New paper published in 'Cell Reports Medicine' (17/04/2026)

New multi-national collaborative study analysing small extracellular vesicles (sEVs) from 206 plasma samples across three independent cohorts (UK, Denmark, NZ), and identifying 45 consistently altered proteins and 20 microRNAs in glioma patients. Our machine learning models achieved AUCs of 0.931-0.971, with 100% accuracy in an independent longitudinal cohort for our proteomic signature.

Critically, biomarker levels normalised with treatment and rose at recurrence—suggesting real potential for monitoring disease progression. Early evidence also shows promise in distinguishing IDH-wildtype glioblastoma from IDH-mutant gliomas (prognoses: 9 months vs up to 207 months).

This minimally invasive blood test could ultimately replace risky brain surgery, delivering answers in days rather than weeks. Next step: standardising procedures for routine clinical use.