overall survival depending on the site of MET and on selected subgroup for stage IV BC. a One or ≥2 involved sites of MET (n = 4717), b One specific MET (n = 3057), c ≥2 specific MET (n = 3863 MET for 1660 patients: each patient is counted at least twice and therefore the sum is 232.7%), d only bone MET over 4 periods (n = 1668), e only liver MET over 4 periods (n = 404) and f subgroups with different combinations of prognostic factors. All survival curves, which do not reach the end of x-axis, include to 10 patients at risk