Background
Glioblastoma and primary CNS lymphoma (PCNSL) dictate different neurosurgical strategies; it is critical to distinguish between them preoperatively. However, current imaging modalities do not effectively differentiate them. We aim to examine the use of DWI and T1-weighted dynamic contrast-enhanced (DCE)-MRI as potential discriminative tools.
Methods
We retrospectively reviewed 18 PCNSL and 36 matched glioblastoma patients with pretreatment DWI- and DCE-MRI. VOIs were drawn around the tumor on contrast-enhanced T1WI and FLAIR images; these images were transferred onto coregistered ADC maps to obtain ADC, and onto DCE perfusion maps to obtain plasma volume (Vp) and permeability transfer constant (Ktrans). Histogram analysis was performed to determine mean (ADCmean) and relative ADCmean (rADCmean), and relative 90th percentile values for plasma volume (rVp90%tile) and permeability transfer constant (rKtrans90%tile). Non-parametric tests were used to assess differences and ROC analysis was performed for optimal threshold calculations.
Results
The enhancing component of PCNSL was found to have significantly lower ADCmean (1.1 × 103 vs 1.4 × 103; p<0.001) and rADCmean (1.5 vs 1.9; p<0.001) and rVp90%tile (3.7 vs 5.0; p<0.05) than the enhancing component of glioblastoma, but not significantly different rKtrans90%tile (5.4 vs 4.4; p=0.83). The non-enhancing portions of GBM and PCNSL did not differ in these parameters. Based on ROC analysis, mean ADC provided the best threshold (AUC 0.83) to distinguish primary CNS lymphoma from glioblastoma, which was not improved with normalized ADC or addition of perfusion parameters.
Conclusion
ADC was superior to DCE-MRI perfusion alone or in combination in differentiating PCNSL from glioblastoma.