Tracer uptake in relation to histological grade
The median MET T/N ratios in grade Ⅱ (n=12), grade Ⅲ (n=24), and grade Ⅳ (n=45) gliomas were 2.51 (interquartile range [IQR] 1.82-3.13), 5.03 (3.58-6.39), and 5.99 (4.64-7.39), respectively (Table 2). There were significant differences in the median MET T/N ratios between grade Ⅱ and Ⅲ gliomas (P<0.001) and grade Ⅱ and Ⅳ gliomas (P<0.001), but not between grade Ⅲ and Ⅳ gliomas. The median FLT T/N ratios in grade Ⅱ, grade Ⅲ, and grade Ⅳ gliomas were 2.24 (IQR 1.69-2.62), 6.25 (4.59-8.69), and 15.24 (11.11-20.49), respectively. There were significant differences in the median FLT T/N ratios between grade Ⅱ and Ⅲ gliomas (P<0.001) and grade Ⅲ and Ⅳ gliomas (P<0.001).
Tracer uptake in relation to histopathological classification
The median MET T/N ratios in DA (n=9), OD (n=3), AA (n=14), AO (n=10), and GBM (n=45) were 2.03 (IQR 1.83-2.97), 2.86 (2.51-3.23), 4.65 (3.25-5.99), 5.07 (4.72-7.21), and 5.99 (4.64-7.39), respectively (Table 2). OD was omitted from the statistical analysis because of the small number of cases. There were significant differences of the median MET T/N ratios between DA and AA (P=0.016), DA and AO (P=0.009), and DA and GBM (P<0.001; Fig. 1A). The median FLT T/N ratios in DA, OD, AA, AO and GBM were 2.35 (IQR 1.69-2.60), 2.14 (1.91-3.42), 7.16 (4.77-9.41), 5.82 (4.26-6.79), and 15.24 (11.11-20.46), respectively. There were significant differences in the median FLT T/N ratios between DA and AA (P=0.002), DA and AO (P=0.005), DA and GBM (P<0.001), AA and GBM (P=0.002), and AO and GBM (P<0.001; Fig. 1B).
Tracer uptake in relation to IDH1 mutation status
In all gliomas, the median MET T/N ratios in IDH1-mutant and IDH1-wildtype tumours were 3.6 (IQR 2.84-5.59) and 5.91 (4.57-7.35), respectively. The median MET T/N ratio in IDH1-wildtype tumours was significantly higher than that in IDH1-mutant tumours (P<0.001; Fig. 2A). The median FLT T/N ratio in IDH1-mutant and IDH1-wildtype gliomas were 4.18 (IQR 2.28-6.39) and 14.7 (8.98-20.38), respectively. Again, the median FLT T/N ratio in IDH1-wildtype tumours was significantly higher than that in IDH1-mutant tumours (P<0.001; Fig. 2B). There was a significant overlap of MET T/N ratios between IDH1-mutant and IDH1-wildtype tumours. On the other hand, the overlap of FLT T/N ratios between IDH1-mutant and IDH1-wildtype tumours was small. ROC analysis for differentiating IDH1-mutant tumours from IDH1-wildtype tumours showed that the area under the curve (AUC) of the FLT T/N ratio (AUC 0.911, 95% CI 0.847-0.975; Fig. 3B) was significantly larger than that of the MET T/N ratio (AUC 0.727, 95% CI 0.607-0.847; Fig. 3A) (P<0.01). When the cutoff value of the FLT T/N ratio in the ROC curve was set at 6.74, the sensitivity for the differential diagnosis was 92.3%, and the specificity was 75.9% (Fig. 3B). On the other hand, the sensitivity for the differential diagnosis was 88.5%, and the specificity was 51.7% when the cutoff value of the MET T/N ratio was set at 3.72 (Fig. 3A).
When analyzing the uptake values in 36 grade Ⅱ and Ⅲ gliomas separately, the median MET T/N ratios in IDH1-mutant and IDH1-wildtype tumours were 3.59 (IQR 2.87-5.28) and 5.14 (3.97-6.06), respectively. There was no significant difference in the median MET T/N ratios between IDH1-mutant and IDH1-wildtype tumours (Fig. 4A). In the same population, the median FLT T/N ratios in IDH1-mutant and IDH1-wildtype tumours were 3.43 (IQR 2.20-5.82) and 7.56 (6.00-12.94), respectively. The median FLT T/N ratio in IDH1-wildtype tumours was significantly higher than that in IDH1-mutant tumours (P=0.005; Fig. 4B).
When analyzing the uptake values in 13 non-enhancing tumours (9 grade Ⅱ and 4 grade Ⅲ) separately, the median MET T/N ratios in IDH1-mutant and IDH1-wildtype tumours were 2.09 (IQR 1.87-3.44) and 3.21 (2.52-3.59), respectively. There was no significant difference in the median MET T/N ratios between IDH1-mutant and IDH1-wildtype tumours (Fig. 5A). In the same population, the median FLT T/N ratios in IDH1-mutant and IDH1-wildtype tumours were 1.77 (IQR 1.68-2.33) and 6.00 (4.23-6.38), respectively. The median FLT T/N ratio in IDH1-wildtype tumours was significantly higher than that in IDH1-mutant tumours (P=0.028; Fig. 5B).
Tracer uptake in relation to histological grade in IDH1-mutant and wildtype tumours
In the present study, there were only 2 IDH1-mutant GBMs among the 29 IDH1-mutant tumours. We therefore excluded the 2 IDH1-mutant GBMs from the analysis. In the remaining 27 IHD1-mutant tumours, the median MET T/N ratios were 2.86 (IQR 1.95-3.28) and 4.83 (3.48-6.39) for the grade Ⅱ (n=11) and Ⅲ (n=16) gliomas, respectively. The median MET T/N ratio in the grade Ⅲ gliomas was significantly higher than that in the grade Ⅱ gliomas (P=0.002; Fig. 6A). In the same population, the median FLT T/N ratios in the grade Ⅱ and Ⅲ gliomas were 2.14 (IQR 1.69-2.64) and 5.36 (3.90-7.09), respectively. Again, the median FLT T/N ratio in the grade Ⅲ gliomas was significantly higher than that in the grade Ⅱ gliomas (P<0.001; Fig. 6B).
In the 52 IDH1-wildtype tumours, there was only 1 IDH1-wildtype DA. We therefore excluded the DA from the analysis. In the remaining 51 IHD1-wildtype tumours, the median MET T/N ratios were 5.41 (IQR 4.11-6.20) and 6.19 (4.63-7.41) for the grade Ⅲ (n=8) and Ⅳ (n=43) gliomas, respectively. There was no significant difference of the median MET T/N ratios between grade Ⅲ and Ⅳ gliomas (Fig. 6C). In the same population, the median FLT T/N ratios in the grade Ⅲ and Ⅳ gliomas were 8.10 (IQR 6.58-14.53) and 15.65 (10.35-20.77), respectively. The median FLT T/N ratio in the grade Ⅳ gliomas was significantly higher than that in the grade Ⅲ gliomas (P=0.029; Fig. 6D).