Gliomas exhibit a complex tumour microenvironment (TME) that contributes to heterogeneity and treatment resistance. Noninvasive methods that probe both intracellular and extracellular compartments are therefore important. This study tests the feasibility and complementarity of combining multi b‑value diffusion‑weighted MRI (DW‑MRI) with the Vascular, Extracellular, and Restricted Diffusion for Cytometry in Tumours (VERDICT) biophysical model and diffusion‑weighted MR spectroscopy (DW‑MRS) to characterise tumour cells and the surrounding TME in the same clinical session.
Fourteen patients with newly diagnosed glioma (WHO grades 2 to 4) underwent scanning at 3 T. Among these, 10 patients had single‑voxel DW‑MRS acquired in addition to multi‑shell DW‑MRI. The cohort included 4 IDH wildtype and 10 IDH mutant tumours. Tumours were automatically segmented into enhancing, non‑enhancing, and oedema regions using a previously validated pipeline. The protocol was designed to be clinically feasible for routine or research brain MRI sessions.
Multi‑shell DW‑MRI data were fitted using the VERDICT framework to estimate biologically relevant microstructural parameters. Reported VERDICT outputs included intracellular volume fraction (fIC), cell radius, extracellular diffusivities, and free‑water fraction (fFW). These parameters aim to separate contributions from intracellular restricted diffusion, extracellular hindered diffusion, and free water, providing metrics that relate to cell density and extracellular matrix properties.
Single‑voxel DW‑MRS was used to derive metabolite‑specific apparent diffusion coefficients (ADCs) for total N‑acetylaspartate (tNAA), total creatine (tCr), and tCho (choline). These ADCs provide diffusion‑weighted measures specific to neuronal and glial metabolites rather than to bulk tissue water. DW‑MRS results are reported as descriptive statistics and trends in the manuscript.
The study compared DW‑MRI VERDICT parameters and DW‑MRS metabolite ADCs between tumour regions (enhancing, non‑enhancing, oedema) and contralateral normal‑appearing white matter (NAWM), and according to IDH mutation status. Key findings from these comparisons were:
VERDICT metrics differentiated IDH‑wildtype from IDH‑mutant gliomas. IDH‑wildtype lesions exhibited significantly higher intracellular volume fraction (fIC) and lower extracellular diffusivities in both enhancing and non‑enhancing tumour regions compared with IDH‑mutant tumours.
DW‑MRS showed a descriptive trend toward reduced tNAA ADC in tumour tissue versus contralateral NAWM, which the authors interpret as consistent with neuronal loss in tumour regions.
A descriptive trend toward increased tCho ADC in tumours was reported, suggested to reflect glial activation or altered choline compartment properties in the TME.
There was a descriptive trend for decreased tNAA ADC in IDH‑wildtype tumours relative to IDH‑mutant tumours; DW‑MRS results are presented as trends rather than as definitive group differences.
The investigators assessed associations between metabolite ADCs from DW‑MRS and VERDICT‑derived DW‑MRI parameters using Pearson correlations. In non‑enhancing tumour regions they identified significant relationships between tumoral tCho ADC and several VERDICT metrics:
Positive correlations between tCho ADC and fEES (the extracellular‑extravascular space fraction) and extracellular diffusivity, suggesting that higher choline diffusion aligns with a larger or more permissive extracellular compartment.
Negative correlations between tCho ADC and both ADC (bulk water diffusion metric) and fFW (free‑water fraction), indicating complex relationships between metabolite diffusion and water‑based diffusion indices in infiltrative regions.
These correlations are presented as evidence of complementarity between metabolite‑specific diffusion measures and modelled water diffusion microstructure parameters.
The combined protocol permits simultaneous interrogation of intracellular and extracellular compartments: VERDICT provides cell‑level and extracellular matrix characterisation from multi‑shell DW‑MRI, while DW‑MRS delivers metabolite‑specific diffusion indices that relate to neuronal (tNAA) and glial (tCho) compartments. The observed higher fIC and lower extracellular diffusivities in IDH‑wildtype tumours point to increased cellularity and altered extracellular properties in that subgroup. Trends in tNAA and tCho ADCs support the interpretation that DW‑MRS can detect neuronal loss and glial compartment changes in tumour tissue. Correlations between tCho ADC and VERDICT metrics in non‑enhancing regions further support the modalities’ complementarity for probing the infiltrative TME.
This work is presented as a proof‑of‑concept preprint and has not been peer reviewed. DW‑MRS findings are reported as descriptive trends in a modest sample (10 with DW‑MRS), and the manuscript provides statistical comparisons for VERDICT measures. The authors state that all data produced are available upon reasonable request. Local ethical approval was obtained from the University College London ethics committee (05/Q0502/101 (34772‑PHYSICS01)), and written consent procedures and reporting guidelines were followed. The authors declared no competing interests.
Combining multi b‑value DW‑MRI with VERDICT modelling and DW‑MRS is feasible within a clinically realistic protocol and yields complementary measures of glioma cellularity, extracellular diffusion properties, and metabolite‑specific diffusion. VERDICT differentiated IDH status via higher fIC and lower extracellular diffusivity in IDH‑wildtype tumours, while DW‑MRS provided metabolite‑specific trends consistent with neuronal loss (tNAA) and glial changes (tCho). Correlations between modalities in infiltrative regions suggest added value from concurrent acquisition; further validation in larger, peer‑reviewed cohorts is required.