# Developing a novel reference region for [18F]PI-2620-PET imaging to facilitate the assessment of 4-repeat tauopathies

**Authors:** Lukas Frontzkowski, Johannes Gnörich, Mattes Gross, Amir Dehsarvi, Sebastian N. Roemer-Cassiano, Carla Palleis, Sabrina Katzdobler, Anna Dewenter, Anna Steward, Davina Biel, Fabian Hirsch, Zeyu Zhu, Johannes Levin, Andrew W. Stephens, Andre Müller, Norman Koglin, Gérard N. Bischof, Gabor G. Kovacs, Günter U. Höglinger, Matthias Brendel, Nicolai Franzmeier

PMC · DOI: 10.1007/s00259-025-07396-8 · European Journal of Nuclear Medicine and Molecular Imaging · 2025-06-10

## TL;DR

This study develops a new reference region for PET imaging to better assess 4R tauopathies like PSP, improving diagnosis and monitoring.

## Contribution

A data-driven temporo-orbital white matter reference region is proposed for [18F]PI-2620 PET imaging in 4R tauopathies.

## Key findings

- The temporo-orbital white matter reference outperformed cerebellar references in diagnosing PSP.
- PSP patients showed significantly higher [18F]PI-2620 signal in basal ganglia using the new reference.
- The novel reference showed stronger associations between PET signal and clinical disease severity.

## Abstract

Progressive supranuclear palsy (PSP) is a fatal 4-repeat (4R) tauopathy with progressive movement phenotypes. In-vivo 4R tau biomarkers are therefore crucial for PSP diagnosis, monitoring, and treatment evaluation. The tau-PET tracer [18F]PI-2620 binds to 4R tau and shows increased uptake in PSP-associated regions (e.g., globus pallidus), and is therefore a candidate 4R tau biomarker. However, commonly used cerebellar tau-PET reference regions show regional proximity to cerebellar 4R tau deposits in PSP, confounding semiquantitative [18F]PI-2620 assessments. Therefore, we employed bias-free image-derived input function (IDIF) PET quantification to identify an optimized data-driven reference region for assessing 4R tau in PSP.

Dynamic [18F]PI-2620 PET (60 min) was acquired in 58 PSP-Richardson Syndrome (PSP-RS) and 18 healthy controls (HC). IDIF-modelling with carotid timeseries derived total distribution volume (VT). Iteratively normalizing VT images to atlas-based white matter (WM), we identified reference candidates maximizing PSP-RS vs. HC pallidum differences. The best-performing WM references were combined to a temporo-orbital WM reference, validated in PSP-nonRS (n = 54), HC (n = 18), and disease controls (α-synucleinopathies, n = 21; Alzheimer’s disease (AD, n = 22) using VT-ratios (VTr) and 20-40min static standardized uptake value ratios (SUVr).

Using the data-driven temporo-orbital WM reference, PSP patients showed significantly higher basal ganglia [18F]PI-2620 signal vs. HC compared to cerebellar normalization. Receiver operating curve (ROC) analysis confirmed higher diagnostic accuracy using the temporo-orbital WM reference. Pallidum [18F]PI-2620 showed significant associations with clinical disease severity exclusively when using the novel temporo-orbital WM reference.

A data-driven temporo-orbital WM reference optimizes [18F]PI-2620 PET assessment for PSP diagnosis, outperforming conventional cerebellar references used in tau-PET imaging.

The online version contains supplementary material available at 10.1007/s00259-025-07396-8.

## Linked entities

- **Chemicals:** [18F]PI-2620 (PubChem CID 145722629)
- **Diseases:** Progressive supranuclear palsy (MONDO:0019037), Alzheimer’s disease (MONDO:0004975)

## Full-text entities

- **Diseases:** tauopathies (MESH:D024801)
- **Chemicals:** [18F]PI-2620 (MESH:C000710692)

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12589379/full.md

## Figures

6 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12589379/full.md

## References

5 references — full list in the complete paper: https://tomesphere.com/paper/PMC12589379/full.md

---
Source: https://tomesphere.com/paper/PMC12589379