Diagnostic performance of spectroscopic and perfusion MRI for distinction of brain tumors

OBJECTIVE To assess the value of spectroscopic and perfusion MRI for glioma grading and for distinguishing glioblastomas from metastases and from CNS lymphomas. METHODS The authors examined 79 consecutive patients with first detection of a brain neoplasm on nonenhanced CT scans and no therapy prior to evaluation. Spectroscopic MRI; arterial spin-labeling MRI for measuring cerebral blood flow (CBF); first-pass dynamic, susceptibility-weighted, contrast-enhanced MRI for measuring cerebral blood volume; and T1-weighted dynamic contrast-enhanced MRI were performed. Receiver operating characteristic analysis was performed, and optimum thresholds for tumor classification and glioma grading were determined. RESULTS Perfusion MRI had a higher diagnostic performance than spectroscopic MRI. Because of a significantly higher tumor blood flow in glioblastomas compared with CNS lymphomas, a threshold value of 1.2 for CBF provided sensitivity of 97%, specificity of 80%, positive predictive value (PPV) of 94%, and negative predictive value (NPV) of 89%. Because CBF was significantly higher in peritumoral nonenhancing T2-hyperintense regions of glioblastomas compared with metastases, a threshold value of 0.5 for CBF provided sensitivity, specificity, PPV, and NPV of 100%, 71%, 94%, and 100%. Glioblastomas had the highest tumor blood flow values among all other glioma grades. For discrimination of glioblastomas from grade 3 gliomas, sensitivity was 97%, specificity was 50%, PPV was 84%, and NPV was 86% (CBF threshold value of 1.4), and for discrimination of glioblastomas from grade 2 gliomas, sensitivity was 94%, specificity was 78%, PPV was 94%, and NPV was 78% (CBF threshold value of 1.6). CONCLUSION Perfusion MRI is predictive in distinguishing glioblastomas from metastases, CNS lymphomas and other gliomas vs MRI and magnetic resonance spectroscopy.