RT Journal Article SR Electronic T1 Metabolomic profiling of clinical sputum samples reveals novel biomarkers for the early identification of lung cancer patients JF European Respiratory Journal JO Eur Respir J FD European Respiratory Society SP P350 VO 44 IS Suppl 58 A1 Simon Cameron A1 Keir Lewis A1 Manfred Beckmann A1 Gordon Allison A1 Robin Ghosal A1 Paul Lewis A1 Luis Mur YR 2014 UL http://erj.ersjournals.com/content/44/Suppl_58/P350.abstract AB INTRODUCTION: Developing novel screening and diagnosis methodologies will allow earlier diagnosis of lung cancer (LC), increasing the effectiveness of clinical interventions. We tested the potential of metabolomic profiling on sputum to differentiate between patients with and without LC.METHODS: Patients: We collected and processed the spontaneous sputum of 34 patients with suspected LC, alongside 33 healthy controls from staff at Swansea University, with no history of lung disease. Of the 34 patients, 23 were subsequently diagnosed with lung cancer (16 NSCLC, six SCLC, and one radiological diagnosis) at various stages of disease progression. Procedure: The 67 samples were profiled using linear quadrupole ion trap mass spectrometery (MS), with metabolite identifications confirmed using gas chromatography MS and fourier transform ion cyclotron resonance MS.RESULTS: Hierarchical cluster and principal component analyses clearly separated the clinically and non-clinically acquired samples. Analysis based on area under the receiver operating characteristic curve (AUC) revealed differential metabolites for patients with and without LC, with a maximum AUC value of 0.8 for a metabolite ratio that is able to differentiate between those clinical samples with LC from those without.CONCLUSION: This preliminary analysis suggests sputum is a viable sample, and metabolomics has potential as a diagnostic and/or discriminator tool. Furthermore, it can identify specific key metabolites that could aid our understanding of the molecular pathogenesis of LC and possibly guide treatment targets.