KEY POINTS
- The systematic review and meta-analysis included 14 studies, 687 patients and 896 locally recurrent brain metastases treated with repeat stereotactic radiotherapy after previous stereotactic treatment. All included studies were retrospective observational series or case series.
- Repeat treatment doses ranged from 15.5 to 26.5 Gy in 1-5 fractions, with single-fraction treatment accounting for 64.6% of repeat procedures. The most common regimen was approximately 18 Gy in one fraction, and the median BED10 across studies was 50.4 Gy.
- Pooled one-year local control was 76% (95% CI 73-79%), with low statistical heterogeneity (I²=23.9%). A competing-risk sensitivity analysis produced a similar estimate of 75%.
- Higher repeat-treatment BED10 was significantly associated with better one-year local control (p=0.007), while larger gross tumor volume was associated with worse control (p=0.027). The resulting tumor-control probability models showed a significant dose-response relationship.
- Pooled one-year overall survival was 62% (95% CI 52-70%), although heterogeneity was substantial (I²=78.7%). Higher BED10 and smaller tumor volume were also associated with better one-year survival in exploratory meta-regression.
- The major trade-off was radiation necrosis. The pooled incidence of any radiation necrosis was 21% (95% CI 18-25%), while symptomatic radiation necrosis occurred in 10% (95% CI 7-13%). Median time to necrosis across reporting studies ranged from 2.8 to 12 months.
- The dose-response analysis used study-level aggregate data rather than individual patient data, and definitions of recurrence and radiation necrosis varied substantially between institutions. The authors therefore describe the TCP modeling as hypothesis-generating rather than a basis for a definitive repeat-SRS dose prescription.
CLINICAL TAKEAWAY
Repeat stereotactic radiotherapy is a reasonable salvage option for selected locally recurrent brain metastases, with approximately three-quarters controlled at one year. The approximately 10% symptomatic necrosis risk is clinically important, and the apparent benefit from higher BED10 must be balanced against normal-brain dose and lesion volume rather than interpreted as a simple mandate for dose escalation.