Key points are not available for this paper at this time.
Abstract Purpose 2-18FFluoro-2-deoxy-D-glucose (FDG)–positron emission tomography (PET)/computed tomography (CT) has been suggested as an imaging modality to diagnose polymyalgia rheumatica (PMR). However, the applicability of FDG-PET/CT remains unclear, especially following glucocorticoid administration. This study aimed to investigate the diagnostic accuracy of FDG-PET/CT before and during prednisolone treatment, as well as following short-term prednisolone discontinuation. Methods Treatment naïve suspected PMR patients were clinically diagnosed at baseline and subsequently had an FDG-PET/CT performed. Patients diagnosed with PMR were administered prednisolone following the first FDG-PET/CT and had a second FDG-PET/CT performed after 8 weeks of treatment. Subsequently, prednisolone was tapered with short-term discontinuation at week 9 followed by a third FDG-PET/CT at week 10. An FDG-PET/CT classification of PMR/non-PMR was applied, utilizing both the validated Leuven score and a dichotomous PMR score. The final diagnosis was based on clinical follow-up after 1 year. Results A total of 68 and 27 patients received a final clinical diagnosis of PMR or non-PMR. A baseline FDG-PET/CT classified the patients as having PMR with a sensitivity/specificity of 86%/63% (Leuven score) and 82%/70% (dichotomous score). Comparing the subgroup of non-PMR with inflammatory diseases to the PMR group demonstrated a specificity of 39%/54% (Leuven/dichotomous score). After 8 weeks of prednisolone treatment, the sensitivity of FDG-PET/CT decreased to 36%/41% (Leuven/dichotomous score), while a short-term prednisolone discontinuation increased the sensitivity to 66%/60%. Conclusion FDG-PET/CT has limited diagnostic accuracy for differentiating PMR from other inflammatory diseases. If FDG-PET/CT is intended for diagnostic purposes, prednisolone should be discontinued to enhance diagnostic accuracy. Trial registration ClinicalTrials.gov (NCT04519580). Registered 17th of August 2020.
Nielsen et al. (Tue,) studied this question.