Metabolite Profiling of Blood Serum for the Early Detection of Recurrent Breast Cancer

We report here on the development of a monitoring test for recurrent breast cancer using metabolite profiling by combining powerful bio-analytical and multivariate statistical methods. We applied a combination of two dimensional gas chromatography-mass spectrometry (GCxGC-MS) and nuclear magnetic resonance (NMR) spectroscopy methods to investigate metabolite profiles of 257 serial blood serum samples from breast cancer patients. A total of 116 samples were from recurrent breast cancer patients and 141 samples from breast cancer patients with no evidence of disease (NED). Data were analyzed to identify metabolite markers that distinguish patients with cancer recurrence from those without the recurrence. A set of eleven bio-markers (7 from NMR and 4 from GCxGC-MS) were shortlisted from an analysis of logistic regression model using 5-fold cross validation. A PLS-DA model built using these markers provided a sensitivity of 86% and a specificity of 84% (AUROC =0.88). One of the important outcomes of this study is that 55% of the patients correctly predicted to have recurrence on average 13 months before clinical diagnosis, which represents an improvement over the current breast cancer monitoring assay CA 27.29. This study clearly demonstrates that the combination of two advanced bio-analytical methods, GCxGC-MS and NMR, provides a powerful approach to develop a blood based diagnostic test for the early detection of recurrent breast cancer. Validation of the results of this study will pave a new way for earlier detection of breast cancer and creates the potential for a simple and sensitive test that may well improve patient survival.