King's College London

Research portal

Supervised clustering for TSPO PET imaging

Research output: Contribution to journalReview articlepeer-review

Julia Schubert, Matteo Tonietto, Federico Turkheimer, Paolo Zanotti-Fregonara, Mattia Veronese

Original languageEnglish
Pages (from-to)257-268
Number of pages12
JournalEuropean Journal of Nuclear Medicine and Molecular Imaging
Issue number1
Early online date29 Mar 2021
E-pub ahead of print29 Mar 2021
PublishedDec 2021

Bibliographical note

Funding Information: Open access funding provided by King's College London. This study was supported in part by the National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London and by the Intramural Research Program of the National Institute of Mental Health, National Institutes of Health (project number ZIAMH002852). Publisher Copyright: © 2021, The Author(s).

King's Authors


PURPOSE: This technical note seeks to act as a practical guide for implementing a supervised clustering algorithm (SVCA) reference region approach and to explain the main strengths and limitations of the technique in the context of 18-kilodalton translocator protein (TSPO) positron emission tomography (PET) studies in experimental medicine.

BACKGROUND: TSPO PET is the most widely used imaging technique for studying neuroinflammation in vivo in humans. Quantifying neuroinflammation with PET can be a challenging and invasive procedure, especially in frail patients, because it often requires blood sampling from an arterial catheter. A widely used alternative to arterial sampling is SVCA, which identifies the voxels with minimal specific binding in the PET images, thus extracting a pseudo-reference region for non-invasive quantification. Unlike other reference region approaches, SVCA does not require specification of an anatomical reference region a priori, which alleviates the limitation of TSPO contamination in anatomically-defined reference regions in individuals with underlying inflammatory processes. Furthermore, SVCA can be applied to any TSPO PET tracer across different neurological and neuropsychiatric conditions, providing noninvasivequantification of TSPO expression.

METHODS: We provide an overview of the development of SVCA as well as step-by-step instructions for implementing SVCA with suggestions for specific settings. We review the literature on SVCAapplications using first- and second- generation TSPO PET tracers and discuss potential clinically relevant limitations and applications.

CONCLUSIONS: The correct implementation of SVCA can provide robust and reproducible estimates of brain TSPO expression. This review encourages the standardisation of SVCA methodology in TSPO PET analysis, ultimately aiming to improve replicability and comparability across study sites.

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454