Electronic Nose that sniffs out diseases

Sumaiya Afroj Muna
IEEE SB KUET
Published in
5 min readJan 13, 2021
Image from Freepik.com

Have you ever heard about doctor dogs?

Maybe I have guessed your answer. Because we commonly see search and rescue (SAR) dogs. But quite surprisingly, dogs can serve the medical sector also. Dogs are mostly known for their sense of smell. Their genetics and body structure makes them perfectly suited for sniffing. And after having extra and extreme training, they can detect diseases like Malaria, diabetics, cancer, Parkinson’s, Alzheimer’s, Asthma, and even Covid-19; just by smelling with their nose.

Sounds quite stunning, ain’t it?

And the astonishing thing is, diagnosing diseases by smelling is not a new concept. In 400 BC, ancient medical practitioners recognized that different disease causes different odors of body excretions like urine, sputum, sweat, and vaginal fluid.

“In persons affected with phthisis, if the sputa which they cough up, have a heavy smell when poured upon coals, the case will prove fatal”(Aphorisms, V).

So in middle age doctors used their noses for sensing specific odors to diagnose disease. But in the modern era instead of using noses made of flesh and blood, they are using noses made of electronic components, known as the electronic nose.

The development of e-nose started as far back as the 1950s, and over the last decades, it has undergone major technical and commercial development.

So, what is an electronic nose?

An electronic nose (image by Simon Bootsma)

The e-nose is an electronic version of the human olfactory system. As the name describes, e-nose smells and analyzes odor. The odors are the summation of chemical molecules. An electronic nose consists of some chemical sensor arrays. So more precisely, the e-nose senses chemical molecules associated with odor by this sensor array. Then the pattern-recognition system of e-nose recognizes the complex and simple odors.

Different kinds of odors have different chemical molecule patterns. That’s why an orange scent is non-identical to a mango scent, and a mango scent is non-identical to coffee. The E-nose is capable of capturing these differences between odors. But the question is,

Do odors contain diagnostic clues?

The answer is: Yes.

It is quite unbelievable that the study says, only our breathe odor contains 60% information about diseases. When any organ or cell of our body is affected by any disease, it creates toxins. The toxins are mostly in the form of gas that is exhaled through the breath. That’s why when our body is affected by any disorder, breathe odor changes. But not only breath odor but also other excretion sites like skin, sweat, stool, urine, sputum odor changes in different diseases.

The following table below shows how the odor of our body sites changes in different diseases.

From the table data, we can realize odor is a fingerprint of diseases that changes from disease to disease.

So if we can detect the difference between these fingerprints, the clinical diagnosis will be more precise, easy, and low-cost. E-nose does this work for us. It recognizes the difference and analyzes them. So let’s have a closure look at-

How does an electronic nose work?

An e-nose detects the difference between odors, the same way the human olfactory system detects. When any substance releases its odor or smell, it releases chemical molecules. In our nose, there are several receptors. Each is different sized and shaped. When we breathe in, molecules associated with odors do reactions with these receptors. These reactions generate different signal patterns. Then the brain processes the signals.

Working principle of human nose

And in the case of an electronic-nose, it has eight pairs of sensors working as the receptor. When the sensor arrays of the e-nose came near to any odor, the chemical molecule reacts with the arrays creating different signal patterns, and after pre-processing send them to the computer. The computer then analyzes the signal patterns by pattern recognition method.

Working principle of an electronic nose

In summary, the e-nose has three parts. Those are:

· Sample delivery System

· Detection System

· Computing System

v Sample delivery System: The main component of this system is sample data. An e-nose contains a big fat database. With the help of doctors all over the world, e-nose manufacturing companies collect different odor data patterns from several patients who have different diseases.

Doctors taking breathe odor samples from patients (images from Google search)

Then the data is stored in a cloud called Breathe cloud. So when a company has a good database, an e-nose is simply the summation of some electronic component.

v Detection System: The sensor arrays made up this system. The process is simple; the odor molecules change the chemical condition of sensors and results in a signal pattern. The pattern is then processed in the computing system described below. Commonly MOSFET, conducting polymers, quartz crystal microbalance, piezoelectric sensors, Metal Oxide sensors are used in the electronic nose as a sensor array.

v Computing System:

Flow diagram of computing system

The signal patterns send to the computing system are further analyzed and processed in a computer. The input signal pattern undergoes pattern classification in this part. When the input pattern matches any pre-existing one from the database, it gives a digital output.

The output of an e-nose requires to be analyzed. And this is done by graphical analysis, data analysis, and network analysis.

Data analysis process flow diagram

So after all this process of data analysis, e-nose is done with disease detection.

To conclude, now we can imagine a day when we will be able to cut off expenses in clinical diagnosis. Maybe in the future e-nose will be so available that, a patient will be able to send his odor data from home to the cloud. And get his diagnosis result by 10 or 15 minutes. Its application will be more feasible for doctors in the unprivileged area to ensure the easy and low-cost diagnosis of all kinds of diseases.

--

--