Background: Advanced ER+ breast cancer patients have reported prolongation of stable disease when treated with a CDK4/6 inhibitor as a monotherapy or in combination with endocrine treatment. However ~20% of patients are intrinsically resistant and all patients eventually acquire resistance to these therapies. There is a critical need to identify biomarkers that accurately predict response and resistance to CDK4/6 inhibitors. ER-positivity, luminal patterns of gene expression, Rb function, overexpression of cyclin D1, cyclin E, CDK6 and low levels of p16 are biomarkers that do not accurately match clinical outcomes. Methods: We performed a retrospective study analyzing plasma-based metabolites from a baseline (pre-dose) and ~2 months post treatment of 21 women with estrogen-receptor-positive (ER+) metastatic breast cancer treated with CDK4/6 inhibitors. Results: By correlating the metabolite expression profiles to clinical outcomes we were able to identify a metabolic signature that could differentiate the CDK4/6 responders and resistant patients with a predictive accuracy of > 90%. Further we were able to identify independent signatures predictive of response for individual CDK4/6 inhibitors palbociclib and ribociclib. Conclusions: The results of this study could lead to a paradigm shift in the administration of CDK4/6 inhibitors wherein prior to treatment and during treatment patient plasma is screened to determine whether that individual patient is responsive or resistant to a CDK4/6 inhibitor.