Evaluation of a combination of SIFT-MS and multivariate data analysis for the diagnosis of Mycobacterium bovis in wild badgers

The currently accepted 'gold standard' tuberculosis (TB) detection method for veterinary applications is that of culturing from a tissue sample post mortem. The test is accurate, but growing Mycobacterium bovis is difficult and the process can take up to 12 weeks to return a diagnosis. In this paper we evaluate a much faster screening approach based on serum headspace analysis using selected ion flow tube mass spectrometry (SIFT-MS). SIFT-MS is a rapid, quantitative gas analysis technique, with sample analysis times of as little as a few seconds. Headspace from above serum samples from wild badgers, captured as part of a randomised trial, was analysed. Multivariate classification algorithms were then employed to extract a simple TB diagnosis from the complex multivariate response provided by the SIFT-MS instrument. This is the first time that such multivariate analysis has been applied to SIFT-MS data. An accuracy of TB discrimination of approximately 88% true positive was achieved which shows promise, but the corresponding false positive rate of 38% indicates that there is more work to do before this approach could replace the culture test. Recommendations for future work that could increase the performance are therefore proposed.