Objective:To assess the diagnostic values of ¹¹C-methionine (MET) PET/CT semiquantitative parameters for detecting recurrence in patients who were diagnosed with suspicious recurrence by MRI after resection of supratentorial gliomas.

Methods:A total of 164 patients (107 males, 57 females, age 6-74 years; high-grade 94, low-grade 63, unclear 7) with supratentorial gliomas who underwent ¹¹C-MET PET/CT between June 2015 and June 2017 in Beijing Tiantan Hospital were enrolled respectively. All patients were with suspicious recurrence after surgery showed by MRI and followed up for 6 months at least. The final diagnosis was determined with histopathological analysis or clinical follow-up. The maximum and mean standardized uptake value (SUV_max and SUV_mean), tumor-to-background ratios (TBR) of SUV_max and SUV_mean (TBR_max and TBR_mean) were recorded and compared between patients with recurrence or without recurrence using independent-sample t test. Receiver operating characteristic (ROC) curves were drawn to determine the threshold, and the diagnostic sensitivity and specificity of each parameter were calculated.

Results:According to the clinical diagnosis, there were 139 patients with recurrence and 25 without recurrence. SUV_max, SUV_mean, TBR_max and TBR_mean were significantly higher for patients with recurrence than those without recurrence (4.19±1.95 vs 2.59±1.18, 2.34±1.08 vs 1.46±0.72, 2.95±1.17 vs 1.83±0.79, 2.64±1.11 vs 1.59±0.71; t values: 5.126-6.183, all P<0.01), but there was no difference in the areas under the ROC curve (AUC) for diagnosis of recurrence with the 4 parameters (z values: 0.265-1.674, all P>0.05), for which the optimal cut-off values were 3.05, 1.65, 1.96 and 1.79, respectively, and the corresponding sensitivities/specificities for the diagnosis of recurrence were 67.6%(94/139)/100%(25/25), 67.6%(94/139)/100%(25/25), 79.9%(111/139)/100%(25/25), 74.8%(104/139)/100%(25/25), respectively. Patients with (n=81) or without (n=13) recurrence had different semiquantitative parameters in high-grade glioma group (t values: 5.137-5.871, all P<0.01), and the AUC for TBR_mean was greater than that for SUV_mean (0.858 vs 0.802; z=1.982, P<0.05). Patients with (n=54) or without (n=9) recurrence in low-grade glioma group showed significant difference in the 4 parameters (t values: 2.730-7.009, all P<0.01), while the AUCs of the 4 parameters were not significantly different (z values: 0.444-1.407, all P>0.05). Among 37 patients with recurrence confirmed by pathology, there were no significant differences in the semiquantitative parameters between the high-grade and low-grade glioma groups and AUCs were not different either (t values: 1.387-1.937, z values: 0.106-1.752, all P>0.05).

Conclusions:Semiquantitative parameters of ¹¹C-MET PET/CT are equally accurate in the differentiation of recurrence from radiation injury in patients with gliomas, while TBR_mean was superior than SUV_mean in patients with the high-grade gliomas. Among the patients with recurrence confirmed by pathology, the value of the semiquantitative parameter is limited for the identification of high- and low- grade gliomas.