Background and Purpose
Standard assessment criteria for brain tumors that only includes anatomic imaging, continues to be insufficient. While numerous studies have demonstrated the value of DSC-MRI perfusion metrics for this purpose, they have not been incorporated due to a lack of confidence in the consistency of DSC-MRI metrics across sites and platforms. This study addresses this limitation with a comparison of multi-site/multi-platform analyses of a shared brain tumor patient DSC-MRI dataset.
Methods
DSC-MRI data was collected, after a preload and during a bolus injection of gadolinium contrast agent using a GRE-EPI sequence (TE/TR=30/1200ms; flip angle=72°). Forty-nine low-grade (LG:n=13) and high-grade (HG:n=36) glioma datasets were uploaded to the cancer imaging archive (TCIA). Datasets included a predetermined AIF (arterial input function), enhancing tumor ROIs and ROIs necessary to create normalized relative cerebral blood volume (nRCBV) and cerebral blood flow (nCBF) maps.
Seven sites computed 20 different perfusion metrics. Pair-wise agreement between sites was assessed with Linn’s-Concordance-Correlation-Coefficient (LCCC). Distinction of LG from HG tumor was evaluated with the Wilcoxon-rank-sum test followed by ROC analysis to identify the optimal threshold(s) based on sensitivity(SN) and specificity(SP).
Results
For nRCBV and nCBF 93% and 94% of entries showed good or excellent cross-site agreement (0.8≤LCCC≤1.0). All metrics could distinguish LG from HG tumor. Optimum thresholds were determined for pooled data (nRCBV=1.4, SN:SP=90%:77%; nCBF=1.58, SN:SP=86%:77%).
Conclusion
Using DSC-MRI data, obtained after a preload of contrast agent, substantial consistency resulted across sites for brain tumor perfusion metrics with a common threshold discoverable for distinguishing LG from HG tumor.