Brain metastases (BMs) occur in 40% of patients with lung cancer. The activity of immunotherapy in these patients, however, remains controversial, as the cornerstone treatment is radiotherapy (RT). Because RT is associated with adverse events that may impair the quality of life, the possibility of substituting it with a single systemic approach is attractive. Therefore, we performed a systematic review and meta-analysis to evaluate the potential benefit of immune checkpoint inhibitors (ICIs) in patients with NSCLC with untreated BM (unBM).

Studies that enrolled patients with NSCLC treated with ICIs and specifically allowed for unBM were identified by searching the EMBASE, PubMed, Cochrane, and other databases. The outcomes evaluated were intracerebral overall response rate (icORR) and intracerebral disease control rate (icDCR) for unBM, and grades 3 and 4 toxicity rate.

We included 12 studies with a total of 566 individuals in the final analysis. Anti-programmed cell death protein-1 therapy seems to be active in the central nervous system, with an icORR of 16.4% (95% confidence interval [CI]: 9.8%-24%; I = 33.17%) and an icDCR of 45% (95% CI: 33.4%-56.9%; I = 46.91%). In the meta-analysis for icORR (risk ratio = 1.26, 95% CI: 0.57-2.79) and icDCR (risk ratio = 0.88, 95% CI: 0.55-1.43) we did not observe any difference among patients with BM who were treated with RT before ICI start and those who were treated with ICI only.

ICI seems to be effective as a single treatment for active BM in selected patients with advanced NSCLC.