Bladder cancer (BC) is the 10th most common cancer in the world. It has no available screening methods for the general population. This leaves diagnosis to be made late on, with the start of symptoms and this is reflected in the poor survival rate.
The use of VOCs that emanate from human waste is providing new perspective for the early detection cancer. VOCs emitted from the human body can be qualitatively and quantitatively measured. Gas chromatography-mass spectrometry (GC-MS) is considered as the golden standard but, due to its high cost and long analysis time, it may not be as efficient for screening. An alternative method is to use an electronic nose, which is designed to mimic the biological olfactory system. This instrument provides real time analysis, is less expensive and easily portable. This study describes the use of an electronic nose to measure VOC biomarkers emanating from urine to diagnose BC.
Here an AlphaMOS FOX 4000 electronic nose was used to analyse samples from 15 BC and 41 non-BC patients, recruited from University Hospital Coventry and Warwickshire, UK. The FOX 4000 consisted of 18 Metal-oxide based sensors. The change in resistance before and after exposure to a urine sample was used as the feature and the data was processed in ‘R’ using two machine learning classifiers. The data was analysed using a 10-fold cross validation with the resultant probabilities used to calculate statistical parameters. The results showed a high separation between BC and non-cancerous with sensitivity and specificity of 0.93 and 0.88 respectively using Sparse Logistic Regression and 0.93 and 0.76 using Random Forest classifier. We concluded that the electronic nose shows potential for discriminating bladder cancer from non-cancer samples. This is the first study in our knowledge which used the AlphaMOS FOX 4000 for this purpose.
