Multicentric ^sup 68^Ge phantom-based evaluation of inter-scanner variability among several Discovery IQ PET/CT scanners

• Published in: The Journal of nuclear medicine (1978) Vol. 58; p. 437
• Main Authors: De Ponti, Elena, Spadavecchia, Chiara, Vallot, Delphine, Morzenti, Sabrina, Pieczonka, Anna, Buraczewska, Katarzyna, Reynes-Llompart, Gabriel, Caselles, Olivier
• Format: Journal Article
• Language: English
• Published: New York Society of Nuclear Medicine 01.05.2017
• Subjects: Accuracy; Algorithms; Computed tomography; Evaluation; Health care; Intelligence; Isotopes; Medical imaging; Nuclear medicine; Positron emission; Positron emission tomography; Reconstruction; Scanners; Tomography; Variability
• ISSN: 0161-5505; 1535-5667

Objectives: Evaluation of accurate and reproducible SUV measurements is an important issue in clinical PET imaging for assessing therapy response for reliable detection of lesion's uptake changes over time. Aim of this study was to evaluate the inter-scanner variability among several PET/CT scanners affecting tracer uptake quantification. Methods: A pre-filled 68Ge-NEMA phantom based on a standard NEMA IQ phantom body (Data Spectrum - sphere to background activity ratio 4:1) was circulated across 5 centres all equipped with a Discovery IQ PET/CT scanner (GE Healthcare, Milwaukee, WI, USA). Acquisitions were performed 1) following NEMA-NU-2 procedures; each centre adjusted acquisition time according to the age of the phantom to meet NEMA-NU-2 prescription and 2) according to each site clinical protocol. The NEMA-NU-2 contrast recovery coefficient (CR) and background variability (BKG) from standard OSEM and regularized reconstruction (RR) algorithm were evalulated. Results: The highest variability was found in small-sized spheres: CR differences among the centres decreased systematically as the sphere dimension increased. CR ranged from 26.97 to 31.50 (10 mm sphere) and from 71.73 to 75.08 (37 mm sphere) for the OSEM NEMA reconstruction, while from 34.42 to 42.86 (10 mm sphere) and from 82.23 to 83.75 (37 mm sphere) for the RR NEMA one. Clinical reconstruction protocols gave higher variability: CR ranged from 15.64 to 30.68 (10 mm sphere) and from 64.51 to 73.16 (37 mm sphere) for the OSEM clinical reconstruction, while from 13.78 to 20.90 (10 mm sphere) and from 72.08 to 76.78 (37 mm sphere) for the RR clinical one. BKG ranged as follows: from 7.92 to 8.55 (10 mm sphere) and from 2.27 to 2.56 (37 mm sphere) for the OSEM NEMA reconstruction, while from 7.96 to 9.68 (10 mm sphere) and from 2.16 to 2.79 (37 mm sphere) for the RR NEMA one. Using clinical protocols BKG ranged from 3.16 to 4.08 (10 mm sphere) and from 1.45 to 1.92 (37 mm sphere) for the OSEM clinical reconstruction, while from 2.92 to 3.58 (10 mm sphere) and from 1.35 to 1.67 (37 mm sphere) for the RR clinical one. Conclusion: The use of a long half-life isotope phantom circulating across several centres permitted to remove phantom filling variability thus being a reliable tool to evaluate inter-scanner variability in a multicentre prospective. RR algorithm provided a smaller variability for CR and BKG evaluation performed with the same scanner in different centers.