Differential impact of prognostic parameters in hormone receptor-positive lobular breast cancer.

ABSTRACT

BACKGROUND: Invasive lobular breast cancer (BC) is the second most common BC subtype. Prognostic parameters (tumor classification, lymph node status, histologic grade, Oncotype DX recurrence score [RS], progesterone receptor status, and Ki67 index) were retrospectively studied in a large, prospective clinical trial encompassing 2585 patients who had hormone receptor-positive early BC (the West German Study Group PlanB trial). METHODS: BCs were centrally reviewed and classified as lobular (n = 353; 14%) or nonlobular (n = 2232; 86%). The median follow-up was 60 months. Five-year disease-free survival (DFS) estimates were obtained using the Kaplan-Meier method. Prognostic parameters were evaluated using Cox proportional hazard models. RESULTS: Lobular BC was associated with higher tumor classification, higher lymph node status, lower histologic grade, lower Ki67 index, and low or intermediate RS. The prevalence of high RS (RS range, 26-100) was 3-fold lower in patients who had lobular BC compared with those who had nonlobular BC (8% vs 24%; P < .001). However, 5-year DFS estimates for lobular and nonlobular BC were similar (92.1% and 92.3%, respectively; P = .673). In multivariate analyses, prognostic parameters for DFS in lobular BC included grade 3 (hazard ratio, 5.06; 95% CI, 1.91-13.39) and a pathologic lymph node status (pN) of pN3 (hazard ratio, 12.16; 95% CI, 3.87-38.24), but not RS. By contrast, prognostic parameters in nonlobular BC included grade 3 (hazard ratio, 1.65; 95% CI, 1.11-2.44), pN3 (hazard ratio, 3.68; 95% CI, 1.60-8.46), and high RS (hazard ratio, 2.49; 95% CI, 1.69-3.68). CONCLUSIONS: Lobular BC is associated with low and intermediate RS, although 5-year DFS is similar to that of nonlobular BC. The effect of the RS in lobular BC appears to be distinct from that in nonlobular BC. For risk assessment, the RS needs to be complemented by clinicopathologic parameters for therapy decision making.