In a post a few weeks ago, we highlighted HeroRats, African-pouch rats, and their use for TB diagnoses. As described in that article, these rats are trained to distinguish positive TB samples verses negative TB samples presumably due to the release of volatile organic compounds (VOC) produced by M. tb. Not surprisingly, review of the scientific literature on VOCs shows that VOCs are produced by various diseases and infections, including TB.
A recent study published in Tuberculosis by Phillips et al. investigated the ability to identify biomarkers of active pulmonary tuberculosis from the breath of human patients. Sputum samples, chest x-rays and breath samples including medical histories of 226 patients from the US, Philippines, and the UK were collected. Triplicate smears and cultures were made of each sputum sample. Breath VOCs were collected by a portable breath collection apparatus which contained sorbent traps. The sorbent traps were then subjected to automated thermal desorption, gas chromatography and mass spectroscopy to separate the volatile compounds trapped in the sorbent traps. Sputum cultures were used as the ‘gold standard’.
Results obtained suggested an ~85% accuracy rate of detection when sputum smears, cultures, and chest x-rays were all positive or all negative. Accuracy for breath VOCs decreased when single criterion for disease, smear, culture or chest x-ray were used. Of interesting note, agreement of positive verses negative between sputum smear, culture and chest x-rays was low, highlighting the existing challenge in accurately diagnosing patients for TB treatment. Only sputum smear and cultures displayed moderate agreement, whereas chest x-rays displayed very little correlation with sputum smear or culture.
According to the results, a series of potential breath VOC biomarkers were identified which were similar but not the same as breath VOCs identified in a previous study by Phillips et al. in 2007. These compounds included derivatives of cyclohexane, benzene, decane and heptanes. However, at present the challenge still remains to identify specific to M. tb VOCs in abundant concentrations with instrumentation that is less expensive and technical than a GC/MS. The main author, Michael Phillips is President and CEO of a Messana Research, Inc., a company which specializes in the development of breath tests for disease.
An alternative diagnostic tool to distinguish VOCs produced by M. tb may be found in electronic nose technology. Electronic noses are gas sensor arrays which are designed to analyze and characterize complex gas mixtures similar to the human olfactory system. In 2004, Pavlou et al. published a paper on the ability of using an electronic nose in combination with a neural network system (think artificial intelligence, the ability for the electronic nose to be ‘trained’ and ‘learn’) to distinguish between positive and negative TB samples. Experiments performed on small samples sizes suggested that the electronic nose could detect positive TB samples with 96-100% accuracy.
Fend et al. in 2006 expanded the use of electronic nose technology to a larger cohort of patient samples. The authors determined the detection limit to be 1 X 104 mycobacterial ml-1. Unfortunately, a complete separation of negative verses positive samples showed some overlap between negative and positive TB samples. Overall, the electronic nose had an 89% accuracy rate for positive TB identification, and was ~90% specific and sensitive compared to culture samples. Whereas GC/MS can separate and identify individual VOCs, the electronic nose identifies a pattern or signature of combined VOCs for a particular disease or bacteria in this case. The cost of electronic nose technology is unclear; however its reported simplicity and sensitivity are encouraging for further development.
In conclusion, the ability of mycobacterium to release VOCs which can be identified as a unique signature offers tremendous potential for new diagnostics that would enable faster detection of TB in patients and sooner treatment.
Compared to the use of African-pouched rats, do the two technologies presented here, breath analysis using GC/MS or electronic noses, offer any advantages? Disadvantages? What limitations do you foresee for either technology? Could costs be brought low enough to allow for wide-distribution and use? What additional thoughts or comments do you have? Please share your comments below.