Magnetic Resonance Fingerprinting Guided Extended Resection in Glioblastomas

Glioblastomas (GBs) are aggressive malignant brain tumors with a median survival of less than 15 months . Infiltration of cancer beyond the tumor margins causes recurrence in nearly 100% of GBs; however, this cannot be measured by current imaging techniques . Availability of reliable and reproducible infiltration prediction maps at initial diagnosis will open new treatment opportunities such as targeted surgery or escalated radiation therapy (RT). On clinical contrast enhanced (CE) magnetic resonance imaging (MRI) scans, a typical GB demonstrates an enhancing mass with central necrosis and an extensive surrounding, peritumoral region with bright signal on T2-weighted(w) and FLAIR (Fluid attenuation inversion recovery) images. This bright, peritumoral T2/FLAIR region is known to contain vasogenic edema and tumor infiltration, as it is well known that GBs infiltrate beyond the enhancing tumor margins. Since there is a clear link between extent of tumor resection and survival the challenge for neurosurgeons is maximizing resection of tumor, while avoiding neurological injury. Typically, the central region of the tumor can be safely resected with minimal risk. The challenge lies in maximal safe resection along the tumor margins as it infiltrates normal brain. MR Fingerprinting is a quantitative imaging (QI) scan developed at CWRU that provides rapid quantification of multiple tissue properties, such as T1 and T2 relaxation maps, with high reproducibility and excellent tissue characterization. Our preliminary analysis of retrospective data of 60 GB participants with MRF+MRI scans with targeted 5-aminolevulenic acid (5-ALA) tissue sampling demonstrates an AUC of 0.8 for MRF/MRI model for GBM infiltration prediction in peritumoral region .