Accuracy of 18F-fluoroethyltyrosine PET for detecting high grade glioma recurrence: size matters

Background: The joint EANM/EANO/RANO practice guidelines/SNMMI procedure standards for imaging of gliomas using PET with radiolabelled amino acids (Law et al., 2019) represent an important step toward high-quality PET imaging standards. To differentiate treatment-related changes from recurrent disease by means of FET-PET, the guidelines recommend a mean tumour-to-background ratio (TBR) of 2.0. In this study, we evaluated accuracy of this cut-off in a natural cohort of patients with high grade gliomas.

Methods: A retrospective analysis of our clinical data base (years 2017 to 2018) identified 46 consecutive cases with WHO III and IV gliomas, who were referred to FET-PET for differential diagnosis of tumour recurrence and who were operated within 4 weeks (median 1.1 week) following imaging. Due to uncertainty in measuring the size of a purely ring-shape contrast enhancement (CE), only cases with a nodular aspect of CE were selected for further analyses (n=33). The maximal size of CE lesions was measured on T1 images by an experienced radiologist.

Results: On a histopathological examination glioma recurrence was diagnosed in 31 patients. As treatment-related changes were found only in 2 patients, no values of specificity could be calculated. Among the patients with a recurrence a TBR < 2.0 was measured in 9 patients, resulting in a cut-off sensitivity of 71 %. This subgroup showed significantly smaller CE lesions (median 8 mm) relative to the remaining 22 subjects (median 22 mm, p<0.001). TBR significantly correlated with the lesion size (Spearman r=0.62, p<0.001). There was no association with the tumour grade. When the analyses were restricted to subjects with a lesion size of > 10 mm (permutation test), sensitivity increased to 90 %.

Conclusions: The recommended cut-off provides a rather low sensitivity in detecting high grade glioma recurrence, likely due to partial volume effects in smaller lesions. Caution should be undertaken when interpreting FET-PET uptake in lesions of 10 mm size and smaller, i.e. below a double spatial resolution (FWHM) of a PET scanner. Analyses of further cases are under way.