Prediction of Chemotherapeutic Efficacy in Non-Small Cell Lung Cancer by Serum Metabolomic Profiling

Purpose: No validated biomarkers that could identify the subset of patients with lung adenocarcinoma who might benefit from chemotherapy have yet been well established. This study aimed to explore potential biomarker model predictive of efficacy and survival outcomes after first-line pemetrexed plus platinum doublet based on metabolomics profiling. Experimental Design: In total, 354 consecutive eligible patients were assigned to receive first-line chemotherapy of pemetrexed in combination with either cisplatin or carboplatin. Prospectively collected serum samples before initial treatment were utilized to perform metabolomics profiling analyses under the application of LC/MS-MS. Binary logistic regression analysis was carried out to establish discrimination models. Results: There were 251 cases randomly sorted into discovery set, the rest of 103 cases into validation set. Seven metabolites including hypotaurine, uridine, dodecanoylcarnitine, choline, dimethylglycine, niacinamide, and L-palmitoylcarnitine were identified associated with chemo response. On the basis of the seven-metabolite panel, a discriminant model according to logistic regression values g(z) was established with the receiver operating characteristic curve (AUC) of 0.912 (Discovery set) and 0.909 (Validation set) in differentiating progressive disease (PD) groups from disease control (DC) groups. The median progression-free survival (PFS) after chemotherapy in patients with g(z) <= 0.155 was significantly longer than that in those with g(z) > 0.155 (10.3 vs. 4.5 months, P < 0.001). Conclusions: This study developed an effective and convenient discriminant model that can accurately predict the efficacy and survival outcomes of pemetrexed plus platinum doublet chemotherapy prior to treatment delivery. (C) 2018 AACR.