The distinctive cortical uptake of the tracer (18)F-FDDNP (2-(1-{6-[(2-fluoroethyl(methyl)amino]-2-naphthyl}ethylidene)malononitrile) in Alzheimer's disease (AD) is believed to be because of its binding to both neurofibrillary tangles (NFTs) and highly fibrillar senile plaques. We therefore investigated the binding of a tracer concentration of (3)H-FDDNP to brain sections containing AD hallmark pathologies. Semi-adjacent sections were labelled with (3)H-PIB (Pittsburgh compound-B, 2-[4'-(methylamino)phenyl]-6-hydroxybenzothiazole) and (14)C-SB13 (4-N-methylamino-4'-hydroxystilbene) for comparison. Neocortical sections containing widespread senile plaques and cerebrovascular amyloid angiopathy, produced a sparse and weak labelling following incubation with (3)H-FDDNP. Furthermore, in sections containing NFTs, there was no overt labelling of the pathology by (3)H-FDDNP. In contrast, sections labelled with (3)H-PIB displayed extensive labelling of diffuse plaques, classical plaques, cerebrovascular amyloid angiopathy and NFTs. (14)C-SB13 produced a broadly similar binding pattern to PIB. Radioligand binding assays employing in vitro generated amyloid-beta peptide fibrils demonstrated a approximately 10-fold reduced affinity for (3)H-FDDNP (85.0 +/- 2.0 nM) compared with (3)H-PIB (8.5 +/- 1.3 nM). These data provide an alternative mechanistic explanation for the observed low cortical uptake of (18)F-FDDNP in AD; in that the ligand is only weakly retained by the hallmark neuropathology because of its low affinity for amyloid structures.