Background: O-(2-[¹⁸F]fluoroethyl)-l-tyrosine positron emission tomography ([¹⁸F]FET PET) scanning is used in routine clinical management and evaluation of gliomas with a recommended 4 h prior fasting. Knowledge of test–retest variation of [¹⁸F]FET PET imaging uptake metrics and the impact of accidental protein intake can be critical for interpretation. The aim of this study was to investigate the repeatability of [¹⁸F]FET-PET metrics and to assess the impact of protein-intake prior to [¹⁸F]FET PET scanning of gliomas. Results: Test–retest variability in the non-protein group was good with absolute (and relative) upper and lower limits of agreement of + 0.15 and − 0.13 (+ 9.7% and − 9.0%) for mean tumour-to-background ratio (TBR_mean), + 0.43 and − 0.28 (+ 19.6% and − 11.8%) for maximal tumour-to-background ratio (TBR_max), and + 2.14 cm³ and − 1.53 ml (+ 219.8% and − 57.3%) for biological tumour volume (BTV). Variation was lower for uptake ratios than for BTV. Protein intake was associated with a 27% increase in the total sum of plasma concentration of the l-type amino acid transporter 1 (LAT1) relevant amino acids and with decreased standardized uptake value (SUV) in both healthy appearing background brain tissue (mean SUV − 25%) and in tumour (maximal SUV − 14%). Oral intake of 24 g of protein 1 h prior to injection of tracer tended to increase variability, but the effects on derived tumour metrics TBR_mean and TBR_max were only borderline significant, and changes generally within the variability observed in the group with no protein intake. Conclusion: The test–retest repeatability was found to be good, and better for TBR_max and TBR_mean than BTV, with the methodological limitation that tumour growth may have influenced results. Oral intake of 24 g of protein one hour before a [¹⁸F]FET PET scan decreases uptake of [¹⁸F]FET in both tumour and in healthy appearing brain, with no clinically significant difference on the most commonly used tumour metrics.