br Introduction br Globally breast cancer BC is the second
Globally, breast cancer (BC) is the second most common type of cancer and a major cause of human morbidity and mortality, dis-proportionately affecting women . It is reported that BC alone
accounts for 25% of all cancer cases and 15% of all cancer deaths among females . The American Cancer Society (2018) strongly re-commends that women with an average risk of developing BC, as de-termined by a family history-based risk assessment, undergo regular screening mammography beginning at age 45 . Although
Abbreviations: BC, breast cancer; PLS-DA, partial least squares-discriminant analysis; ROC, receiver operating characteristic; miRNA, microRNA; FHCRC, Fred Hutchinson Cancer Research Center; QC, quality control; GLM, general linear model; FDR, false discovery rate; VIP, variable importance in projection; AUROC, area under receiver operating characteristic; EFA, exploratory factor analysis; IPA, Ingenuity Pathway Analysis; ER, Z-Gly-Gly-Arg-AMC acetate receptor; PR, progesterone receptor; HER2, human epidermal growth factor receptor 2; CV, coefficient of variation; FC, fold change; CI, confidence interval; OPLS-DA, orthogonal partial least squares-dis-criminant analysis; Prodh, proline dehydrogenase
Corresponding author at: Arizona Metabolomics Laboratory, College of Health Solutions, Arizona State University, 13208 E. Shea Blvd, Scottsdale, AZ 85259, USA.
E-mail address: [email protected] (H. Gu).
1 Author contributions: PJ and DW contributed equally to this project.
Available online 03 December 2018
mammography has been shown to have high sensitivity (93%) for the detection of symptomatic BC , it is far less effective for the detection of early-stage BC. As such, it has been suggested that regular physical examination is of comparable importance, and perhaps the best method of early detection . Sentinel lymph node biopsy remains the gold standard for detection of BC with lymph node involvement, but the major disadvantages include invasiveness, potential risk of complica-tion, and the inherent inability for detection of early-stage BC . The average 5-year BC survival rate is roughly 90%, but can be as high as 99% for those diagnosed and treated with early-stage, localized disease (stages I and II), which regrettably accounts for only 61% of BC patients . Therefore, noninvasive detection methods with high sensitivity and specificity, which would enable the early diagnosis and timely treat-ment of BC, are still critically needed.
A number of new BC detection methods have been developed at the molecular level, particularly based on genetic/phenotypic testing, computer-aided technology and biomarker identification, such as im-munohistochemistry , and serum circulating microRNA (miRNA) profiling [9,10], which have shown some favorable evidence of en-hanced BC detection. Although efforts to typify ER/PR and HER2 status in order to predict BC pathogenesis and disease progression are pro-mising methods of early detection, they often lead only to the assertion of broad probabilities .
A characteristic feature of cancer is its abnormal metabolism. Consequently, recent cancer studies have endeavored to monitor levels of differential metabolites from biologically relevant pathways, re-sulting in improved BC subtype identification and diagnosis . Me-tabolomics, the scientific study of comprehensive sets of metabolites present in biological samples, offers new avenues for advanced disease biomarker discovery [11–27]. Mass spectrometry-based metabolic profiling has emerged as a powerful analytical platform for analysis of metabolic alterations caused by various cancers, which has led to substantial advances in cancer diagnosis, pathogenesis clarification, and identification of potential drug targets for clinical treatment [28–30]. Previous efforts to typify an associated metabolic profile of BC have typically assumed global profiling approaches to differentiate disease patients from healthy controls, employing gas chromato-graphy–mass spectrometry (GC–MS) , nuclear magnetic resonance (NMR) spectroscopy , flow injection analysis-tandem mass spec-trometry (FIA-MS/MS) , and liquid chromatography time-of-flight mass spectrometry (LC-TOF-MS) . In contrast to global approaches, some metabolomics studies, citing well-known alterations in cancer metabolism such as the Warburg effect and glutamine addiction, have argued for MS-based testing of particular metabolites associated with decreased oxidative phosphorylation and increased glycolysis and lactic acid fermentation , while others have emphasized the need to therapeutically target glutamine addiction as breast cancer cells require glutamine that is vital to cancer cell growth and proliferation . These metabolic differences, which can be hallmarks of all cancers, also suggest that metabolomics is a highly promising approach to discover significant risk factors for BC and develop sensitive and specific BC biomarkers.