Breath mass ion biomarkers of breast cancer. (9th January 2017)

• Record Type: Journal Article
• Authors: Phillips, Michael; Cataneo, Renee N; Lebauer, Cassie; Mundada, Mayur; Saunders, Christobel
• Is Part Of: Journal of breath research. Volume 11:Number 1(2017:Mar.)
• Publication Date: 2017-01-09
• Subjects: breath -- volatile organic compound -- breast cancer -- mass ions -- cancer detection; Volatile organic compounds -- Analysis -- Periodicals; Clinical chemistry -- Periodicals; Bad breath -- Periodicals; Bad breath -- Treatment -- Periodicals; Bad breath -- Diagnosis -- Periodicals; 616.0756
• DOI: 10.1088/1752-7163/aa549b
• Languages: English
• ISSNs: 1752-7155

Abstract

Abstract: Breath volatile organic compounds (VOCs) contain biomarkers of breast cancer that are detectable with gas chromatography mass spectrometry (GC MS). However, chemical identification of breath VOC biomarkers may be erroneous because spectral matching can misidentify their structure. Breath mass ions detected with GC MS have been proposed as intrinsically robust biomarkers because they can be identified without spectral matching. We investigated whether breath mass ion biomarkers could identify breast cancer. We re-analyzed data from a previous study of breath VOCs in 54 women with biopsy-proven breast cancer and in 204 healthy controls. Subjects were randomly assigned to a training set (2/3) and a test set (1/3). Chromatograms were processed with metabolomic analysis software (XCMS in R) in order to generate a table listing retention times with their associated ion masses and intensities, and binned into a series of 5 s retention time segments. In the training set, mass ions in each time segment were ranked according to their diagnostic accuracy i.e. the area under curve (AUC) of the receiver operating characteristic (ROC) curve. We employed multiple Monte Carlo simulations to select the biomarker mass ions in each time segment that identified breast cancer with greater than random accuracy and combined those with the highest diagnostic accuracy in a predictive algorithm using multivariate weighted digital analysis (WDA). We then employed this algorithm to predict the diagnosis in the test set. The training set WDA algorithm employing 21 mass ion biomarkers identified breast cancer with ROC curve AUC = 0.79. In the test set, this algorithm predicted breast cancer with ROC curve AUC = 0.77. Breath mass ions biomarkers accurately identified women with breast cancer and could potentially be used in early diagnosis and treatment monitoring.