Cerebral neurofibrillary tangles (NFTs) accumulate in a predictable sequence decades before the clinical symptoms of Alzheimer's disease emerge, and the degree of tangle degeneration correlates with the severity of cognitive impairment. A valid in vivo marker of tangle burden, therefore, would be useful for presymptomatic and symptomatic disease detection and treatment monitoring. Recent advances using positron emission tomography (PET) indicate the feasibility of in vivo imaging that provides a combined signal of both neurofibrillary tangles and senile plaques. Such results are encouraging that a tangle-specific marker will be found; however, several methodological issues first need to be addressed, including scanner spatial resolution in the relatively small brain regions where tangles accumulate. NFT-specific imaging probes will need to be lipophilic in order to cross the blood-brain barrier and neuronal membranes and have a high binding affinity to NFTs with minimal nonspecific binding, which would result in a high signal-to-background ratio in PET images.