Predicting prognosis in metastatic HR+/HER2-negative disease might be of clinical value. Here, we developed and validated a prognostic biomarker in 765 patients (pts) recruited in 2 phase III trials evaluating endocrine-based therapies.
PAM50 and clinical data were available from 821 pts treated with letrozole+placebo/lapatinib in the first-line setting (EGF3008). Among them, 484 pts were selected based on HER2-negative disease and no prior endocrine therapy or relapse ≥6 months since tamoxifen discontinuation. To derive a prognostic model, the following variables were evaluated: 1) PAM50 subtypes, signatures and genes 2) ECOG, 4) visceral disease, 5) num. metastasis, 6) biopsy-type and 7) age. Using the variables, a progression-free survival (PFS) cox model was evaluated in 2/3 pts using Elastic Net (Monte Carlo CV). C-index of each model was estimated in 1/3 pts. The final model was tested in 261 pts treated with exemestane+placebo/everolimus (BOLERO2). PAM50 was performed in FFPE tumors (∼80% primary).
In EGF3008, prognostic models that integrated PAM50 and clinical variables yielded superior C-index values compared to models with PAM50-only or clinical variables-only. The final model (PAM50MET) combined 21 variables, including 2 PAM50 subtypes, the Basal signature, 14 genes and 4 clinical variables. In EGF3008, the optimized cutpoint was associated with PFS (hazard ratio=0.41; p < 0.0001) and overall survival (OS; hazard ratio=0.41; p < 0.0001). In BOLERO2, PAM50MET score was associated with PFS (p = 0.004) and OS (p < 0.0001). Using the same cutpoint, PAM50MET-low was associated with better PFS (hazard ratio=0.72; p = 0.028) and OS (hazard ratio=0.51; p < 0.0001). The median PFS and OS in PAM50MET-low was 6.9 and 36.5 months compared to 5.2 and 23.4 months in PAM50MET-high.
PAM50MET could help identify pts with HR+/HER2-negative metastatic disease candidates, especially in the first-line setting, for endocrine therapy-only vs. endocrine therapy + CDK4/6 inhibitor vs. new treatment strategies. Further validation of PAM50MET in pivotal clinical trials that have evaluated endocrine-based therapies is justified.
NCT00073528; NCT00863655.
Aleix Prat.
These trials was sponsored by Novartis Pharmaceuticals Corporation. This work was also supported by funds from the Spanish Society of Medical Oncology (to A.P.), Instituto de Salud Carlos III - PI13/01718 (to A.P.), Pas a Pas (to A.P.), Save the Mama (to A.P.), Instituto de Salud Carlos III - PI16/00904 (to A.P.), a Career Catalyst Grant (CCR13261208) from the Susan Komen Foundation (to A.P.), and Premio Jóven Investigador de la Fundación AstraZeneca (to A.P.).
A. Prat: Research funding: NanoString Technologies, GlaxoSmithKline, Susan G. Komen Foundation; Consulting/advisory relationship: NanoString Technologies, Novartis; Scientific advisory board: Novartis Pharmaceuticals Corporation, Pfizer, Roche. J.C. Brase: Employee: Novartis. S.R.D. Johnston: Consulting or advisory role: Eli Lilly, AstraZeneca, Puma Biotechnology; Speakers’ bureau: Pfizer, Novartis; Research funding: Pfizer (Inst). E.M. Ciruelos: Honoraria for consultancy and speaker: Lilly, Roche, Novartis, Pfizer. J.S. Parker: Co-inventor on patents related to the PAM50, licensed to NanoString Technologies, Inc. All other authors have declared no conflicts of interest.