Open Access Highly Accessed Original research

Assessment of acquisition protocols for routine imaging of Y-90 using PET/CT

Thomas Carlier12*, Thomas Eugène1, Caroline Bodet-Milin12, Etienne Garin3, Catherine Ansquer12, Caroline Rousseau24, Ludovic Ferrer24, Jacques Barbet2, Frédéric Schoenahl5 and Françoise Kraeber-Bodéré124

Author Affiliations

1 Nuclear Medicine Department, University Hospital of Nantes, Place Alexis Ricordeau, Nantes, 44093, France

2 Nantes-Angers Cancer Research Center, UMR 892 INSERM and UMR 6299 CNRS, 8 quai Moncousu BP 70721, Nantes, 44007, France

3 Nuclear Medicine Department, Cancer Center, Rue de la Bataille Flandres Dunkerque, Rennes, 35042, France

4 Nuclear Medicine Department, Integrated Center of Oncology, Boulevard Jacques Monod, Saint-Herblain, 44805, France

5 Molecular Imaging, Siemens Healthcare, Hartmnannstrasse 16, Erlangen, 91052, Germany

For all author emails, please log on.

EJNMMI Research 2013, 3:11  doi:10.1186/2191-219X-3-11

Published: 17 February 2013



Despite the early theoretical prediction of the 0+-0+ transition of 90Zr, 90Y-PET underwent only recently a growing interest for the development of imaging radioembolization of liver tumors. The aim of this work was to determine the minimum detectable activity (MDA) of 90Y by PET imaging and the impact of time-of-flight (TOF) reconstruction on detectability and quantitative accuracy according to the lesion size.


The study was conducted using a Siemens Biograph® mCT with a 22 cm large axial field of view. An IEC torso-shaped phantom containing five coplanar spheres was uniformly filled to achieve sphere-to-background ratios of 40:1. The phantom was imaged nine times in 14 days over 30 min. Sinograms were reconstructed with and without TOF information. A contrast-to-noise ratio (CNR) index was calculated using the Rose criterion, taking partial volume effects into account. The impact of reconstruction parameters on quantification accuracy, detectability, and spatial localization of the signal was investigated. Finally, six patients with hepatocellular carcinoma and four patients included in different 90Y-based radioimmunotherapy protocols were enrolled for the evaluation of the imaging parameters in a clinical situation.


The highest CNR was achieved with one iteration for both TOF and non-TOF reconstructions. The MDA, however, was found to be lower with TOF than with non-TOF reconstruction. There was no gain by adding TOF information in terms of CNR for concentrations higher than 2 to 3 MBq mL−1, except for infra-centimetric lesions. Recovered activity was highly underestimated when a single iteration or non-TOF reconstruction was used (10% to 150% less depending on the lesion size). The MDA was estimated at 1 MBq mL−1 for a TOF reconstruction and infra-centimetric lesions. Images from patients treated with microspheres were clinically relevant, unlike those of patients who received systemic injections of 90Y.


Only one iteration and TOF were necessary to achieve an MDA around 1 MBq mL−1 and the most accurate localization of lesions. For precise quantification, at least three iterations gave the best performance, using TOF reconstruction and keeping an MDA of roughly 1 MBq mL−1. One and three iterations were mandatory to prevent false positive results for quantitative analysis of clinical data.

Trial registration

http://IDRCB 2011-A00043-38 P101103 webcite

90-Y; PET; Quantification; Detectability; Radioembolization