Reported brain SRS injury patterns vary widely between metastasis and AVM cohorts

Across 8,385 patients, reported radiation-related injury ranged from 6.8–30.4% in metastasis cohorts and 3.4–38.4% in AVMs, using noncomparable endpoints.

KEY POINTS

  • This PRISMA-based systematic review evaluated radiation-induced brain injury after stereotactic radiosurgery for brain metastases (BM) and arteriovenous malformations (AVMs). The search identified 30 eligible studies, of which 16 independent primary clinical series comprising 8,385 patients contributed to the quantitative descriptive synthesis: 3,567 patients with brain metastases and 4,818 patients with AVMs. Fourteen additional investigations informed the qualitative assessment of imaging, management, and potential biological mechanisms.
  • Reported radiation-related injury rates varied considerably across clinical settings. Individual study event frequencies ranged from 6.8% to 30.4% in brain metastasis cohorts and 3.4% to 38.4% in AVM cohorts. Importantly, these were not pooled estimates of radiation necrosis incidence: studies combined different definitions of radiation necrosis, radiation-induced imaging changes, adverse radiation effects, and delayed complications. Differences in follow-up, patient- versus lesion-level analysis, and competing mortality prevented valid direct comparison.
  • Brain metastasis cohorts generally reported radiation-related injury within the first one to two years after SRS or fractionated stereotactic RT. Symptomatic injury occurred in approximately 8.5–10.0% of patients in selected contemporary series. Larger irradiated volumes, melanoma or renal cell carcinoma histology, and concomitant immune checkpoint inhibitors were identified as possible risk modifiers. Normal-brain V12Gy values around 7.9–8.5 cm³ appeared in selected dose-response evidence, but the authors explicitly caution against interpreting these as universally validated safety thresholds.
  • AVM cohorts demonstrated a much longer spectrum of delayed complications, including cyst formation and chronic encapsulated hematoma occurring up to 24 years after radiosurgery. In one long-term series, symptomatic cysts or chronic encapsulated hematomas occurred in approximately 3% of patients, with reported cumulative rates of 3.7% at ten years and 9.4% at 15 years. These complications are not biologically identical to conventional radiation necrosis and cannot be directly compared with earlier post-SRS injury in brain metastasis populations.
  • Dose-volume exposure and lesion anatomy emerged as recurring risk considerations, although no single predictive model was validated. In AVM cohorts, larger nidus volume, particularly >10 cm³, higher marginal dose, and previous embolization were associated with delayed complications, with prior embolization showing an approximate OR of 2.1 in selected evidence. In brain metastasis cohorts, irradiated volume and tumor characteristics were more frequently emphasized, reflecting differences in treatment geometry and clinical context.
  • Deep periventricular white matter may represent a particularly vulnerable anatomical region. Selected brain metastasis series reported radiation necrosis rates as high as 41% in periventricular deep white matter, compared with approximately 6.1% in cortical locations. However, these observations came from small and heterogeneous datasets, including a subgroup of only 22 periventricular lesions, and do not establish an independent anatomical dose constraint or demonstrate intrinsically different tissue radiosensitivity.
  • The review supports individualized assessment rather than universal surveillance or treatment rules. Perfusion MRI and amino-acid PET may help distinguish radiation injury from tumor progression when conventional imaging is inconclusive. Bevacizumab remains a treatment option for selected patients with symptomatic radiation necrosis, particularly when corticosteroids are ineffective or poorly tolerated, but routine prophylactic anti-VEGF treatment is unsupported. No universally validated pathology-specific surveillance schedule or dose constraint could be recommended from the available evidence.

CLINICAL TAKEAWAY

Radiation-related brain injury after SRS is heterogeneous, with reported event frequency and timing strongly influenced by treatment geometry, anatomical location, underlying disease, and follow-up methods. The evidence supports careful dose-volume assessment and awareness of very late complications after AVM radiosurgery, but does not justify direct comparison of necrosis rates between brain metastases and AVMs or adoption of universal surveillance schedules.

SOURCE

Clinical Oncology

Browse more research Suggest a correction