Identification of diseases by their """"""""scent"""""""" has the potential to revolutionize diagnostics. We believe that many diseases have their unique """"""""scent"""""""", which can be identified by non-invasive, cheap methods. Each """"""""scent"""""""" is determined by unique patterns of specific volatile organic compounds (VOC). We have developed an """"""""electronic nose"""""""" (or e-nose) that can identify and discriminate different VOC pattens with high accuracy. In this proof-of-concept study to be implemented in Botswana, we will focus on the identification of M. tuberculosis (MTB) through the recognition of is VOC patterns under laboratory and field conditions. First, we will determine the sensitivity and specificity of the e-nose on cultures and sputum samples with known concentrations of MTB, as well as on clinical sputum samples from patients with confirmed pulmonary tuberculosis (PTB) under laboratory, controlled conditions. Then, we will test the performance of the e-nose on the identification of MTB on the breath of patients with active PTB. These data will be used to identify the VOC patterns that best screen and diagnose MTB under laboratory and field conditions, aimed at assisting the non-specialist health care provider in clinical decision making. Since the only requirement for an e-nose i the partial selectivity of the sensors, it is possible to assemble electronic noses withany available sensor technology. This study will be the first one to comprehensively determine the relative performance of a wide variety of sensors and of e-noses platforms build with different number of such sensors under different environmental conditions.
The diagnosis of clinical conditions through their scent by identifying their unique patterns of specific volatie organic compounds (VOC) has the potential to revolutionize diagnostics by providing rapid point-of- care diagnosis of diseases as varied as cancers or infections. Early detectio of diseases will allow the possibility of earlier, cheaper treatment leading to decreased morbidity and mortality. In addition, availability of cheap, non-invasive diagnosis could facilitate the implementation of large-scale screening programs for primary and secondary prevention. The impact of such technology can be even greater in low-income countries by circumventing the need for high-infrastructure and expensive treatment programs. In the case of tuberculosis, identification of its VOC pattern and the VOC patterns of the host/pathogen interaction in the breath can allow earlier diagnosis and treatment of the disease, identification of patients with high burden of disease, patients with higher likelihood of having poor outcomes, resistant strains, and co- infections with other pathogens. Earlier and better treatment and easier case detection would likely lead to decrease TB transmission at the community level.
