Background: Bone metastasis (BM) occurs in about 40% of patients with metastatic lung cancer. Recently, BM was associated with decreased OS to nivolumab in previously-treated NSCLC (Landi L et al, P1.01.53, 19^th WCLC, 2018). CCTG BR.34 (NCT03057106) was an open-label, randomized phase II clinical trial that randomized 301 patients with treatment-naïve, high-risk, stage IVA/B NSCLC without sensitizing EGFR or ALK alterations (1:1) to durvalumab plus tremelimumab with or without platinum doublet chemotherapy. First, 109 patients accrued with stage IVB, or selected IVA disease. Then 192 patients accrued with any stage IVA/B disease. In CCTG BR.34, median OS was not significantly different: 16.6 mo in the chemotherapy plus immunotherapy (C+IO) arm, vs 14.1 mo in the IO alone arm (HR 0.88, p = 0.46) (Leighl NB et al, J Thor Oncol, 2021). However, in BR.34 PFS was significantly longer in the C+IO arm (7.7 mo) compared to the IO alone arm (3.2 mo) (HR 0.67, 95% CI, 0.52 ? 0.88). Here we analyzed the effect of BM on outcomes in BR.34. Methods: The 301 patients in the trial were characterized by the presence of BM at study entry (129-yes, 172-no). BM effect was evaluated on trial outcomes (OS, PFS, and ORR) using Cox/logistic regression analysis. Multivariable analysis was performed adjusting for the clinical and molecular covariates available. Results: In univariate analysis of the entire study population, median OS was significantly shorter for patients with BM vs those without BM (10. 9 vs 18.7 mos, HR 1.68, p = 0.001), as was median PFS (3.4 vs 7.2 mos, HR 1.82, p < 0.0001), and lower ORR (29.5% vs 45.9%, OR 0.52, p = 0.003), respectively. There was no evidence of differential association of BM with treatment arms for OS (p = 0.23), PFS (p = 0.84), and ORR (p = 0.25, Breslow-Day test). In multivariate analysis (MVA), BM remained significantly associated with worse OS (HR 1.44, p = 0.026), PFS (HR 1.69, p < 0.0001), and ORR (OR 0.52, p = 0.01). In MVA for OS: TMB, histology type, race, and ECOG were also significant; but age, smoking history, and PD-L1 IHC status were not significant. Conclusions: In CCTG BR.34 the presence of BM at trial entry was associated with significantly shorter OS, PFS, and lower ORR. BM is therefore a significant adverse prognostic factor in high-risk, stage IVA/B NSCLC treated with durvalumab and tremilimumab (with or without platinum doublet chemotherapy). If confirmed in a larger phase III trial, BM should be considered as an important new stratification factor in all clinical trials of immune checkpoint inhibitor (ICI) therapy. We and others have reported that molecules arising in the bone microenvironment (e.g: IL-8, PTHrP, TGF-b, sclerostin, and activin A) cause immunosuppression in cancer, and future trials should evaluate the addition of targeted therapies against these factors in combination with the ICIs in patients with BM.