Diffusion and perfusion-weighted imaging in the structural and functional characterization of diffuse gliomas for differential diagnosis and treatment planning

Abstract

Diffuse infiltrative gliomas are the most common primary brain tumors in adult population and represent a range from slow-growing lesions to highly invasive tumors with poor prognosis. The growth potential and aggressiveness of gliomas depend on World Health Organization (WHO) grade and molecular features. The standard radiographic characterization of glioma is based on magnetic resonance imaging (MRI); a widely utilized examination for both the initial diagnosis and for ongoing post-treatment management of these patients. Despite the growing success and number of applications in neuro-oncologic imaging there are still many challenges that need to be solved with regard to diagnosis, preoperative delineation and posttreatment monitoring of gliomas. The objectives of this thesis were to evaluate the value of the advanced MR techniques, as perfusion and diffusion imaging, to characterize diffuse gliomas with respect to WHO grading systems, morphologic and genetic features and to find new imaging prognostic biomarkers in order to individualize medical management. In study I we retrospectively evaluated the ability of Apparent Diffusion Coefficient (ADC) and relative Cerebral Blood Flow (rCBV) parameters derived from whole-tumor normalized histograms to stratify progression-free survival (PFS) and overall survival (OS) in patients with diffuse gliomas grade II and III. In patients with oligodendrogliomas we found that tumors with heterogeneous perfusion signatures and high average perfusion values were associated with longer PFS, while in patients with astrocytomas, heterogeneous perfusion distribution was associated with poorer outcomes. We did not find a significant association between ADC from diffusion MRI and patient survival. Our results indicate that perfusion MRI might serve as an independent factor to predict prognosis in patients with diffuse gliomas. Study II was based on the same retrospective material as for study I. The aim was to determine whether the rCBV and ADC parameters could help differentiate genetically defined oligodendrogliomas from astrocytomas and to distinguish between WHO grade II and grade III diffuse gliomas. Several large clinical trials have demonstrated that patients with oligodendrogliomas (Isocitrate dehydrogenase (IDH)-mutant and 1p19q codeleted gliomas) derive more benefit from chemotherapy when compared to their genetic astrocytoma counterparts. Differentiation between these two entities affects treatment strategy and prognosis. We found that patients with oligodendrogliomas showed significantly higher microvascularity and higher vascular heterogenity than patients with astrocytomas. Among diffuse gliomas, oligodendrogliomas revealed a higher cellular density. Combined use of ADC and rCBV histogram parameters had superior diagnostic performance to identify oligodendroglial tumors. Thus, imaging-based biomarkers of vascularity and cellularity may constitute a non-invasive supplement to histopathologic and molecular genetic markers and provide important information to guide future treatment. In study III we prospectively included forty-two patients with untreated IDH wild-type glioblastoma to evaluate the ability of MRI-based Restriction Spectrum Imaging (RSI) to estimate the level of cellularity (cell density in tissue) in the different tumoral zones. We also investigated the prognostic value of RSI to assess the potentially aggressive behavior of the tumor and compared it to established diffusion metrics such as mean diffusivity (MD) and fractional anisotropy (FA). The highest RSI-cellularity index was measured in the contrastenhanced zone with a negative gradient from the tumor core to the periphery of the peritumoral zone. Shorter survival outcomes were significantly associated with higher RSIcellularity index in the contrast-enhanced zone, but also in the peri-enhancing zone and near peri-tumoral zone. In contrast, MD and FA in the near peri-tumoral zone did not show any predictive value to survival outcome, which may indicate a more severe affection of MD and FA values by extracellular edema than for RSI-cellularity index. RSI provided a promising prognostic biomarker to depict tumor infiltration in the peritumoral brain zone, which can be helpful to optimize surgical procedures and radiation field mapping with significant benefits for treatment. In conclusion, our results suggest that both perfusion and diffusion MRI provide reliable non-invasive biomarkers of glioma status and the information from the two imaging techniques appear to be complementary. Additionally, to further implementation of these modalities in a diagnostic work-flow, multicenter studies are warranted that would assist in standardizing imaging protocols as well as postprocessing procedures.