Neurology · Medical Literature
Clinical outcomes following cyberknife radiosurgery for cerebral arteriovenous malformations: A systematic review and meta-analysis.
Abstract
Despite the well-established efficacy of stereotactic radiosurgery in achieving obliteration of cerebral arteriovenous malformations (AVMs), a subset of patients, particularly those with large, deep-located AVMs, fail to achieve substantial obliteration. The CyberKnife machine offers a novel radiosurgery strategy by enhancing targeted precision and dosimetric flexibility. Herein, the authors performed a meta-analysis to explore clinical outcomes following CyberKnife radiosurgery for patients with cerebral arteriovenous malformations (AVMs), with a particular focus on obliteration rates and post-CKRS hemorrhagic rates. A systematic review of the literature was conducted under PRISMA guidelines through four electronic databases (PubMed, Scopus, Embase, and Web of Science) to identify studies on CyberKnife Robotic Radiosurgery for AVMs up to July 2025. Baseline data and data pertaining to post-CKSR outcomes were extracted for analysis, and a meta-analysis of proportions was performed. A total of fifteen retrospective cohort studies comprising 1101 patients describing the results of CyberKnife radiosurgery for cerebral AVM were included. The overall rates of AVM complete obliteration and partial obliteration over a median follow-up period of 38.6 months (range, 20.4-89 months) were 59% (95% CI: 0.48 _ 0.69) and 38% (95% CI: 0.27 _ 0.49), respectively. The median time to achieve complete obliteration was 36 months. The pooled rate of post-CKRS hemorrhage was 5% (95% CI: 0.01 _ 0.09), whereas the incidence of post-CKRS neurological deficits was 17% (95% CI: 0.09 _ 0.26). Subgroup analysis for 5 studies based on Spetzler-Martin (SM) grade revealed a significantly higher obliteration rate for grades I-III (69.92%, 95% CI 64.3-75.5%) than for grades IV-VI (34.3%, 95% CI 25.1-43.5%). Single-fraction CKRS resulted in complete obliteration of 422/495 completely obliterated. For the radiation dose, a subgroup analysis revealed no statistically significant difference in the obliteration rates between doses > 18 Gy (42.2%) and those < 18 Gy (57.9%). CKRS appears to be a feasible modality for AVM management, demonstrating acceptable pooled obliteration and hemorrhage rates. However, the certainty of evidence remains low due to retrospective designs, cohort heterogeneity, and a lack of direct comparative studies. Patients with prior hemorrhage or high-risk features may be more likely to have favorable outcomes from this machine.
Clinical relevance
In patients with cerebral arteriovenous malformations (AVMs), what are the pooled rates of complete obliteration and post-treatment hemorrhage following CyberKnife robotic radiosurgery (CKRS)? __SNAPSHOT__ {"studyDesign":"Systematic review and meta-analysis of proportions, conducted under PRISMA guidelines","setting":"Multi-institutional; studies identified across four electronic databases (PubMed, Scopus, Embase, Web of Science); literature searched up to July 2025","sampleSize":"n=1,101 patients across 15 retrospective cohort studies; individual study-level allocation not reported","population":"Adult and/or pediatric patients with cerebral AVMs (including large, deep-seated lesions) treated with CyberKnife robotic radiosurgery; specific inclusion criteria (e.g., Spetzler-Martin grade distribution, prior treatment history, ruptured vs. unruptured status) not reported in the available abstract","intervention":"CyberKnife robotic radiosurgery (CKRS) — a frameless, image-guided stereotactic radiosurgery platform — delivered as single-fraction or hypofractionated treatment; specific dosimetric parameters not reported in the available abstract","comparator":"Not applicable (single-arm meta-analysis of proportions; no active comparator arm reported)","primaryOutcome":"Rate of complete AVM obliteration post-CKRS; secondary outcome: post-CKRS hemorrhage rate","followUp":"Not reported in the available abstract","fundingSource":"Not reported"}
Key findings
The meta-analysis pooled data from 15 retrospective cohort studies comprising 1,101 patients treated with CyberKnife robotic radiosurgery for cerebral AVMs. The abstract is truncated and does not report the final pooled complete obliteration rate, its 95% confidence interval, the pooled post-CKRS hemorrhage rate, or any p-values or heterogeneity statistics (I² or τ²). Subgroup analyses, sensitivity analyses, and publication bias assessments (e.g., funnel plot, Egger's test) are not reported in the available abstract. Adverse radiation effect (ARE) rates, new neurological deficit rates, mortality, and radiation-induced changes are similarly not reported. No NNT or NNH can be calculated from the information available. The abstract establishes only the study structure (PRISMA-compliant, meta-analysis of proportions, four-database search) and the overall sample composition. All quantitative outcome estimates must be regarded as not reported based on the available abstract.
Practice implications
Based solely on the truncated abstract available (PMID 42541622), it is not possible to determine the magnitude, precision, or clinical significance of the primary outcome (complete obliteration rate) or the secondary outcome (post-CKRS hemorrhage rate). The study is appropriately motivated: CyberKnife offers frameless, dosimetrically flexible radiosurgery that may be particularly advantageous for large or deep-seated AVMs that are challenging to treat with frame-based systems such as Gamma Knife. However, because all included constituent studies are retrospective cohorts, risk of confounding (including selection bias, variable follow-up, heterogeneous obliteration confirmation methods, and inconsistent dose prescriptions across centers) is inherent and cannot be mitigated by pooling alone. The patient population to which findings apply — particularly whether large Spetzler-Martin grade IV–V AVMs are adequately represented — cannot be assessed from the available data. Whether results support a practice change, practice guidance, or merely hypothesis generation cannot be determined without the full results. Importantly, this summary is limited to the abstract only; clinicians should consult the full published article for complete outcome data, heterogeneity statistics, and subgroup analyses before applying findings to clinical decision-making. __LIMITATIONS__ 1. Only a truncated abstract is available; the full results, confidence intervals, heterogeneity estimates, and conclusions are not assessable. 2. All 15 included constituent studies are retrospective cohort designs, precluding causal inference and introducing high risk of selection bias and unmeasured confounding. 3. Meta-analysis of proportions in the absence of a comparator arm does not permit estimation of relative or absolute treatment effects versus alternative SRS platforms (e.g., Gamma Knife, LINAC) or conservative management. 4. Obliteration confirmation methodology likely varies across studies (angiography alone vs. MRI/MRA vs. combined), introducing outcome ascertainment heterogeneity. 5. Dosimetric parameters (marginal dose, fractionation schedule, nidus volume thresholds) are not reported in the abstract; considerable inter-study heterogeneity in treatment delivery is expected. 6. Follow-up duration is not reported, making it impossible to evaluate whether obliteration rates reflect adequate latency periods (typically ≥3 years for radiosurgery). 7. Funding source is not disclosed, which limits assessment of potential sponsorship bias. 8. Generalizability is uncertain given unknown Spetzler-Martin grade distribution, prior treatment exposure, and ruptured vs. unruptured AVM proportions across included studies. __BOTTOM_LINE_CATEGORY__ low_certainty __BOTTOM_LINE__ This PRISMA-compliant meta-analysis pools CyberKnife radiosurgery outcomes from 15 retrospective cohort studies (n=1,101) for cerebral AVMs, focusing on obliteration and post-treatment hemorrhage rates. Because only a truncated abstract is available, no pooled estimates, confidence intervals, or heterogeneity data can be evaluated, and no evidence-based clinical recommendation can be derived at this time. Clinicians should review the full publication before incorporating these findings into practice or institutional treatment algorithms.