Distinction between glioma progression and post-radiation change by combined physiologic MR imaging

Abstract

Magnetic resonance (MR) diffusion-weighted imaging (DWI), dynamic susceptibility contrast-enhanced perfusion imaging (DSC), and MR spectroscopy (MRS) techniques provide specific physiologic information that may distinguish malignant glioma progression from post-radiation change, yet no single technique is completely reliable. We propose a simple, multiparametric scoring system to improve diagnostic accuracy beyond that of each technique alone. Fifteen subjects with lesions suspicious for glioma progression following radiation therapy who had also undergone 3-tesla DWI, DSC, and MRS studies of the lesion were retrospectively reviewed. Minimum apparent diffusion coefficient (ADC) ratio, maximum regional cerebral blood volume (rCBV) ratio, and maximum MRS choline/creatine (Cho/Cr) and choline/N-acetyl-aspartate (Cho/NAA) metabolic peak-height ratios were quantified within each lesion. Each parameter (ADC ratio, rCBV ratio, and combined Cho/Cr and Cho/NAA ratios) was scored as either glioma progression (one point) or radiation change (zero point) based upon thresholds derived from our own data. For each lesion, the combined parameters yielded a multiparametric score (0 to 3) for prediction of tumor progression or post-radiation change. Optimum thresholds for ADC ratio (1.30), rCBV ratio (2.10), and either combined Cho/Cr (1.29) and Cho/NAA (1.06) yielded diagnostic accuracies of 86.7%, 86.7%, and 84.6%, respectively (p < 0.05). A combined multiparametric score threshold of 2 improved diagnostic accuracy to 93.3% (p < 0.05). In this small series combining 3-T DWI, DSC, and MRS diagnostic results using a simple, multiparametric scoring system has potential to improve overall diagnostic accuracy in distinguishing glioma progression from post-radiation change beyond that of each technique alone.