Getting reliable information for disease detection is a major challenge for existing diagnostic technologies, typically requiring invasive and time-consuming tests involving bodily fluids such as blood and urine. Over the past 5-10 years, molecules in the breath have been identified as biomarkers of certain diseases such as bladder or lung cancers, specific infectious diseases, and neurodegenerative diseases. Dogs have been trained to quickly detect these biomarkers through smell, and they help clinicians identify affected patients. However, the training process is time consuming, labor intensive, and not suitable for scale-up. Meanwhile, efforts to mimic the performance of canine olfactory systems typically involve large sensor arrays that require careful calibration or utilize expensive and laborious nanofabrication. This project aims to create a portable breathalyzer with excellent diagnostic potential, which we will develop for the beachhead markets of lung cancer and malaria detection. The device consists of an array of off-the-shelf and cost-effective chemiresistive sensors, which are in themselves limited by noise and inaccuracies, but are elevated through the use of our patented bio-inspired machine learning (ML) algorithms based on ‘sniffing sequences.

​

​

​

​