Abstract 452: Novel MRI-based biomarker for early assessment of glioma recurrence

Abstract

Abstract Gliomas continue to be the most common form of brain malignancy in adult patients. Even with advancements in the clinical management of these patients, assessment of tumor recurrence continues to be based on late or serial changes in tumor volume as measured by CT or MRI [1; 2]. The purpose of this study was to develop and evaluate a new voxel-based method, referred as parametric response mapping (PRM), for early detection of brain tumor progression using standard MR images (FLAIR and contrast-enhanced T1-weighted). PRM results were evaluated relative to standard MRI-based criteria of clinical progression assessment in patients with Grade III/IV gliomas. Fourteen patients with grades III/IV glioma underwent MRI before and during treatment (every 2 months). MRI scans, acquired on a 1.5T or a 3T clinical scanner, consisted of fluid-attenuated inversion recovery imaging (FLAIR) and contrast-enhanced T1-weighted (CE-T1w) images. All images were spatially registered to CE-T1w pre-treatment scans and tumors were manually contoured on the T1w-Gd images. Subsequent to PRM analysis, FLAIR images were normalized to the signal using the white matter (rFLAIR). PRM, applied to the whole brain volume on rFLAIR, was performed by first generating voxel-based difference maps at each longitudinal follow-up scan using the baseline as the subtrahend. Baseline was defined as either the pre-treatment or subsequent rFLAIR image. Individual voxels were classified based on the extent of change observed in the rFLAIR difference maps. Three classifications were used: voxels with a significant increase in rFLAIR values (PRMrFLAIR+), significant decrease (PRMrFLAIR-) or statistically unchanged (PRMrFLAIR0). Disease recurrence was defined by PRM as the sum of red voxels that exceed 20% of the tumor volume as delineated on T1w-Gd. PRMrFLAIR detected tumor recurrence on average 18±3 weeks earlier than the MacDonald criteria. It was also observed that PRMrFLAIR+ found outside the tumor CE-T1w regions revealed the spatial location at which the tumor would progress on subsequent scans based on current standard progression criteria. These results show that the PRM approach provides for the early detection and spatial depiction of brain tumor progression prior to detection by currently available conventional MRI-based criteria. While further validation of PRM for detection of tumor recurrence is currently underway, the ability to detect tumor progression significantly earlier than currently possible is anticipated to have a major impact on patient care. [1] Macdonald, D. R., et al. (1990). J Clin Oncol 8(7): 1277-1280. [2] Wen, P. Y., et al. (2010). J Clin Oncol 28(11): 1963-1972. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 452. doi:1538-7445.AM2012-452