Image Quality and Semiquantitative Measurements on the Biograph Vision PET/CT System: Initial Experiences and Comparison with the Biograph mCT

In May 2018, the Biograph Vision PET/CT system was installed at the University Medical Center Groningen. This study evaluated the initial experiences with this new PET/CT system in terms of perceived image quality and semiquantitative analysis in comparison to the Biograph mCT as a reference. Methods: In total, 20 oncologic patients were enrolled and received a single 3 MBq/kg injected dose of 18F-FDG followed by a dual-imaging PET scan. Ten patients were scanned on the Biograph mCT first, whereas the other 10 patients were scanned on the Biograph Vision first. The locally preferred clinically reconstructed images were blindly reviewed by 3 nuclear medicine physicians and scored (using a Likert scale of 1-5) on tumor lesion demarcation, overall image quality, and image noise. In addition, these clinically reconstructed images were used for semiquantitative analysis by measurement of SUVs in tumor lesions. Images acquired using reconstructions conform with the European Association of Nuclear Medicine Research Ltd. (EARL) specifications were also used for measurements of SUV in tumor lesions and healthy tissues for comparison between systems. Results: The 18F-FDG dose received by the 14 men and 6 women (age range, 36-84; mean ± SD, 61 ± 16 y) ranged from 145 to 405 MBq (mean ± SD, 268 ± 59.3). Images acquired on the Biograph Vision were scored significantly higher on tumor lesion demarcation, overall image quality, and image noise than images acquired on the Biograph mCT (P < 0.001). The overall interreader agreement showed a Fleiss κ of 0.61 (95% confidence interval, 0.53-0.70). Furthermore, the SUVs in tumor lesions and healthy tissues agreed well (within 95%) between PET/CT systems, particularly when EARL-compliant reconstructions were used on both systems. Conclusion: In this initial study, the Biograph Vision showed improved image quality compared with the Biograph mCT in terms of lesion demarcation, overall image quality, and visually assessed signal-to-noise ratio. The 2 systems are comparable in semiquantitatively assessed image biomarkers in both healthy tissues and tumor lesions. Improved quantitative performance may, however, be feasible using the clinically optimized reconstruction settings.