The Analysis of Small Molecule Metabolite Profiles in the Blood as a Biomarker of Lung Cancer.

SESSION TITLE: Lung Cancer Posters ISESSION TYPE: Original Investigation PosterPRESENTED ON: Wednesday, October 29, 2014 at 01:30 PM - 02:30 PMPURPOSE: Determine if differences exist in the profile of small molecule metabolites in the serum of patients with stage I-III non-small cell lung cancer when compared to matched controls.METHODS: Serum of patients with biopsy confirmed untreated stage I-III non-small cell cancer and at risk controls was collected. 2 controls were selected for each lung cancer patient through propensity matching that included age, gender, smoking history, COPD, DM, and lipids. Samples were prepared and divided into two fractions, one for analysis by UPLC-MS/MS and one for analysis by GC/MS. Compounds were identified by comparison to library entries of purified standards or recurrent unknown entities. Two-sample t-tests were used for comparison, with the level of significance adjusted for multiple comparisons. A prediction model was developed using random forest methodology.RESULTS: The serum of 284 subjects was analyzed. The mean age was 67 and 48% were female. 94 patients had lung cancer (50 adenocarcinoma, 44 squamous cell), 44% stage I, 17% stage II, and 39% stage III. Cancer patients were slightly older (68.7 vs. 66.2, p=0.013); all other matched variables were not significantly different. 534 metabolites were identified in 8 metabolite super-pathways and 73 sub-pathways. The concentration of 149 metabolites differed significantly (q-values < 0.05) between the cancer and control groups (70 lower, 79 higher). Reductions in phenolic compounds, elevated transulfuration pathway activity, and elevations of fatty acids were noted. A model containing 36 metabolites had an AUC of 0.837 (sensitivity 70.2%, specificity 89.5%) for distinguishing lung cancer from control.CONCLUSIONS: Patients with stage I-III non-small cell lung cancer have different serum metabolite profiles than matched controls.CLINICAL IMPLICATIONS: Metabolite profiles point to altered lung cancer metabolic processes. They may be developed into a diagnostic biomarker.DISCLOSURE: Robert Mahoney: Employee: Employee of the commercial entity that processed the specimens Tracy Holt: Employee: Employee of the commercial entity that processed the specimens Kirk Pappan: Employee: Employee of the commercial entity that processed the specimens The following authors have nothing to disclose: Peter Mazzone, Xiaofeng Wang, Mary Beukeman, Qi Zhang, Meredith SeeleyThe project aimed to determine if there are serum metabolites that are different in lung cancer patients when compared to matched controls. This was a research project and is not approved for any purpose.