BackgroundEstablished biomarkers of immune checkpoint inhibitor (ICI) response in metastatic non-small cell lung cancer (mNSCLC), such as PD-L1 and tumor mutational burden (TMB), do not identify all patients with durable response. While biomarkers spanning multiple data modalities have been proposed to address this unmet need, additional evidence is required for use in the clinic. Here, we performed a comparative study assessing the association between previously described biomarkers of ICI response and outcomes in a real-world mNSCLC cohort.MethodsUsing the Tempus database, we analyzed de-identified records of non-squamous EGFR-negative and ALK fusion-negative mNSCLC patients treated with first-line ICI regimens and profiled with targeted-panel DNA-seq and whole-exome-capture RNA-seq. ICI-related biomarkers were calculated following published methods using the Tempus IO platform (table 1). Immune biomarkers included features describing tumor biology (including PD-L1 IHC, STK11 and KEAP1 mutations, APOBEC signature1), HLA (HLA-LOH, HLA A*03 genotype), neoantigens (TMB, neoantigen burden), and RNA-based signatures of immune activity such as IFNγ tumor inflammation signature,2 IMPRES,3 Cytotoxic score,4 NRS,5tertiary lymphoid structure (TLS),6 7 and others. Real-world time to progression (rwTTP) was defined as the interval from ICI start to the first progression event, censored on last known physician encounter. Cox proportional hazards (Cox PH) models were fitted to evaluate the relationship between each biomarker and rwTTP.ResultsIn this mNSCLC cohort (N=199), multiple RNA-based signatures related to immune infiltration, TLS, cytotoxic and cytolytic activity, and IFNγ tumor inflammation signature (TIS) were significantly associated with longer rwTTP (P < 0.001). HLA- and tumor-related biomarkers had no significant associations with longer rwTTP. As expected, TMB and PD-L1 IHC were significantly associated with longer rwTTP (P < 0.05). In multivariable Cox PH analyses controlling for PD-L1 IHC and TMB, several immune signatures remained significantly associated with rwTTP (table 1). Notably, many RNA-based signatures were highly correlated despite non-overlapping genes and functions, underscoring a broad profile of immune infiltration and activation is likely associated with ICI response.ConclusionsHere we demonstrate that RNA-based immune signatures are significantly associated with clinical benefit and may supplement well-established ICI biomarkers in therapy selection. These immune RNA-based signatures need to be prospectively validated in future studies.ReferencesWang S, Jia M, He Z, Liu XS. APOBEC3B and APOBEC mutational signature as potential predictive markers for immunotherapy response in non-small cell lung cancer. Oncogene. 2018;37(29):3924–3936.Ayers M, Lunceford J, Nebozhyn M, et al. IFN-γ-related mRNA profile predicts clinical response to PD-1 blockade. J Clin Invest. 2017;127(8):2930–2940.Auslander N, Zhang G, Lee JS, et al. Robust prediction of response to immune checkpoint blockade therapy in metastatic melanoma published correction appears in Nat Med. 2018 Dec;24(12):1942. Nat Med. 2018;24(10):1545–1549.Lau D, Khare S, Stein MM, et al. Integration of tumor extrinsic and intrinsic features associates with immunotherapy response in non-small cell lung cancer. Nat Commun. 2022;13(1):4053.Huang AC, Orlowski RJ, Xu X, et al. A single dose of neoadjuvant PD-1 blockade predicts clinical outcomes in resectable melanoma. Nat Med. 2019;25(3):454–461.Cabrita R, Lauss M, Sanna A, et al. Tertiary lymphoid structures improve immunotherapy and survival in melanoma published correction appears in Nature. 2020 Apr;580(7801):E1. Nature. 2020;577(7791):561–565.Andersson A, Larsson L, Stenbeck L, et al. Spatial deconvolution of HER2-positive breast cancer delineates tumor-associated cell type interactions. Nat Commun. 2021;12(1):6012.Abstract 181 Table 1ICI biomarker associations with rwTTP in ICI-treated mNSCLC cohort (n=199). Cox PH regression results are shown for each biomarker as a single predictor (Univariate Cox PH HR, P-value) and as a predictor with TMB and PD-Ll IHC included in a multivariable Cox PH model (Multivariable Cox PH HR, P-value).