Brain Tumor Reporting and Data System (BT-RADS) Score: A Novel Predictor of the Karnofsky Performance Scale and Progression-free Survival in Glioblastoma Patients
Michele Persico1, Tsion Assaye2, Michael Essien2, Brent D. Weinberg2
1Neurology, 2Radiology and Imaging Services, Emory University
Objective:
We aim to evaluate the Brain Tumor Reporting and Data System Score (BT-RADS) predictive value on functional outcomes assessed through the Karnofsky Performance Scale (KPS) and progression-free survival (PFS) in glioblastoma (GB) patients. 
Background:
The BT-RADS is a novel eight-point classification system that standardizes magnetic resonance imaging (MRI) reporting in GB patients. While previous studies have demonstrated BT-RADS predictive value in the prognosis and overall survival of high-grade glioma patients, its association with functional outcomes and progression-free survival (PFS) in GB patients is unclear. 
Design/Methods:
BT-RADS from MRI ~3 months post-diagnosis after chemo-radiation therapy (BT-RADS_A), and ~6 months post-diagnosis after the 6-month maintenance temozolomide (TMZ) cycle (BT-RADS_B) were analyzed. BT-RADS was classified as low (BT-RADS score 0-3a) and high (BT-RADS score 3b-4). 
Results:
In a univariate logistic regression analysis, BT-RADS_A was a significant predictor (p=0.04) in discriminating an unfavorable KPS from a favorable one. In a multivariate logistic regression, in which the best model fit was the model containing BT-RADS_A and MGMT tumor status (model p value=0.03), the ROC analyses indicate that BT-RADS_A is not better than chance in differentiating an unfavorable KPS from a favorable one (model AUC=0.65; 95% CI=0.55, 0.76). The discriminatory ability of BT-RADS scores alone in predicting KPS outcomes was modest, as indicated by the AUC. This underscores the need for larger sample sizes.  In a final model of BTRADS_A and BT-RADS_B, accounting for MGMT tumor status, the risk of tumor progression was 50% lower (p value=0.003) with a low BT-RADS (0-3a) in the BT-RADS _A group and 38% lower (p value=0.03) with a low BT-RADS (0-3a) in the BT-RADS_B group. 
Conclusions:
Considering the above-discussed findings, BT-RADS may represent a new tool in the clinical management of glioblastoma patients by informing on disease progression, prognosis, and quality of life.
10.1212/WNL.0000000000208837
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.