Multivariate ROC analysis showing AUC±SE values for various combinations of FDG-PET and MRI parameters in the detection of glioma recurrencea
| rCBVmean (0.680 ± 0.011) | ADCmean (0.752 ± 0.085) | Cho/Cr (0.861 ± 0.080) | ADCmean + rCBVmean (0.781 ± 0.079) | ADCmean + Cho/Cr (0.912 ± 0.051) | rCBVmean + Cho/Cr (0.860 ± 0.083) | ADCmean + rCBVmean + Cho/Cr (0.913 ± 0.053) | |
|---|---|---|---|---|---|---|---|
| ADCmean (0.752 ± 0.085) | 0.781 ± 0.079 | ||||||
| Cho/Cr (0.861 ± 0.080) | 0.860 ± 0.083 | 0.912 ± 0.051 | 0.913 ± 0.053b | ||||
| TBRmax (0.827 ± 0.078) | 0.831 ± 0.081 | 0.848 ± 0.072 | 0.894 ± 0.059 | 0.850 ± 0.073 | 0.935 ± 0.044 | 0.889 ± 0.061 | 0.932 ± 0.046 |
| TBRmean (0.888 ± 0.059) | 0.884 ± 0.063 | 0.888 ± 0.058 | 0.935 ± 0.044 | 0.897 ± 0.057 | 0.935 ± 0.046b.c | 0.928 ± 0.047 | 0.932 ± 0.048b |
↵a Each entry represents the AUC±SE of a combination of parameters mentioned in the respective rows and columns. Values in parentheses represent AUC±SE of that particular parameter or combination. Blank cells are left to avoid repetition of values.
↵b A combination of FDG-PET and MR imaging parameters has a better AUC±SE than a combination of MR imaging parameters.
↵c Note that the maximum AUC was achieved with a combination of ADCmean, Cho/Cr, and TBRmean.