Critical to development of new therapies for Alzheimer’s disease (AD) is the ability to detect clinical or pathological change over time. Clinical outcome measures typically used in therapeutic trials have unfortunately proven to be relatively variable and somewhat insensitive to change in this slowly progressive disease. For this reason, development of surrogate biomarkers that identify significant disease-associated brain changes are necessary to expedite treatment development in AD. Since AD pathology is present in the brain many years prior to clinical manifestation, ideally we want to develop biomarkers of disease that identify abnormal brain structure or function even prior to cognitive decline. Magnetic resonance imaging, fluorodeoxyglucose positron emission tomography, new amyloid imaging techniques, and spinal fluid markers of AD all have great potential to provide surrogate endpoint measures for AD pathology. The Alzheimer’s disease neuroimaging initiative (ADNI) was developed for the distinct purpose of evaluating surrogate biomarkers for drug development in AD. Recent evidence from ADNI demonstrates that imaging may provide more sensitive, and earlier, measures of disease progression than traditional clinical measures for powering clinical drug trials in Alzheimer's disease. This review discusses recently presented data from the ADNI dataset, and the importance of imaging in the future of drug development in AD.