- Design
- Systematic review and meta-analysis, mostly retrospective cohorts
- Population
- Patients with metastatic NSCLC and oligoprogression on checkpoint inhibitors
- Primary outcome
- Progression-free and overall survival
- Effect
- PFS HR 0.45 (95% CI 0.33–0.62); OS HR 0.41 (0.27–0.63); low certainty
A meta-analysis examined local ablative treatment, mostly stereotactic body radiotherapy, for metastatic NSCLC progressing at a few sites during checkpoint inhibitor therapy, with the same immunotherapy continued afterwards. Of 22 studies, four (one randomised subgroup and three retrospective cohorts) contributed to each pooled analysis.
Local ablation was associated with better progression-free survival (HR 0.45, 95% CI 0.33–0.62) and overall survival (HR 0.41, 0.27–0.63), with no heterogeneity.
The rationale is that focal progression may reflect a resistant clone rather than failure of the whole treatment. The evidence is low certainty and largely observational; patients selected for ablation may differ in ways that favour survival. Randomised trials are ongoing.
- Consider stereotactic ablation for one to a few progressing sites when other disease remains controlled on immunotherapy
- Continue the same immunotherapy after ablation in such patients
- Discuss in a multidisciplinary meeting — selection matters
- Await randomised data before treating this as standard
Why it matters
It may keep a working immunotherapy line in use rather than abandoning it for a few progressing sites.
Don't overread it
The evidence is mostly observational and low certainty; selection bias could explain part of the benefit.
The statistics, in plain English
A hazard ratio of 0.41 for overall survival suggests about 60% lower risk of death at any time, but with only four contributing studies, mostly retrospective, the estimate is fragile.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for oncology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free