This meta-analysis concluded that neutron beam therapy did not improve survival of high-grade glioma patients; there was no definitive conclusion regarding carbon therapy. Unclear study quality assessment and analysis of a small number of studies and small sample sizes made the reliability of these conclusions is unclear.

To investigate carbon ion therapy for the treatment of high-grade glioma.

MEDLINE and CancerLit (1960 to 2007) and The Cochrane Library (issue 1, 2006) were searched. Search terms were reported. Reports by governmental organisations (National Cancer Institute, International Agency for Research on Cancer), scientific societies (American Society of Clinical Oncology, European Society for Therapeutic Radiology and Oncology) co-operative groups for clinical research (European Organisation for Research and Treatment of Cancer) or agencies for health technology assessment (Agence Nationale d'Accreditation et d'Evaluation en Sante, Swedish Council on Technology Assessment in Health Care, National Institute for Clinical Excellence) were searched. No language restrictions were applied.

Randomised controlled trials (RCTs) or acceptable equivalent study designs of any duration with less than 10% loss to follow-up that compared neutron beam therapy with photon therapy (reference treatment) in adult patients with histologically confirmed high-grade glioma were eligible for inclusion. Neutron or light ion beam therapy could be delivered either with or without photon therapy and neutron beam therapy as a boost or as a large field irradiation; associated surgery or chemotherapy was also acceptable. For inclusion, published data had to allow intention-to-treat analysis. All of the included patients underwent some surgical procedure. The primary endpoint was overall survival. Secondary endpoints were: local control for treatment efficiency assessed through autopsy results (percentage of remaining tumours) for neutrons and imaging for carbon ions; median overall survival; median progression-free survival; and treatment-related toxicity.

Prospective and retrospective non-RCTs were included for light ion therapy (carbon and neon) and assessed the effects of carbon or neon therapy alone or mixed with photons. Mean or median age was 53 years. Most studies were single centre. The proportion of males ranged from 38% to 54%. Patient age ranged from less than 40 to more than 50, where reported. Most patients had Grade III or Grade IV glioma. Doses ranged from 13Gy to 50Gy.

Two reviewers independently assessed studies for inclusion.

The authors stated that methodological quality and absence of bias was assessed (number of reviewers not reported). There was no description of the validity assessment, but treatment allocation, description of patient characteristics and availability of intention-to-treat analysis were described for each study within the text.

Overall mortality or survival rates were extracted to calculate relative risks (RR) and 95% confidence intervals (CIs). For carbon or neon ion therapy, overall median survival at two years was read from the printed survival curves.

Data were extracted by one reviewer and checked by another; discrepancies were resolved by a third reviewer.

Relative risks and associated 95% CIs (for neutron beam therapy studies) were pooled using a fixed-effect model. Statistical heterogeneity was assessed using the Cochran Q test and factors that may have explained heterogeneity were investigated using post-hoc analyses, otherwise a random-effects model was used. Sensitivity analyses were planned with and without quasi-randomised trials. Publication bias was assessed using funnel plots. The results for overall median survival for carbon or neon ion therapy studies were reported narratively.

Four RCTs that assessed neon beam therapy were included in the review (n=324, range 34 to 166). The process of treatment allocation was described in three studies. Patient characteristics were poorly described in three studies. Intention-to-treat analyses were performed in three studies.

There was a 7% increase in mortality at 12 months with neutron beam therapy compared with photon therapy (RR 1.07, 95% CI 0.95 to 1.20; four studies). This was not statistically significant. Similar results were seen at 24 months. There was no evidence of heterogeneity or publication bias. Two phase I/II studies for carbon and neon ion therapy for glioblastoma reported 10% and 31% two-year overall survival and 13.9 and 19 months median survival.

Three studies reported on toxicity, with complete and permanent epilation in the neutron beam arm compared with scanty or patchy regrowth of finer hair in the photon therapy arm.

Further results were reported in the review.

This meta-analysis suggested that neutron beam therapy did not improve survival of high-grade glioma patients; there was no definitive conclusion regarding carbon therapy.

The review question was defined in terms of participants, intervention, study design and outcomes. All languages were included, which reduced the possibility of language bias. Some sources of unpublished studies were searched and there was no evidence of publication bias (the authors acknowledged that the assessment may have been underpowered). Study selection and data extraction were performed in duplicate, which reduced risks of error and bias; this was not reported for validity assessment. Some aspects of study validity were assessed, but did not appear to have been performed systematically. The number of studies included was small and sample sizes were small, so the reliability of these studies was unclear. Heterogeneity was assessed and some possible sources were investigated.

Unclear assessment of study quality and the small number of studies and small sample sizes in the analysis made the reliability of the authors' conclusions unclear.

Practice: The authors stated that carbon therapy may be administered in first intention to patients diagnosed with relapsing glioblastoma and in second intention as light-ion complementary dose in initial treatment.

Research: The authors stated that well-designed prospective controlled trials on carbon therapy were needed to confirm the concept and early benefits of dose escalation with high linear energy transfer particles (further details reported).

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.