Differentiation of treatment-related changes from tumour progression: a direct comparison between dynamic FET PET and ADC values obtained from DWI MRI

BackgroundFollowing brain cancer treatment, the capacity of anatomical MRI to differentiate neoplastic tissue from treatment-related changes (e.g., pseudoprogression) is limited. This study compared apparent diffusion coefficients (ADC) obtained by diffusion-weighted MRI (DWI) with static and dynamic parameters of O-(2-[F-18]fluoroethyl)-L-tyrosine (FET) PET for the differentiation of treatment-related changes from tumour progression.Patients and methodsForty-eight pretreated high-grade glioma patients with anatomical MRI findings suspicious for progression (median time elapsed since last treatment was 16weeks) were investigated using DWI and dynamic FET PET. Maximum and mean tumour-to-brain ratios (TBRmax, TBRmean) as well as dynamic parameters (time-to-peak and slope values) of FET uptake were calculated. For mean ADC calculation, regions-of-interest analyses were performed on ADC maps calculated from DWI coregistered with the contrast-enhanced MR image. Diagnoses were confirmed neuropathologically (21%) or clinicoradiologically. Diagnostic performance was evaluated using receiver-operating-characteristic analyses or Fisher's exact test for a combinational approach.ResultsTen of 48 patients had treatment-related changes (21%). The diagnostic performance of FET PET was significantly higher (threshold for both TBRmax and TBRmean, 1.95; accuracy, 83%; AUC, 0.890.05; P<0.001) than that of ADC values (threshold ADC, 1.09x10(-3)mm(2)/s; accuracy, 69%; AUC, 0.73 +/- 0.09; P=0.13). The addition of static FET PET parameters to ADC values increased the latter's accuracy to 89%. The highest accuracy was achieved by combining static and dynamic FET PET parameters (93%). Moreover, in contrast to ADC values, TBRs <1.95 at suspected progression predicted a significantly longer survival (P=0.01).Conclusions Data suggest that static and dynamic FET PET provide valuable information concerning the differentiation of early treatment-related changes from tumour progression and outperform ADC measurement for this highly relevant clinical question.