Introduction. Differentiation between recurrent glioblastoma (GB) and late treatment-related damage with conventional magnetic resonance imaging (MRI) is difficult since both entities have similar imaging characteristics mimicking tumor progression. In this study, we evaluated the diagnostic accuracy of supplementary O-(2-[18F]-fluoro-Ethyl)-L-tyrosine (18F-FET) PET imaging. Methods. In this retrospective study, 103 patients (68 males and 35 females; median age 60 years (range, 30-80 years) undergone a total of 115 18F-FET PET scans were selected. 18FET-PET scans were performed as 20 min static PET/CT acquisitions and evaluated fused to T1 post-contrast MRI with measurement of maximum tumor-to-brain ratio (TBRmax) and biological tumor volume (BTV). Inclusion criteria were 1) histopathologically proven GB; 2) previous surgical resection/biopsy followed by concomitant and adjuvant radio/chemotherapy; 3) increasing contrast-enhancing lesions on T1 MRI where the differentiation between disease progression or treatment-related damage was uncertain; 4) presentation later than 6 months after radiation therapy; 5) 18FET-PET scan for supplementary assessment; and 6) a histopathological evaluation following re-resection/biopsy. Receiver operating characteristics (ROC) analysis was used to determine the best cutoff value of TBRmax. Overall survival (OS) was defined as the interval from the date of 18F-FET PET scan to date of last clinical and/or MRI follow-up, or date of death. For correlation analysis for TBRmax and OS, the Spearman’s rank correlation coefficient was applied. Results. The median time interval from radiation therapy until radiological progression was 12.5 months (range, 6.0-93.5 months). Subsequent surgical interventions were performed following 88 (76.5%; 76 patients) of 115 18F-FET PET scans. Recurrent GB was confirmed in 83 (93.3%) cases, while treatment-related changes were seen in 5 (5.7%). TBRmax was significant increased in patients with recurrent tumor compared to patients with late treatment-related changes (3.3 ± 0.9 versus 1,9 ± 0.8; p < 0.005), while there was no significant difference in BTV (21.8 ± 27.1 versus 15.5 ± 21.2; p = 0.257). ROC analysis yielded an optimal cut-off value of 2.8 for TBRmax to differentiate between recurrent tumor and late treatment-related damage with a sensitivity of 67%, a specificity of 100%, and an AUC of 0.860. Median OS was 7.5 months (range, 1,5-50.5 months) for all patients. TBRmax was inversely associated with OS in patients (n = 27) who did not undergo subsequent surgical or oncologic treatment (p < 0.0001). Conclusion. The results suggest that static 18F-FET PET may be an useful noninvasive tool to distinguish recurrent tumor from late treatment-related damage in patients with GB.
|Publication status||Published - 2018|
|Event||Scandinavian Society of Neuro-Oncology - Scandic Nidelven, Trondheim, Norway|
Duration: 3 May 2018 → 5 May 2018
|Conference||Scandinavian Society of Neuro-Oncology|
|Period||03/05/2018 → 05/05/2018|