Multicentre quantitative (68)Ga PET/CT performance harmonisation

PURPOSE: Performance standards for quantitative (18)F-FDG PET/CT studies are provided by the EANM Research Ltd. (EARL) to enable comparability of quantitative PET in multicentre studies. Yet, such specifications are not available for (68)Ga. Therefore, our aim was to evaluate (68)Ga-PET/CT quantification variability in a multicentre setting. METHODS: A survey across Dutch hospitals was performed to evaluate differences in clinical (68)Ga PET/CT study protocols. (68)Ga and (18)F phantom acquisitions were performed by 8 centres with 13 different PET/CT systems according to EARL protocol. The cylindrical phantom and NEMA image quality (IQ) phantom were used to assess image noise and to identify recovery coefficients (RCs) for quantitative analysis. Both phantoms were used to evaluate cross-calibration between the PET/CT system and local dose calibrator. RESULTS: The survey across Dutch hospitals showed a large variation in clinical (68)Ga PET/CT acquisition and reconstruction protocols. (68)Ga PET/CT image noise was below 10%. Cross-calibration was within 10% deviation, except for one system to overestimate (18)F and two systems to underestimate the (68)Ga activity concentration. RC-curves for (18)F and (68)Ga were within and on the lower limit of current EARL standards, respectively. After correction for local (68)Ga/(18)F cross-calibration, mean (68)Ga performance was 5% below mean EARL performance specifications. CONCLUSIONS: (68)Ga PET/CT quantification performs on the lower limits of the current EARL RC standards for (18)F. Correction for local (68)Ga/(18)F cross-calibration mismatch is advised, while maintaining the EARL reconstruction protocol thereby avoiding multiple EARL protocols.