see Recurrent Glioblastoma treatment.
Recurrent high-grade glioma, especially glioblastoma, remains one of the greatest challenges in neuro-oncology due to inevitable recurrence and limited therapeutic options.
| Treatment | Median OS (recurrence) | Comments |
|---|---|---|
| Surgery | 9–12 months | Only for selected, high-functioning patients |
| Re-irradiation | 6–10 months | Best if >6 months since initial RT |
| Bevacizumab alone | 4–6 months | Improves symptoms, not OS |
| Lomustine | 6–8 months | Modest benefit |
| BEV + lomustine | ~9 months | ↑PFS, unclear OS benefit |
| Clinical trial | Variable | Best option in qualified centers |
In a systematic review and meta-analysis published in Therapeutic advances in neurological disorders of non-randomized, two-arm clinical trials comparing re-irradiation + bevacizumab versus bevacizumab alone in patients with recurrent high-grade gliomas (rHGG), aiming to demonstrate superiority in survival outcomes. Hammed et al. claim that the combination therapy improves overall survival (OS) and progression-free survival (PFS) without increasing grade ≥3 toxicity, and propose it as a potentially standard salvage approach 1).
💣 Critical Appraisal
1. Low-Level Evidence in Disguise
This is a meta-analysis based entirely on retrospective, non-randomized studies, with all the biases and heterogeneity that come with them. The use of ROBINS-I and II to “evaluate” these studies does not mitigate the inherent confounding by indication, selection bias, and lack of treatment standardization. The authors repeatedly conflate association with causation, ignoring the fact that hazard ratios in observational cohorts are not treatment effects.
2. Pooling Apples and Oranges
The methodology pools data across studies with wildly heterogeneous re-irradiation protocols (doses, volumes, pulsed vs. non-pulsed), variable use of BEV (dose, schedule), and differing baseline characteristics. Combining such studies and then reporting pooled HRs is statistical theater with no clinical applicability. Worse yet, subgroup analysis by age, gender, and performance status in this context becomes data dredging, not hypothesis testing.
3. Toxicity Reporting = Wishful Thinking
The authors claim no increase in grade 3 toxicity — without acknowledging that toxicity data are the weakest part of retrospective trials, often underreported and inconsistently graded. The meta-analysis fails to stratify by radiation volume or cumulative dose, which are critical in assessing toxicity in re-irradiation contexts.
4. Clinical Relevance? Minimal at Best
While the paper reports a statistically significant HR for OS (0.69), the absolute median survival gains are not reported. No mention is made of quality of life, neurocognitive outcomes, or steroid dependence, which are central in rHGG salvage treatment. Improving OS by a few weeks while ignoring the trade-offs is misleading at best and harmful at worst.
5. Prospective Trials? Good Luck
The authors end by calling for prospective trials — a noble gesture, but an empty one given that they have already drawn definitive therapeutic conclusions. This dual messaging — “Here is the answer, but someone else should prove it” — is the hallmark of pseudo-precision medicine built on methodological sand.
🧯 Final Verdict
This study offers an illusion of progress in recurrent high-grade glioma treatment. It recycles weak retrospective data through the statistical blender of meta-analysis and pours out hazard ratios that look impressive but mean very little. It is a classic case of overconfident conclusions from underpowered, heterogenous data.
Until proper randomized evidence emerges, clinicians should resist the temptation to treat these results as practice-changing. The paper is more editorial decoration than solid evidence.