Glioma Recurrence Versus Radiation Necrosis: Single-Session Multiparametric Approach Using Simultaneous O-(2-18F-Fluoroethyl)-L-Tyrosine PET/MRI.

PURPOSE: This study aimed to investigate the potential of hybrid gadolinium (Gd)-enhanced F-fluoroethyl-L-tyrosine (F-FET) PET/MRI in distinguishing recurrence from radiation necrosis using simultaneously acquired multiple structural and functional parameters.
METHODS: Twenty-six patients (5 female and 21 male patients; mean ± SD age, 51.58 ± 15.97 years) with single or multiple contrast-enhancing brain lesions (n = 32) on MRI after surgery and radiation therapy were evaluated with simultaneously acquired Gd-enhanced F-FET PET/MRI. They were then followed up with resurgery and histopathological diagnosis (n = 9) and/or clinical/MRI- or PET/MRI-based imaging follow-up (n = 17). PET/MR images were analyzed using manually drawn regions of interest over areas of maximal contrast enhancement and/or FET uptake. Maximum target-to-background ratio (TBRmax), mean target-to-background ratio (TBRmean), and choline-to-creatine (Cho/Cr) ratios as well as normalized mean relative cerebral blood volume (rCBVmean) and mean apparent diffusion coefficient (ADCmean) were determined. The accuracy of each parameter individually and in various possible combinations for differentiating recurrence versus radiation necrosis was evaluated using 2-tailed independent samples Student t test, multivariate analysis of variance, and multivariate receiver operating characteristic analysis. Positive histopathological finding and long-term imaging/clinical follow-up suggestive of disease progression served as criterion standard.
RESULTS: Of 26 patients, 19 were classified as recurrence, with 7 patients showing radiation necrosis. Individually, TBRmax, TBRmean, ADCmean, and Cho/Cr ratios as well as normalized rCBVmean was significant in differentiating recurrence from radiation necrosis, with an accuracy of 93.8% for TBRmax, 87.5% for TBRmean, 81.3% for ADCmean, 96.9% for Cho/Cr ratio, and 90.6% for normalized rCBVmean. The accuracy of both normalized rCBVmean and ADCmean was improved in combination with TBRmax or Cho/Cr ratio. However, TBRmax (or TBRmean) with Cho/Cr ratio yielded the highest accuracy, approaching up to 97%. Furthermore, maximum area under the curve is achieved with the combination of TBRmean, CBV, and Cho/Cr values.
CONCLUSIONS: Our findings suggest that FET uptake with Cho/Cr ratio and normalized rCBVmean could be most useful to distinguish primary glioma recurrence from radiation necrosis. Hybrid simultaneous multiparametric F-FET PET/MRI might play a significant role in the evaluation of patients with suspected glioma recurrence.