NIMG-44. ROLE OF PRE-TREATMENT TUMOR DYNAMICS AND IMAGING RESPONSE IN DISCRIMINATING GLIOBLASTOMA SURVIVAL FOLLOWING GAMMA KNIFE. (6th November 2017)

• Record Type: Journal Article
• Authors: Singleton, Kyle W; Johnston, Sandra K; Rickertsen, Cassandra R; De Leon, Gustavo; Kunkel, Lauren R; Rockhill, Jason; Mrugala, Maciej; Bendok, Bernard; Patel, Naresh; Porter, Alyx; Swanson, Kristin R
• Is Part Of: Neuro-oncology. Volume 19(2017)Supplement 6
• Page Start: vi151
• Page End: vi152
• Publication Date: 2017-11-06
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/nox168.619
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: INTRODUCTION: Glioblastoma (GBM) is a notoriously deadly cancer with poor prognosis following tumor progression. Stereotactic radiosurgery (SRS) is advantageous for localized treatment of new tumor burden thus controlling cumulative radiation dose to normal tissue. However, post-treatment effect and tumor progression are difficult to decipher due to transient effects seen on subsequent contrast-enhanced MRI's. In this work, we investigated Days Gained (DG), our previously reported personalized response metric, for discriminating overall survival following gamma knife SRS. METHODS: A retrospective analysis was performed on 17 GBM patients who received gamma knife SRS at first tumor recurrence. Tumor volumes were segmented on T1GD and FLAIR MRI at three imaging timepoints (2 prior to therapy, 1 at approximately 1-month post-SRS). Pre-treatment tumor growth velocity (Vpre) was computed and used to estimate DG. Kaplan-Meier and Cox proportional hazard analysis compared DG and Vpre for discriminating survival since the start of treatment. RESULTS: Kaplan-Meier and Cox analysis identified thresholds of DG that discriminate survival for T1Gd (DG>34.5, p=0.008) and FLAIR (DG>11.6, p=0.026). This significance was retained when treating DG as a continuous variable for FLAIR (HR=0.996, p=0.030) but not for T1Gd (p=0.32). An Vpre threshold of 14.6 mm/year (p=0.021) was found for T1GD which was significant in univariate Cox analysis (HR=1.03, p=0.023) and after controlling for age and sex (HR=1.034, p=0.009). No significant thresholds were found for FLAIR velocity. DISCUSSION: Identified DG thresholds were lower than previous analyses for other treatments. This suggests the typical 1-month follow-up interval after SRS may be suboptimal in identifying clinically significant response measured using DG. Significant discrimination using T1GD velocity is also suggestive that transient enhancement may confound DG. Further analysis is needed to determine if later follow-up images improve prediction. CONCLUSION: Aggressive tumor growth prior to SRS and low Days Gained are poor prognostic indicators for gamma knife.