Impact on survival of glioblastoma patients in relation to the imaging of the peri-surgical area: a multi-parametric diffusion MRI, perfusion MRI, and [11C]MET PET study.

OBJECTIVES

Glioblastoma (GBM) is an aggressive brain tumor with poor prognosis; recurrence near the surgical cavity is common despite multimodal-treatment. Conventional MRI lacks accuracy for early detection of recurrence and cannot reliably differentiate progression from pseudoprogression. This prospective-study evaluated whether <72 h post-surgery imaging biomarkers from diffusion-weighted imaging (DWI), intravoxel incoherent motion (IVIM), dynamic-susceptibility contrast (DSC) MRI, and [11C]Methionine PET (MET PET), could predict progression-free survival (PFS) and overall survival (OS).

METHODS

A prospective cohort of 21 GBM patients underwent multimodal imaging. Metrics from 0 to 5 and 5 to 10 mm peri-surgical rims and contralateral white matter were analyzed with Spearman correlations and Cox regression.

RESULTS

IVIM diffusion coefficient D in the 0-5 mm rim correlated positively with OS (r = 0.821, P = .023), whereas pseudo-diffusion D* in the 5-10 mm rim correlated negatively (r = -0.786, P = .036), consistent with longer survival with lower microvascular perfusion/higher lower tissue density. Cox regression showed a significant association between SUVmax MET PET uptake and OS (HR = 13.022, P = .044), indicating a lower SUVmax MET PET corresponds with a longer OS.

CONCLUSIONS

IVIM-derived metrics and MET PET SUVmax may serve as prognostic biomarkers of OS after GBM surgery. Despite small sample size and heterogeneous acquisition, the associations warrant validation in larger, standardized cohorts and exploration for personalized post-operative risk stratification.

ADVANCES IN KNOWLEDGE: This study shows that IVIM-captured microvascular characteristics and MET PET SUVmax <72 h post-surgery predict OS, outperforming conventional MRI and DSC metrics. It suggests that both vascularity near the cavity and residual-tumor signal may drive recurrence risk more than residual-tumor signal.