Beyond the usual suspects.

All aspects of lung cancer care will benefit from the development of accurate molecular biomarkers. Those that can predict the risk of developing lung cancer will help in the selection of subjects for lung cancer screening. Those that can identify the presence of earlystage lung cancer will help in the evaluation of indeterminate lung nodules. Those capable of prognosticating outcome will assist in planning the aggressiveness of treatment. Those capable of predicting the response to systemic therapies will help to individualize their selection. Over the past decade, molecular biomarker discovery has been dominated by studies assessing the genome, transcriptome, and proteome. Alterations of individual molecules, or patterns of multiple alterations, have been presented as potential biomarkers. To date, with the exception of epidermal growth factor receptor mutation testing 1 and echinoderm microtubule-associated protein-like 4 anaplastic lymphoma kinase identification, 2 none of the molecular biomarkers have been developed to the point of technical and clinical validation, and/or proven clinical utility. Metabolomics is a source of potential molecular biomarkers, which has been relatively under-represented in the literature. Metabolomics refers to an assessment of the products of cellular metabolism. In a systems biology approach, the metabolome can be considered the product of the interaction between the genome, transcriptome, and proteome with an individual’s physiologic status. Traditionally, metabolomics involves the study of nonvolatile endogenous small molecule metabolites such as amino acids, carbohydrates, and lipids, in tissue, blood, or urine. Nonvolatile metabolite profiling can be performed as an overview of the metabolome, or as a targeted look at a metabolic pathway, or a specific metabolite. To date, published studies of nonvolatile metabolite profiling in lung cancer have been small, pilot studies. Differences in metabolite signatures between tumor and adjacent control tissue, 3 as well as serum and urine of lung cancer patients compared with healthy controls, 4–6 have been reported. In addition to nonvolatile metabolites, volatile organic compounds (VOCs) can be produced or degraded as part of cellular metabolism and the handling of oxidative stresses. They can be identified in the headspace gas of tissues, 7 blood, 8 urine, 9 and in the exhaled breath. 10 They can be evaluated for differences in a specific compound, class of compounds, or as a general overview of the spectrum of VOCs present in the biospecimen. VOC profiles in lung cancer have been analyzed with gas chromatography mass spectrometry (GC-MS), other spectrometry technologies, and cross-responsive chemical sensor matrices. GC-MS is capable of identifying selected VOCs and their concentrations, whereas the signal from the sensor technologies reflects the broad mixture of VOCs within the specimen without identifying the specific components of the mixture. These studies have been relatively small and highly variable in their methods, but have shown promising accuracy. In this issue of the Journal Thoracic Oncology, Pelid et al. report on the development of a breath VOC biomarker identified by a cross-responsive sensor matrix using a combination of a gold nanoparticle sensor and a single-walled carbon nanotube sensor, as well as evaluation of selected breath samples with GC-MS. They sampled alveolar breath from 53 lung cancer subjects with early- and late-stage disease as well as 19 control subjects with benign lung nodules. The groups were well matched for smoking histories,