Marcel Wiesweg1, Fabian Mairinger2, Henning Reis2, Moritz Goetz2, Jens Kollmeier3, Daniel Misch3, Susann Stephan-Falkenau4, Thomas Mairinger4, Robert F H Walter5, Thomas Hager2, Martin Metzenmacher6, Wilfried E E Eberhardt6, Gregor Zaun7, Johannes Köster8, Martin Stuschke9, Clemens Aigner10, Kaid Darwiche11, Kurt W Schmid12, Sven Rahmann13, Martin Schuler14. 1. Department of Medical Oncology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; Division of Thoracic Oncology, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany; Genome Informatics, Institute of Human Genetics, University Hospital Essen, University Duisburg -Essen, Hufelandstrasse 55, 45122 Essen, Germany. 2. Institute of Pathology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany. 3. Department of Pneumology, Heckeshorn Lung Clinic, Walterhöferstraße 11, 14165 Berlin, Germany. 4. Institute of Pathology, Helios Klinikum Emil von Behring, Walterhöferstraße 11, 14165 Berlin, Germany. 5. Institute of Pathology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; Department of Pulmonary Medicine, Section of Interventional Pneumology, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany. 6. Department of Medical Oncology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; Division of Thoracic Oncology, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany. 7. Department of Medical Oncology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany. 8. Genome Informatics, Institute of Human Genetics, University Hospital Essen, University Duisburg -Essen, Hufelandstrasse 55, 45122 Essen, Germany. 9. Department of Radiotherapy, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Hufelandstrasse 55, 45122 Essen, Germany. 10. Department of Thoracic Surgery and Endoscopy, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany; German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Hufelandstrasse 55, 45122 Essen, Germany. 11. Department of Pulmonary Medicine, Section of Interventional Pneumology, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany. 12. Institute of Pathology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Hufelandstrasse 55, 45122 Essen, Germany. 13. Genome Informatics, Institute of Human Genetics, University Hospital Essen, University Duisburg -Essen, Hufelandstrasse 55, 45122 Essen, Germany; German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Hufelandstrasse 55, 45122 Essen, Germany. 14. Department of Medical Oncology, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany; Division of Thoracic Oncology, University Medicine Essen - Ruhrlandklinik, University Duisburg-Essen, Tüschener Weg 40, 45239 Essen, Germany; German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Hufelandstrasse 55, 45122 Essen, Germany. Electronic address: martin.schuler@uk-essen.de.
Abstract
OBJECTIVE: Current predictive biomarkers for PD-1 (programmed cell death protein 1)/PD-L1 (programmed death-ligand 1)-directed immunotherapy in non-small cell lung cancer (NSCLC) mostly focus on features of tumour cells. However, the tumour microenvironment and immune context are expected to play major roles in governing therapy response. Against this background, we set out to apply context-sensitive feature selection and machine learning approaches on expression profiles of immune-related genes in diagnostic biopsies of patients with stage IV NSCLC. METHODS: RNA expression levels were determined using the NanoString nCounter platform in formalin-fixed paraffin-embedded tumour biopsies obtained during the diagnostic workup of stage IV NSCLC from two thoracic oncology centres. A 770-gene panel covering immune-related genes and control genes was used. We applied supervised machine learning methods for feature selection and generation of predictive models. RESULTS: Feature selection and model creation were based on a training cohort of 55 patients with recurrent NSCLC treated with PD-1/PD-L1 antibody therapy. Resulting models identified patients with superior outcomes to immunotherapy, as validated in two subsequently recruited, separate patient cohorts (n = 67, hazard ratio = 0.46, p = 0.035). The predictive information obtained from these models was orthogonal to PD-L1 expression as per immunohistochemistry: Selecting by PD-L1 positivity at immunohistochemistry plus model prediction identified patients with highly favourable outcomes. Independence of PD-L1 positivity and model predictions were confirmed in multivariate analysis. Visualisation of the models revealed the predictive superiority of the entire 7-gene context over any single gene. CONCLUSION: Using context-sensitive assays and bioinformatics capturing the tumour immune context allows precise prediction of response to PD-1/PD-L1-directed immunotherapy in NSCLC.
OBJECTIVE: Current predictive biomarkers for PD-1 (programmed cell death protein 1)/PD-L1 (programmed death-ligand 1)-directed immunotherapy in non-small cell lung cancer (NSCLC) mostly focus on features of tumour cells. However, the tumour microenvironment and immune context are expected to play major roles in governing therapy response. Against this background, we set out to apply context-sensitive feature selection and machine learning approaches on expression profiles of immune-related genes in diagnostic biopsies of patients with stage IV NSCLC. METHODS: RNA expression levels were determined using the NanoString nCounter platform in formalin-fixed paraffin-embedded tumour biopsies obtained during the diagnostic workup of stage IV NSCLC from two thoracic oncology centres. A 770-gene panel covering immune-related genes and control genes was used. We applied supervised machine learning methods for feature selection and generation of predictive models. RESULTS: Feature selection and model creation were based on a training cohort of 55 patients with recurrent NSCLC treated with PD-1/PD-L1 antibody therapy. Resulting models identified patients with superior outcomes to immunotherapy, as validated in two subsequently recruited, separate patient cohorts (n = 67, hazard ratio = 0.46, p = 0.035). The predictive information obtained from these models was orthogonal to PD-L1 expression as per immunohistochemistry: Selecting by PD-L1 positivity at immunohistochemistry plus model prediction identified patients with highly favourable outcomes. Independence of PD-L1 positivity and model predictions were confirmed in multivariate analysis. Visualisation of the models revealed the predictive superiority of the entire 7-gene context over any single gene. CONCLUSION: Using context-sensitive assays and bioinformatics capturing the tumour immune context allows precise prediction of response to PD-1/PD-L1-directed immunotherapy in NSCLC.