NIMG-23. BRAIN TUMOR REPORTING AND DATA SYSTEM (BT-RADS) AND QUANTITATIVE TOOLS TO GUIDE ITS IMPLEMENTATION. (11th November 2019)

• Record Type: Journal Article
• Authors: Weinberg, Brent; Ramesh, Karthik; Gurbani, Saumya; Schreibmann, Eduard; Kleinberg, Lawrence; Shu, Hui-Kuo; Shim, Hyunsuk
• Is Part Of: Neuro-oncology. Volume 21(2019)Supplement 6
• Page Start: vi166
• Page End: vi166
• Publication Date: 2019-11-11
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noz175.695
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Glioblastoma is the most aggressive primary adult brain tumor, with median survival of 15 months despite surgery and chemoradiation. MRI is used to guide treatment decisions, but imaging interpretation is challenging because of subtle findings with overlap between treatment effect and disease progression. The most frequently used quantitative brain tumor assessment metric is the Response Assessment in Neuro-Oncology (RANO), but this scoring system requires manual delineation 2-D metrics which can be subjective and is not directly tied to patient management decisions. The Brain Tumor Reporting and Data System (BT-RADS) is a novel, management-based reporting system developed by the Emory brain tumor team to improve quantitation and reduce bias in imaging assessment. BT-RADS scoring incorporates quantitative volumetric changes in contrast-enhanced T1-weighted (CE-T1w) and fluid attenuation inversion recovery (FLAIR) MRI, patient medications, clinical outcome, and treatment dates, and has been aligned closely with management decisions. To improve repeatability and reliability of scoring, we have begun to develop a cloud platform to help physicians utilizing BT-RADS. Longitudinal MRI data from each patient are automatically co-registered using rigid registration and aligned using trilinear interpolation, enabling voxel-to-voxel comparisons across multiple scans. The platform implements a semi-automated algorithm to segment tumor volumes using curvature flow, thresholding, and morphological filtering which expedites clinical review, as physicians simply edit and confirm segmentation accuracy rather than manual segmenting and measuring 2D images. Additionally, radiation dose maps can be overlaid on clinical images to determine what may be treatment effect and in-field vs. out-of-field recurrence. A secure web interface allows easy use by the entire treatment team, e.g. in a "tumor board" setting. Future plans include incorporating clinical and genomic data and fully automating tumor segmentation. The goal of this work is to provide more quantitative and objective follow-up metrics that can guide clinical decision making in glioblastoma patients.