Detecting glioblastoma infiltration beyond conventional imaging tumour margins using MTE-NODDI

Abstract Glioblastoma (GBM) is the most common and aggressive brain tumour with stark resistance to available therapies, leading to relapse and a median survival of <15 months. A key cause of therapy resistance is diffuse infiltration of tumour cells into brain regions surrounding the tumour, which presents a major clinical challenge as existing imaging techniques offer limited detection of the resectable margin. Here, we use diffusion weighted imaging (DWI) and apply the multiple echo time neurite orientation dispersion and density imaging (MTE-NODDI) model as a tool to detect tumour cells in the hard-to-distinguish margin. We used the G144 patient-derived xenograft model, with characteristic invasion along white matter tracts, in combination with MTE-NODDI. Tumour development was monitored, and magnetic resonance imaging (MRI) data were acquired over a 4-week period, starting at 4 weeks after stereotactic injection of tumour cells. MTE-NODDI demonstrated sensitivity to the developing tumour in the invading margin, and changes in measured parameters were apparent from 6 weeks after injection. In comparison to standard DWI, MTE-NODDI showed increased sensitivity to the tumour-associated changes in the margin. Furthermore, extraneurite volume fraction (fen) and neurite density index (NDI) measured from MTE-NODDI correlated with immunohistological measurement of tumour cells. These findings suggest that MTE-NODDI may non-invasively detect infiltrating cells and tumour-induced pathology in margin regions without T2 or DWI changes in a patient-derived mouse model of GBM. MTE-NODDI is clinically translatable and could be a powerful tool for neurosurgeons to maximise surgical resection, resulting in better survival outcomes for patients with GBM.