Imaging MS and Ion Mobility-MS Metabolomics for Detecting Ovarian Cancer

FACUNDO M. FERNANDEZ; DAVID A. GAUL; CHRISTINA M. JONES; MARÍA EUGENIA MONGE; MARTIN R. L. PAINE; JAEYEON KIM; MARTIN M. MATZUK; LONG Q. TRAN; ROMAN MEZENCEV; JOHN F. MCDONALD

Brisbane

Conferencia; 25th Australian and New Zealand Society for Mass Spectrometry and 6th Asia Oceania Mass Spectrometry Conference; 2015

Ovarian cancer (OC) is the fifth leading cause of cancer-related deaths for U.S. women, yet it has the highest mortality rate amongst gynaecological cancers. Lack of symptoms as well as the deficiency of highly specific biomarkers has resulted in only a quarter of OC cases being diagnosed at FIGO stage I. An effective screening strategy for early diagnosis would be particularly advantageous since 5-year survival rates can be as high as 90%.To explore the feasibility of mass spectrometry (MS) metabolome-based OC detection, early stage and normal serum samples were collected and analysed using ultra performance liquid chromatography coupled with ion mobility-mass spectrometry. Metabolites were identified by tandem mass spectrometry and, whenever possible, chromatographic matching with standards. After normalization, alignment, and extraction of spectral features, multivariate statistical analysis employing support vector machine learning methods coupled to recursive feature elimination selected panels of metabolites that differentiated age-matched samples with excellent accuracy. In parallel, tissue samples were examined by matrix-assisted laser desorption/ionization and desorption electrospray ionization MS to find correlations with biofluids.Comparison of metabolic phenotypes of OC with normal metabolic signatures revealed unique alterations in the serum metabolome. One study compared early-stage serous papillary (n=46) vs. normal human patients (n=49). Serous papillary OC is the most commonly diagnosed histopathological subtype of OC, and for high grade serous carcinomas (HGSCs), the most deadly. A second study investigated metabolome alterations in a DKO mouse model of ovarian HGSC (n=14 vs. 11 controls), as a surrogate for studying the disease in humans. A third study focused on more rare types of ovarian cancer such as endometrioid, clear cell, and mucinous (n=23) vs. normal patients (n=40). In all three cases, panels consisting of multiple serum metabolites were found to differentiate between early-stage OC and normal with accuracy, sensitivity, and specificity higher than 90%. Employing 2-dimensional MS imaging on ovarian tissues confirmed the source of biomarkers detected in sera and provided further insight into the biological pathways involved in tumour development. Overall, these results demonstrated that serum metabolites may be useful for detecting early-stage OC and supports conducting studies with larger cohorts.