# Underestimated Centiloid Values in Amyloid PET: A Technical Report on Clinically Relevant Quantification Errors

## TL;DR

This paper highlights how amyloid PET scans can underestimate amyloid levels, leading to incorrect clinical decisions about Alzheimer's treatments.

## Contribution

The paper identifies a specific technical error in Centiloid quantification due to reference region selection in amyloid PET scans.

## Key findings

- Using the whole cerebellum as a reference region can lower Centiloid values by about 7%.
- This underestimation may lead to patients being incorrectly denied anti-amyloid therapy.
- Visual assessment should be used alongside Centiloid values to avoid misjudgments.

## Abstract

Quantitative amyloid positron emission tomography (PET) interpretation using the Centiloid scale is increasingly adopted worldwide to guide eligibility and continuation of anti-amyloid monoclonal antibody therapy. However, discrepancies between visual and quantitative assessments occasionally occur, potentially leading to critical misjudgments in clinical decision-making. We present a representative 18F-florbetapir case in which cortical amyloid deposition was visually evident, yet the calculated Centiloid value was 0, falsely indicating a negative scan. This underestimation likely results from reference region selection: using the whole cerebellum (including white matter) lowers standardized uptake value ratios by approximately 7% compared with cerebellar gray matter, thereby decreasing Centiloid values. Consequently, patients with substantial amyloid burden may be incorrectly deemed ineligible for initiation or continuation of anti-amyloid therapy. Clinicians should therefore interpret Centiloid-based quantification with caution, always corroborating it with expert visual reads. Harmonization of reference region definitions and standardized reporting are urgently needed to prevent inappropriate treatment decisions and ensure the safe, effective use of disease-modifying therapies in Alzheimer's disease.