Article Text
PDF
Abstract
Background Clinically actionable biomarkers of immune checkpoint inhibitor (ICI) response are currently limited to specific mutation profiling, immunohistochemistry staining for PD-L1, and tumor mutational burden. Use of the latter two are challenging, as they are incompletely predictive and lack accepted standards for measurement and interpretation. Transcriptomic associations with response have been reported and may add critical information to an integrated biomarker strategy. There is a need for better understanding of the performance of potential biomarkers across multiple datasets and tumor tissue types.
Methods RNA sequencing FASTQ data files from 12 ICI trials 1-13 and 29 solid tumors in The Cancer Genome Atlas (TCGA)14 were processed using a standardized bioinformatics workflow for quality control, mapping, generation of gene expression matrices, and extraction of immunogenomics features. We evaluated 18 2, 15-22 immunogenomics features that have been published or proposed to associate with clinical response to ICI therapy for correlation with response and survival across these datasets, estimating predictive information from the ICI trials and prognostic information from TCGA dataset results.
Results The MIRACLE score was associated with response and survival in most ICI studies, both overall and within melanoma trials (figures 1). Other immunogenomics features had both lower effect sizes of outcome associations and fewer cohorts in which their outcome associations were statistically significant. Features that were associated with outcome in the ICI studies were generally associated with survival in TCGA as well, whether evaluating all tumor tissue types (figure 2) or melanoma only (figure 3). In melanoma, the TIDE score was associated with response to ICIs, but not with overall survival in TCGA, though the effect size was small. Gene expression signatures built from responders versus non-responders in each trial did not yield generalizable associations with response across other trials. Harmonized gene expression data and immunogenomics features extracted in this project are available for review and further analysis in the CRI iAtlas platform (https://cri-iatlas.org/).
Conclusions The MIRACLE score performed best in both effect size and frequency of studies where its association with outcome was statistically significant. No features gave substantial predictive rather than prognostic information. We expect that integration of transcriptomic features with clinical features and DNA alterations will be required to provide predictive (rather than just prognostic) information. Methods that train models on prioritization of predictive information and generalizability across studies may be required for optimal biomarker development.
Acknowledgements We would like to thank SITC for funding for this work as part of the Sparkathon TimIOS collaborative project. We would also like to thank the authors of the published datasets we have used in the study.
References
Gide TN, et al. Distinct Immune Cell Populations Define Response to Anti-PD-1 Monotherapy and Anti-PD-1/Anti-CTLA-4 Combined Therapy. Cancer Cell, 2019;35(2):238–255 e6.
Hugo W, et al. Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma. Cell, 2016;165(1): 35–44.
McDermott DF, et al. Clinical activity and molecular correlates of response to atezolizumab alone or in combination with bevacizumab versus sunitinib in renal cell carcinoma. Nat Med, 2018;24(6):749–757.
Balar AV, et al. Atezolizumab as first-line treatment in cisplatin-ineligible patients with locally advanced and metastatic urothelial carcinoma: a single-arm, multicentre, phase 2 trial. Lancet, 2017;389(10064): 67–76.
Rosenberg JE, et al. Atezolizumab in patients with locally advanced and metastatic urothelial carcinoma who have progressed following treatment with platinum-based chemotherapy: a single-arm, multicentre, phase 2 trial. Lancet, 2016;387(10031): 1909–20.
Liu D, et al. Integrative molecular and clinical modeling of clinical outcomes to PD1 blockade in patients with metastatic melanoma. Nat Med, 2019;25(12): 1916–1927.
Cloughesy TF, et al. Neoadjuvant anti-PD-1 immunotherapy promotes a survival benefit with intratumoral and systemic immune responses in recurrent glioblastoma. Nat Med, 2019. 25(3): p. 477-486.
Riaz N, et al. Tumor and Microenvironment Evolution during Immunotherapy with Nivolumab. Cell, 2017;171(4): 934–949 e16.
Van Allen EM, et al. Genomic correlates of response to CTLA-4 blockade in metastatic melanoma. Science, 2015;350(6257): 207–211.
10. Zhao J, et al. Immune and genomic correlates of response to anti-PD-1 immunotherapy in glioblastoma. Nat Med, 2019;25(3): 462–469.
11. Kim ST, et al. Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer. Nat Med, 2018;24(9): 1449–1458.
12. Miao D, et al. Genomic correlates of response to immune checkpoint therapies in clear cell renal cell carcinoma. Science, 2018;359(6377): 801–806.
13. Choueiri TK, et al. Immunomodulatory activity of nivolumab in metastatic renal cell carcinoma. Clin Cancer Res, 2016;22(22): 5461–5471.
14. Cancer Genome Atlas Research N, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet, 2013. 45(10): 1113–20.
15. Bindea G, et al. Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer. Immunity 2013;39(4): 782–95.
16. Thorsson V, et al. The immune landscape of cancer. Immunity 2018;48(4): 812–830. e14.
17. Roufas C, et al. The expression and prognostic impact of immune cytolytic activity-related markers in human malignancies: a comprehensive meta-analysis. Front Oncol 2018;8: 27.
18. Hollern DP, et al. B Cells and T follicular helper cells mediate response to checkpoint inhibitors in high mutation burden mouse models of breast cancer. Cell, 2019;179(5): 1191–1206 e21.
19. Bortone DS, et al. Improved T-cell Receptor Diversity Estimates Associate with Survival and Response to Anti-PD-1 Therapy. Cancer Immunol Res 2021;9(1): 103–112.
20. Auslander N, et al. Robust prediction of response to immune checkpoint blockade therapy in metastatic melanoma. Nat Med, 2018;24(10): 1545–1549.
21. Turan T, et al. A balance score between immune stimulatory and suppressive microenvironments identifies mediators of tumour immunity and predicts pan-cancer survival. Br J Cancer, 2021;124(4): 760–769.
22. Jiang P, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med, 2018;24(10): 1550–1558.
Overall survival associations of selected immunogenomics features. Rows represent selected immunogenomics features and columns represent individual datasets. Results from ICI trials are shown in the left panel, and results from TCGA datasets are shown in the right panel. Row/column intersections represent effect size (triangle direction and color) and statistical significance (triangle size) of associations with overall survival. Column-side colorbars show various dataset features for comparison
Predictive versus prognostic information content of selected immunogenomics features. X-axis represents the log10 hazard ratio with 95% confidence interval derived from TCGA data, and Y-axis represents log10 hazard ratio with 95% confidence interval derived from ICI data.
Predictive versus prognostic information content of selected immunogenomics features, melanoma trials only. X-axis represents the log10 hazard ratio with 95% confidence interval derived from TCGA data, and Y-axis represents log10 hazard ratio with 95% confidence interval derived from ICI data.