Methods for the diagnosis and evaluation of stroke and stroke sub-type employ a variety of bio-markers including cellular fibronectin (c-Fn) assembled as a panel for stoke diagnosis and evaluation. Methods are disclosed for selecting markers and correlating their combined levels with a clinical outcome of interest. In various aspects the methods permit early detection and differentiation of stroke subtypes, determination of the prognosis of a patient presenting stroke symptoms, and identification of a patient at risk for early hematoma growth and/or malignant massive cerebral artery infarction. The disclosed methods provide rapid, sensitive and specific assays to greatly increase the number of patients that can receive beneficial stroke treatment and therapy, and to reduce the human and economic costs associated with incorrect stroke diagnosis.