A major limitation in finding therapeutic solutions for Alzheimer's disease (AD) has been the lack of a reliable method for early diagnosis of this devastating disease. Besides the development of biomarkers in biological fluids of patients, the search for a pathology-specific neuroimaging tools is critical at the present stage in which almost 30 million people suffer this disease worldwide. Several interesting approaches have been developed, however their clinical impact has been low. One of the difficulties has been to find the proper molecular tracers to specifically tag pathognomonic lesions in AD brain, including not only amyloid aggregates but also filaments of the modified microtubule-associated protein tau. In this review, we analyze the evidence towards developing pathology-specific diagnostic tools for AD. We analyze the current evidence and clinical implications of new imaging technologies for AD, and how tau hypothesis and the amyloid cascade hypothesis will impact on these scientific efforts in the near future.