Biomarkers that predict responses to oral corticosteroids (OCS) facilitate patient selection for asthma treatment. We hypothesised that asthma patients would respond differently to OCS therapy, with biomarkers and inflammometry predicting response.Adults with stable asthma underwent a randomised controlled cross-over trial of 50 mg prednisolone daily for 10 days (n=55). A six-gene expression biomarker signature (, ,, , and ) in induced sputum, and eosinophils in blood and sputum were assessed and predictors of response were investigated (changes in forced expiratory volume in 1 s (ΔFEV), six-item Asthma Control Questionnaire score (ΔACQ6) or exhaled nitric oxide fraction (Δ)).At baseline, responders to OCS (n=25) had upregulated mast cell gene expression, poorer lung function, and higher sputum and blood eosinophils. Following treatment, and gene expression was reduced, whereas and expression remained unchanged. Receiver operating characteristic (ROC) analysis showed the six-gene expression biomarker signature as a better predictor of clinically significant responses to OCS than blood and sputum eosinophils.The six-gene expression signature including eosinophil and Th2 related mast cell biomarkers showed greater precision in predicting OCS response in stable asthma. Thus, a novel sputum gene expression signature highlights an additional role of mast cells in asthma, and could be a useful measurement to guide OCS therapy in asthma.