Comparative analysis of visual assessment, Centiloid quantification, and clinical follow-up in patients with suspected Alzheimer's disease with [18F]Florbetaben PET imaging

Abstract

Objetive

The detection of amyloid deposition has become a pivotal element in the diagnostic workup of Alzheimer's disease (AD). Among the available PET radiotracers, [¹⁸F]Florbetaben has proven to be a reliable tool for identifying amyloid-beta plaques, validated through both clinical assessments and histopathological correlation. This study aimed to assess the diagnostic accuracy and reliability of two key imaging methods: visual interpretation and semi-quantitative analysis using the Centiloid (CL) scale.

Methods

We included 208 patients who were attended at the outpatient Memory Clinic of a tertiary hospital, with mild cognitive impairment or mild dementia, suspected of AD. We performed receiver-operating-characteristic (ROC) analyses to compare the discriminative power of visual and quantification between AD and non-AD patients, characterized through the previously established standard of truth (SoT) as clinical follow-up.

Results

A total of 198 cases (95.0%) demonstrated concordance between both diagnostic methods (visual and CL-based quantification) and the SoT, with a weighted kappa coefficient k = 0.903 (p < 0.001), indicating almost perfect agreement. When comparing visual diagnosis to the SoT, the sensitivity was 96.8%, specificity was 96.5%, and diagnostic accuracy was 96.6%. Both approaches exhibited excellent diagnostic performance, effectively correlating with the final clinical diagnoses.

Conclusion

Our findings highlight that both visual assessment and semi quantitative analysis are reliable tools for the detection and monitoring of amyloid deposition in AD, particularly for early diagnosis and disease progression tracking.