Document resource
Background Recent expansions in immunotherapy technologies and applications have led to significant improvements in patient outcomes across many tumor types. However, these gains have been limited by tumor immune escape and suppression, which render certain patient populations resistant to these therapies. Here, we present an improved method of integrating immune and tumor biomarkers to classify the interplay between tumor and immune activity in the context of immunotherapy response.Methods A validated clinical assay for comprehensive genomic and immune profiling (CGIP) was performed on 31,100 tumor samples, including the expression analysis of 395 immune genes and variant prevalence for 523 genes. Three previously published gene expression signatures were calculated from this data: cell proliferation (CP), tumor immunogenic signature (TIGS), and cancer testis antigen burden (CTAB). In addition, immunohistochemistry-derived PD-L1 status and tumor mutational burden (TMB) were calculated for each sample. Principle component analysis (PCA) and unsupervised clustering analysis were used to detect emergent biological groups based on these biomarkers. A nearest centroid classification method was used to classify an immunotherapy-treated cohort of 208 patients with non-small cell lung cancer (NSCLC) into one of these groups, and the relationship between these groups and survival was assessed using Kaplan-Meier analysis.Results Applied to the 31,100-patient study cohort, the biomarker integration workflow suggested the existence of four distinct tumor populations, or immune response phenotypes (IRPs), each characterized by the state of the balance between tumor expansion and immune activity: 1) Tumor Dominant, characterized by elevated CTAB, CP, and TIGS; 2) Proliferative, characterized by elevated CP; 3) Inflamed, characterized by elevated TIGS; and 4) Checkpoint, characterized by elevated TMB, PD-L1, TIGS, and CP. These results were consistent with those previously obtained from a smaller, 5450-patient study cohort. Additionally, gene variant analysis revealed distinct patterns of mutation prevalence in each IRP. In the immunotherapy-treated NSCLC cohort, we found significant association between IRP and overall survival [p=0.0092], with the checkpoint IRP demonstrating improved survival over the other IRPs.Conclusions Our study demonstrated the scientific utility of a biomarker integration approach combining PCA with unsupervised clustering to reveal emergent biological groups called IRPs within real-world CGIP data. We then showed how new CGIP samples can be classified into one of these IRPs and demonstrated an association between IRP and survival in immunotherapy-treated NSCLC. We aim to expand and refine these methods to develop treatment decision and clinical trial selection tools that integrate multiple biomarkers to characterize tumor-immune interactions in solid tumors.Ethics Approval Approval for this study was obtained from an independent institutional review board (IRB), the Western Copernicus Group (WCG) (www.wcgclinical.com), under protocol # 1340120. The WCG IRB determined this study met requirements for a waiver of informed consent under 45 CFR 164.512.