Objectives: Lung cancer (LC) has short life expectancy in advanced stages, so an early diagnosis is necessary. We apply a metabolomic approach based on direct infusion of high resolution mass spectrometry (DI-ESI-QTOF-MS) to identify altered metabolites as potential biomarkers of the LC diagnosis. Materials and methods: Serum samples handling consist in the extraction of polar and non-polar metabolites and the use of electrospray ionization mass spectrometry in positive and negative mode. The approach was applied to 21 serum samples from LC and 21 healthy controls (C). Results: Partial least squares discriminant analysis (PLS-DA) presented a clear classification between LC and C groups and a preliminary classification between different histological types of LC. A total of 34 metabolites, including amino acids, fatty acids, lysophospholipids, phospholipids and triacylglycerides were identified. Several metabolic pathways showed the highest impact (i)glycine, serine and threonine; (ii)cysteine and methionine, (iii)arginine and proline, and (iv)glycerophospholipid metabolism. So, analysis of ROC (receiver operator characteristic) curves of urea, L-threonine, L-ornithine, LPC (18:2) and LPC(20:3) showed AUC values greater than 0.75 Conclusions: A robust metabolomic technique DI-ESI-QTOF-MS has been applied to serum samples from LC and C. PLS-DA showed good classification between groups, being possible to identify 34 altered metabolites involved in different metabolic routes. The Pathway and ROC curves analysis stressed aminoacids and phospholipids as the most altered metabolites in LC being urea, L-threonine, L-ornithine, LPC (18:2) and LPC (20:3) potential LC biomarkers.