Objective:The recurrence or progression under endocrine therapy in hormone receptor-positive (HR+)advanced breast cancer (ABC) remained a critical clinical challenge.Chidamide is an oral subtype-selective histone deacetylase (HDAC) inhibitor with multiple functions in tumor growth inhibition and microenvironment modulation via epigenetic reprogramming.The purpose of this study was to evaluate the safety,pharmacokinedcs (PK),and preliminary efficacy of chidamide in combination with exemestane in HR+ ABC patients.Methods:Eligible parents were postnenopausal women with HR+ ABC recurrent or progressed to at least one endocrine therapy.Blood samples were obtained in the run-in period and the first day of combination treatment for PK analysis.In combination treatment,parents were given exemestane 25 mg daily and chidamide 30 mg twice a week (BIW) until progression of disease or intolerable toxicities.A treatment cycle was defined as 4 weeks.Safety,PK parameters,and preliminary efficacy were evaluated.Results:A total of 20 patients were enrolled between July and December,2015.The median number of treatments cycle was 5.2 (20.8 weeks) with 2 patients still on treatment at the data cut-off date of October,2017.The treatment-related adverse events (AE) ≥ grade 3 in more than 2 patients were neutropenia (35％),thrombocytopenia (30％),and leucopenia (20％).The plasma exposure of exemestane was consistent in the presence or absence of chidamide.A slight increase in chidamide exposure was noted in the presence of exemestane,probably due to the inter-and intra-patient variations.The best response in 16 evaluable parents was assessed by Response Evaluation Criteria in Solid Tumors (RECIST),including 4 parents with partial response,10 patients with stable disease.The median progression-free survival (PFS) was 7.6 months.Conclusions:The combination of chidamide with exemestane was generally well tolerated with promising preliminary efficacy in HR+ ABC parents.The overall results from this study encourage further pivotal trial in this patient population.