37P - Can we perform subtyping of non-small cell lung cancer patients by use of lung tissue metabolic fingerprinting?

Date 07 May 2017
Event ELCC 2017
Session Poster Display Session
Topics Thoracic malignancies
Non-small-cell lung cancer
Presenter Michal Ciborowski
Citation Annals of Oncology (2017) 28 (suppl_2): ii9-ii13. 10.1093/annonc/mdx089
Authors M. Ciborowski1, J. Kisluk2, K. Pietrowska1, P. Samczuk1, E. Parfieniuk1, T. Kowalczyk1, M. Kozlowski3, A. Kretowski1, J. Niklinski2
  • 1Clinical Research Centre, Medical University of Bialystok, 15-276 - Bialystok/PL
  • 2Department Of Clinical Molecular Biology, Medical University of Bialystok, 15-269 - Bialystok/PL
  • 3Department Of Thoracic Surgery, Medical University of Bialystok, 15-276 - Bialystok/PL

Abstract

Background

In case of NSCLC patients targeted agents (TAs) and newer chemotherapeutics yielded differences in outcomes according to histologic subgroups. About 62% of adenocarcinoma (ADC) patients have driver mutations in targeted oncogenes and TAs can be successfully administrated to these patients. Similar mutations are rarely observed in squamous cell carcinoma (SCC) and TAs are currently unavailable for this group of patients. A correct histologic classification is important for treatment planning and novel therapeutic options are awaited for SCC patients. Metabolomics has potential for revealing metabolic pathways altered by disease and for biomarkers discovery. The purpose of this study was to evaluate whether LC-MS metabolomics of lung tumor tissue can be used to classify NSCLC patients according to histological subtype: ADC, SCC or large cell carcinoma (LCC).

Methods

The study was performed in the group of 25 NSCLC patients (62±8 years old, BMI=25±2, 32% females) who underwent curative tumor surgery. The patients were classified as ADC (n = 8), LCC (n = 4), and SCC (n = 13). Metabolites were extracted from tumor and adjacent control lung tissue and analyzed by LC-QTOF-MS. Differences between tumor and control tissue were evaluated by paired t-test or Wilcoxon rank test. One-way ANOVA was used to select metabolites discriminating NSCLC subtypes. Multivariate analysis was used for samples classification.

Results

Control and tumor tissue samples were well separated on OPLS-DA model (R2=0.96, Q2=0.86). The main metabolites responsible for separation were lysophospholipids (+51-384%, p-value 0.02-0.0004) and acylcarnitines (+59-477%, 0.05-0.0007). PLS-DA model showed good separation of NSCLC subtypes (R2=0.99, Q2=0.7). Among metabolites responsible for this separation sphingosine (p = 0.03), four acylcarnitines (p-value 0.03-0.001), Lyso PEs 18:1 (p = 0.04) and 16:0 (p = 0.03) as well as Lyso PC 16:1 (p = 0.01) can be mentioned.

Conclusions

NSCLC subtypes can be discriminated based on tumor tissue metabolic fingerprinting. Obtained results are preliminary and require further validation in a larger group of patients.

Clinical trial identification

This study is a research project which was approved by the Ethics Committee of the Medical University of Bialystok (No. R-I-003/262/2004).

Legal entity responsible for the study

Medical University of Bialystok

Funding

National Science Centre, Poland (2014/13/B/NZ5/01256).

Disclosure

All authors have declared no conflicts of interest.