Effect of dose-dense adjuvant chemotherapy in hormone receptor positive/HER2-negative early breast cancer patients according to immunohistochemically defined luminal subtype: an exploratory analysis of the GIM2 trial

Background: Luminal A-like and luminal B-like subtypes have different sensitivity to (neo)adjuvant chemotherapy, but their role in predicting dose-dense (DD) efficacy in the high-risk setting is unknown. In this exploratory analysis of the Gruppo Italiano Mammella 2 (GIM2) trial, we investigated DD efficacy according to luminal-like subtypes. Methods: Patients with node-positive early breast cancer were randomised to receive either DD or standard-interval (SI) anthracycline-based chemotherapy followed by paclitaxel. In our analysis, luminal A-like cohort was identified as having a Ki67 < 20% and a progesterone receptor (PgR) >= 20%; luminal B-like cohort as having a Ki67 >= 20% and/or a PgR < 20%. Results: Out of 2003 patients enrolled in the GIM2 trial, 412 had luminal A-like and 638 luminal B-like breast cancer. After a median follow-up of 7.9 years, disease-free survival (DFS) was 80.8% (95% confidence interval [CI] 76.4-84.5) and 70.5% (66.5-74.2) in luminal A-like and luminal B-like cohorts; overall survival (OS) was 91.6% (88.2-94.1) and 85.1% (81.7-87.9), respectively. We found no significant interaction between treatment and luminal subtype (interaction p = 0.603 and 0.535 for DFS and OS, respectively). When DD efficacy was investigated separately in each cohort, luminal-B like cohort appeared to benefit more from the DD schedule both in terms of DFS (unadjusted hazard ratio [HR] 0.72 [95% CI 0.54-0.96]) and OS (unadjusted HR 0.61 [95% CI 0.40-0.94]), compared with the luminal A-like cohort (unadjusted HR for DFS 0.89 [95% CI 0.59-1.33]; unadjusted HR for OS 0.83 [95% CI 0.45-1.54]). Conclusions: No significant interaction between luminal-like subtype and treatment was observed. Patients in the luminal B-like cohort seemed to benefit more from DD schedule. (C) 2020 Elsevier Ltd. All rights reserved.