18F-flutemetamol PET imaging demonstrates strong concordance with histopathology irrespective of timing and sequence of exams in prospective and retrospective settings, and shows promise as a valuable tool to study and possibly facilitate diagnosis of Alzheimer’s disease (AD), both in patients with suspected normal pressure hydrocephalus, and among the wider population, according to a pooled analysis of four studies presented at the 2012 Human Amyloid Imaging (HAI) conference in Miami, Fla.
The development of molecular imaging techniques to ‘visualize’ amyloid in vivo represents a major achievement in the study of AD, said study presenter and lead author Juha O. Rinne, MD, from the Turku PET Centre at the Turku University Hospital in Turku, Finland, in a statement.
With this analysis, Rinne and colleagues sought to determine the level of association between uptake of the amyloid PET imaging agent 18F-flutemetamol and the level of beta-amyloid measured by immunohistochemical (IHC) and histochemical (HC) staining in a frontal or parietal cortical region biopsy site.
The researchers enrolled 49 patients with suspected normal pressure hydrocephalus (NPH), prospective (27 patients) or retrospective (22 patients) with 18F-flutemetamol PET and cortical brain biopsy during intracranial pressure measurement or ventriculo-peritoneal shunting. 18F-flutemetamol uptake was quantified using standardized uptake value ratio (SUVR) with cerebellar cortex as a reference region.
They evaluated tissue beta-amyloid using the monoclonal antibody 4G8, Thioflavin-S and Bielschowksy silver stain and an overall pathology result.
Rinne and colleagues found that the 18F-flutemetamol SUVRs from the biopsy sites were significantly associated with biopsy specimen beta-amyloid levels using 4G8 (Pearson’s r=0.41). “There was also good correlation between the biopsy specimen beta-amyloid level and uptake of 18F-flutemetamol in the region contralateral to the biopsy site (r=0.42), or with composite cortical 18F-flutemetamol uptake (r=0.44).”
The blinded visual assessment of images showed a high level of agreement between all readers (r=0.86). Using the overall pathology result as the standard of truth, blinded visual assessment of PET images showed by-reader sensitivities of 100 percent, 71 percent, and 93 percent; specificity was 100 percent for each reader.
Overall, according to the authors, the sensitivity and specificity by majority read were 93 percent and 100 percent, respectively.
The study was funded by GE Healthcare, based in Waukesha, Wis.