Spectroscopic MRI-Based Biomarkers Predict Survival for Newly Diagnosed Glioblastoma in a Clinical Trial

SIMPLE SUMMARY: Due to the infiltrative nature of glioblastoma, standard MRI techniques, such as T1-weighted contrast-enhanced (T1w-CE) and T2-weighted fluid-attenuated inversion recovery (FLAIR), imperfectly delineate radiation-targeted tumor volume. With spectroscopic MRI, the ratio of choline elevation, a tumor biomarker and N-acetylaspartate reduction, a healthy neuronal biomarker, can better determine the extent of the tumor. The aim of our secondary analysis was to determine if there was a relationship between survival outcomes and biomarkers identified by spectroscopic MRI for a cohort of 28 glioblastoma patients who received high-dose radiation guided by spectroscopic MRI. We determined that the volume of post-surgical spectroscopically abnormal tissue was a biomarker of overall and progression-free survival, whereas the volume of residual contrast enhancement, determined by T1w-CE MRI, was not. Our results suggest that accurate delineation and treatment of an infiltrative tumor not identified by contrast is a critical component of glioblastoma management and patient survival. ABSTRACT: Despite aggressive treatment, glioblastoma has a poor prognosis due to its infiltrative nature. Spectroscopic MRI-measured brain metabolites, particularly the choline to N-acetylaspartate ratio (Cho/NAA), better characterizes the extent of tumor infiltration. In a previous pilot trial (NCT03137888), brain regions with Cho/NAA ≥ 2x normal were treated with high-dose radiation for newly diagnosed glioblastoma patients. This report is a secondary analysis of that trial where spectroscopic MRI-based biomarkers are evaluated for how they correlate with progression-free and overall survival (PFS/OS). Subgroups were created within the cohort based on pre-radiation treatment (pre-RT) median cutoff volumes of residual enhancement (2.1 cc) and metabolically abnormal volumes used for treatment (19.2 cc). We generated Kaplan–Meier PFS/OS curves and compared these curves via the log-rank test between subgroups. For the subgroups stratified by metabolic abnormality, statistically significant differences were observed for PFS (p = 0.019) and OS (p = 0.020). Stratification by residual enhancement did not lead to observable differences in the OS (p = 0.373) or PFS (p = 0.286) curves. This retrospective analysis shows that patients with lower post-surgical Cho/NAA volumes had significantly superior survival outcomes, while residual enhancement, which guides high-dose radiation in standard treatment, had little significance in PFS/OS. This suggests that the infiltrating, non-enhancing component of glioblastoma is an important factor in patient outcomes and should be treated accordingly.