Perivascular spaces as MRI markers associated with radiation necrosis in brain metastases patients: an exploratory study
In brief
Higher brain MRI perivascular-space grades tracked radiation necrosis in 71 patients
Radiation necrosis developed in 33 of 71 patients after brain radiotherapy. Higher perivascular-space grades in the brain's centrum semiovale were associated with about sixfold and 26-fold higher odds of necrosis, but estimates were imprecise; links in other brain regions also rested on small numbers. This exploratory study needs validation with larger cohorts and detailed dose analysis.
- Journal
- Neuroradiology (Q1)
- Published
- 8 October 2026
- Study design
- Prospective / inception cohort
- Evidence level
- Level 2, Moderate (CEBM 2b)
- Authors
- Carlo Augusto Mallio, Gianfranco Di Gennaro, Ugo Ferrari, Caterina Bernetti, Gloria Perillo, Paolo Matteucci, et al.
- PMID
- 42848065
- DOI
- 10.1007/s00234-026-04216-0
Why clinicians should know about it
- Picked for Radiation Oncology (paper of the day, 11 October 2026): Radiation necrosis risk linked to perivascular spaces
Abstract
BACKGROUND AND PURPOSE: Radiation necrosis (RN) is a significant complication of stereotactic radiotherapy (SRT) for brain metastases. We investigated whether pre-treatment cerebral small vessel disease (cSVD) burden, particularly perivascular spaces (PVSs), is associated with RN risk. MATERIALS AND METHODS: This single-center retrospective study included 71 patients treated with fractionated SRT for brain metastases. RN was diagnosed by consensus using multiparametric MRI including DSC perfusion. Pre-treatment cSVD markers were evaluated according to STRIVE criteria, with PVSs graded using Potter's criteria in the basal ganglia (BG), centrum semiovale (CS), and midbrain (MB). Associations with RN were assessed using Firth's penalized logistic regression adjusted for age and sex. RESULTS: RN developed in 33 patients (46.5%). Higher PVS grades in the CS were significantly associated with increased RN risk (grade 1: OR 6.35, 95% CI 1.86-21.72, p = 0.003; grade 2: OR 26.26, 95% CI 3.64-189.61, p = 0.001). Associations were also observed for higher PVS burden in the BG and MB, although these rested on small subsets with wide confidence intervals. Other cSVD markers showed no significant associations. Confidence intervals were wide due to sparse data in higher PVS grades. CONCLUSIONS: This exploratory study suggests a potential association between pre-treatment PVS burden and RN risk. However, sparse data, wide confidence intervals, and the absence of lesion-level dosimetric adjustment limit conclusions. These preliminary findings warrant validation in larger, prospective cohorts with comprehensive dosimetric analysis.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.