Articles
Development and validation of response markers to predict survival and pleurodesis success in patients with malignant pleural effusion (PROMISE): a multicohort analysis Ioannis Psallidas, Nikolaos I Kanellakis, Stephen Gerry, Marie Laëtitia Thézénas, Philip D Charles, Anastasia Samsonova, Herbert B Schiller, Roman Fischer, Rachelle Asciak, Robert J Hallifax, Rachel Mercer, Melissa Dobson, Tao Dong, Ian D Pavord, Gary S Collins, Benedikt M Kessler, Harvey I Pass, Nick Maskell, Georgios T Stathopoulos, Najib M Rahman
Summary
Background The prevalence of malignant pleural effusion is increasing worldwide, but prognostic biomarkers to plan treatment and to understand the underlying mechanisms of disease progression remain unidentified. The PROMISE study was designed with the objectives to discover, validate, and prospectively assess biomarkers of survival and pleurodesis response in malignant pleural effusion and build a score that predicts survival. Methods In this multicohort study, we used five separate and independent datasets from randomised controlled trials to investigate potential biomarkers of survival and pleurodesis. Mass spectrometry-based discovery was used to investigate pleural fluid samples for differential protein expression in patients from the discovery group with different survival and pleurodesis outcomes. Clinical, radiological, and biological variables were entered into least absolute shrinkage and selection operator regression to build a model that predicts 3-month mortality. We evaluated the model using internal and external validation. Findings 17 biomarker candidates of survival and seven of pleurodesis were identified in the discovery dataset. Three independent datasets (n=502) were used for biomarker validation. All pleurodesis biomarkers failed, and gelsolin, macrophage migration inhibitory factor, versican, and tissue inhibitor of metalloproteinases 1 (TIMP1) emerged as accurate predictors of survival. Eight variables (haemoglobin, C-reactive protein, white blood cell count, Eastern Cooperative Oncology Group performance status, cancer type, pleural fluid TIMP1 concentrations, and previous chemotherapy or radiotherapy) were validated and used to develop a survival score. Internal validation with bootstrap resampling and external validation with 162 patients from two independent datasets showed good discrimination (C statistic values of 0·78 [95% CI 0·72–0·83] for internal validation and 0·89 [0·84–0·93] for external validation of the clinical PROMISE score). Interpretation To our knowledge, the PROMISE score is the first prospectively validated prognostic model for malignant pleural effusion that combines biological and clinical parameters to accurately estimate 3-month mortality. It is a robust, clinically relevant prognostic score that can be applied immediately, provide important information on patient prognosis, and guide the selection of appropriate management strategies. Funding European Respiratory Society, Medical Research Funding-University of Oxford, Slater & Gordon Research Fund, and Oxfordshire Health Services Research Committee Research Grants. Copyright © 2018 Elsevier Ltd. All rights reserved.
Introduction Malignant pleural effusion is a common condition that affects 500–700 individuals per million population annually.1 The global prevalence of cancer is rising, and with improvements in systemic therapy that allow many patients to live longer, the burden of malignant pleural effusion is increasing.2–4 In addition to the clinical significance of this disorder, dissemination of cancer cells in the pleural cavity is a biological hallmark of highly metastatic malignancy regardless of the primary neoplasm, and current guidelines quote a median survival of 3–12 months.1 Pleurodesis, the most common procedure used to treat malignant pleural effusion, prevents fluid accumulation through induction of pleural
inflammation and fibrosis, succeeding in approximately 70% of patients assessed at 1 month.1,5 In the past 10 years, the types of cancer treatment and pleural procedures available for malignant pleural effusion have expanded, leading to improved stratification of patients and progressive individualisation of treatment, but also to diversification of outcomes such as pleurodesis success and survival.6 However, the factors that drive and define malignant progression, resistance to therapy, and mortality in malignant pleural effusion are poorly understood. To this end, a prognostic model using clinical information in combination with blood or pleural fluid biomarkers that predicts survival and pleurodesis response would be of importance for clinical
www.thelancet.com/oncology Published online June 13, 2018 http://dx.doi.org/10.1016/S1470-2045(18)30294-8
Lancet Oncol 2018 Published Online June 13, 2018 http://dx.doi.org/10.1016/ S1470-2045(18)30294-8 See Online/Comment http://dx.doi.org/10.1016/ S1470-2045(18)30361-9 Oxford Centre for Respiratory Medicine, Churchill Hospital, Oxford University Hospitals NHS Foundation Trust, Oxford, UK (I Psallidas PhD, N I Kanellakis MSc, R Asciak MD, R J Hallifax MRCP, R Mercer MRCP, Prof I D Pavord PhD, N M Rahman DPhil); Laboratory of Pleural and Lung Cancer Translational Research (I Psallidas, N I Kanellakis, R Asciak, R J Hallifax, R Mercer, M Dobson BSc, N M Rahman), Oxford Respiratory Trials Unit (I Psallidas, N I Kanellakis, R Asciak, R J Hallifax, R Mercer, N M Rahman), Mass Spectrometry Laboratory, Target Discovery Institute (M L Thézénas MSc, P D Charles PhD, R Fischer PhD, Prof B M Kessler PhD), Centre for Translational Immunology, CAMS-Oxford Institute (Prof T Dong PhD), Nuffield Department of Medicine, University of Oxford, Oxford, UK; Centre for Statistics in Medicine, Botnar Research Centre (S Gerry MSc, Prof G S Collins PhD), Bioinformatics Core, Department of Oncology (A Samsonova PhD), MRC Human Immunology Unit, Weatherall Institute of Molecular Medicine (Prof T Dong), National Institute for Health Research Oxford Biomedical Research Centre (N M Rahman), University of Oxford, Oxford, UK; Institute of Translational Biomedicine, St Petersburg State University, St Petersburg, Russia (A Samsonova);
1
Articles
Comprehensive Pneumology Center and Institute for Lung Biology and Disease, University Hospital, Ludwig-Maximilian University, Munich, Bavaria, Germany (H B Schiller PhD, G T Stathopoulos PhD); Helmholtz Center Munich, Member of the German Center for Lung Research, Munich, Bavaria, Germany (H B Schiller, G T Stathopoulos); Department of Cardiothoracic Surgery New York University Langone Medical Center, New York, NY, USA (Prof H I Pass PhD); Academic Respiratory Unit, School of Clinical Sciences, University of Bristol, Bristol, UK (Prof N Maskell DM); and Laboratory for Molecular Respiratory Carcinogenesis, Department of Physiology, Faculty of Medicine, University of Patras, Patras, Greece (G T Stathopoulos) Correspondence to: Dr Ioannis Psallidas, Oxford Centre for Respiratory Medicine, Churchill Hospital, Oxford University Hospitals NHS Foundation Trust, Oxford OX3 7LE, UK
[email protected]
See Online for appendix
Research in context Evidence before this study We searched PubMed for systematic reviews and research articles with high-throughput screening, retrospective analysis, development of clinical scores, and prospective validation of survival and treatment biomarkers in malignant pleural effusion published up to Dec 5, 2017, with the keywords “biomarkers”, “malignant pleural effusion (MPE)”, ‘’survival’’, and ‘’pleurodesis’’. We applied search filters for “adult” and “English”. No studies had been published that met our criteria. Currently, the factors that drive and define malignant progression, resistance to therapy, and mortality in malignant pleural effusion are poorly understood. A prognostic model using clinical information in combination with blood or pleural fluid biomarkers that predicts survival and pleurodesis response would be of importance for clinical management and improve our understanding of the pathobiology of this condition.
management, and it would improve our understanding of malignant pleural effusion pathobiology. Such a risk stratification system has been proposed: the LENT score predicts patients’ survival on the basis of tumour type, pleural fluid lactate dehydrogenase (LDH), Eastern Cooperative Oncology Group (ECOG) performance status, and blood neutrophil-to-lymphocyte ratio.7 The LENT score was developed with clinical data from three existing patient cohorts combined with biochemical measurements of pleural fluid and has not been prospectively assessed. This study reports the development and validation of the survival and pleurodesis response markers in malignant pleural effusion (PROMISE), the first score dedicated to the prospective assessment of the survival of patients with malignant pleural effusion simultaneously by use of clinical, biological, and radiological parameters.
Methods
Study design and population This multicohort study consisted of a three-step approach (discovery, validation, and prospective assessment) that included label-free quantitative proteomic analyses of cellfree pleural fluid from a prospectively recruited cohort (biological discovery), identification of prognostic clinical variables in published cohorts (clinical discovery), model development, model validation in two separate cohorts, and finally prospective assessment in two independent datasets from ongoing randomised studies in patients with malignant pleural effusion. We used clinical data and pleural fluid samples from three previous multicentre clinical trials, TIME-1,8 TIME-2,9 and TIME-3,10 which were prospectively collected between 2007 and 2014 (between March, 2007, and December, 2009, in TIME-1; between April, 2007, and January, 2009, in TIME-2; and between September, 2
Added value of this study This study, to our knowledge, is the first translational assessment combining discovery technology with high-quality clinical data from five different datasets for the development of a prognostic score for malignant pleural effusion. The results demonstrate the discovery, assessment, and prospective validation of a new prognostic score, including clinical data with pleural fluid biomarkers. Implications of all the available evidence There is now a compelling need for stratification of management of patients with malignant pleural effusion. A prognostic score predicting the 3-month risk of death will provide important information about patient prognosis and guide the selection of appropriate management strategies. Additionally, our results identify four biological pathways with the potential to guide new treatments of malignant pleural effusion.
2009, and June, 2014, in TIME-3). Patients at the first onset of malignant pleural effusion were recruited on identical, clinically robust criteria from hospitals in the UK, the USA, and Canada. All studies recruited patients with stage IV cancer and proven malignant pleural effusion who were at least 18 years old and with expected survival of at least 1 month. Consistent data collected for each study participant included type of primary cancer, pleurodesis outcome, chest x-ray findings, pleural fluid LDH, performance status, full blood count, haemoglobin, previous treatment (with either first-line chemotherapy or radiotherapy), forced expiratory volume in 1 s (FEV1), and C-reactive protein (CRP; the full list of variables evaluated is in the appendix p 3). Survival data for each patient were collected from the trials’ databases and cross-validated with a national registry. We used data from individuals from the Oxford Radcliffe Pleural Biobank (Oxford, UK) and SIMPLE study11 for external validation of the prognostic score. Pleural fluid samples were prospectively collected alongside clinical data. We collected the exact survival data for each patient from the trial databases and cross-validated with a national registry. Ethical and regulatory approval for the study was obtained by the South Central Oxford A Research Ethics Committee (REC reference number 15/SC/0186). Ethics approval for sample analysis for external validation was obtained before the laboratory investigations (Oxford Radcliffe Biobank ethics committee reference 17/A078).
Discovery phase The TIME-2 database was used for the discovery phase. Study participants were stratified on the basis of survival and pleurodesis outcomes. For survival, patients were categorised as having poor (
