Neural network-based early electronic nose lung cancer diagnosis system (1)

Lung cancer is the most common cancer worldwide, one of the rising incidence and mortality rates.

The reason is lung cancer etiology is unknown, the attack time is short, transfer speed, high malignancy, previous difficult diagnosis, to the middle and late lost opportunities for patients undergoing a five-year survival rate of only 15 percent. Early detection can make patients within five years of survival rate to 70 per cent to 80 per cent. Therefore, early detection, early diagnosis and early treatment is to improve survival and reduce mortality from lung cancer. Lung cancer symptoms are usually non-special, almost not be alert to doctors and patients, and common diagnostic method for difficult early detection and early diagnosis.

At present, based on the electronic nose diseases diagnosis is mainly concentrated in the kidney disease, diabetes early diagnosis and some bacteria types and growth phase of recognition.

Electronic nose diseases diagnosis as medical diagnosis-destructive one of the important direction, have already made many achievements, but yet have certified respiratory diagnostic instruments. How to further improve lung cancer diagnosis technology, improve the variety of treatments, etc., has become a recent areas of cancer research worldwide. China also lung cancer as national key research project. Looking for a more advanced equipment and technology in local organizations which can occur in cancer detection and diagnosis, that is, the e-nose system the work to be done. This article from related pathology of respiratory gas detection and collection, gas sensor array of selection and optimization and pattern recognition technology choices, and so on electronic nose lung cancer early diagnosis system for the design of key technologies, and achieved good results.

1 electronic nose lung cancer early diagnosis system structure

Electronic nose is the use of gas sensor array response patterns to identify gas electronic systems.

Electronic nose is mainly composed of gas sampling operations, gas sensor array and signal processing system consists of three functional devices. Electronic nose identify gas main mechanism is the array of each sensor test gases have different sensitivity, the entire sensor array to the response of different gas logo. It is this difference that makes the system according to the sensor response patterns to identify odours.

Electronic nose typical workflow is: first, use breathing gas collection equipment (after breathing gas purification and flow control) to draw to the breathing gas is fitted with an electronic sensor array in the small container room; then, put the sensor array has been initialized to be logging gas exposure, when the volatile compounds (VOC) and sensor activity material surface contact, generate instantaneous response, the response is recorded and delivered to the signal processing unit for analysis, and database to store large amounts of VOC pattern identification, to compare, to determine the type of gas; and, finally, the clean gas flushing sensor activity material surface to remove logging bi gas mixtures.

In the next round of new measurement, the sensor is still required before once again initialization (that is, before work, each sensor is necessary for drying gas or some other reference gas cleaning to reach baseline condition). Electronic nose lung cancer early diagnosis system structure as shown in Figure 1.

2 electronic nose lung cancer early diagnosis system design

This article from lung cancer patients with diseases of the respiratory gas and greater relevance of pathological significance of gas detection, respiratory gas collection equipment, gas sensor array of selection and optimization and pattern recognition technology choices in the design of key technologies.

2.1 breathing gas detection

About 200 kinds of compounds in human breath in reconnaissance, some gas and lung cancer related, so use breath to detect diseases is feasible.

Its advantages are non-intrusive, simple, fast, and it has a high quality of application development potential. And hospitals, and select the right gas sensors and detection methods to detect lung cancer patients breath gases styrene, n-Decane, 11, 22 species of organic volatile (VOCs) characteristics of the gas concentration is very promising for lung cancer in non-invasive diagnosis and monitoring methods.

2.2 breathing gas collection device design

As the breath in and lung cancer related gas concentrations are very low (usually in ppb level), and the traditional way of breath testing is based on the standard after the breath sample programs, and then to gas chromatography-mass Analyzer to determine the types of compounds their analysis process must be a lot of breath sample (approximately 3L breath samples) concentration before reaching instrument can detect limits.

This method does not only equipment is expensive, time-consuming and require a large number of physical samples. The electronic nose the desired analysis costing is not high, and the required amount of breath samples, only 10ml simple operation, fast response (few minutes). Respiratory gases collect in the electronic nose of lung cancer early diagnosis system occupies a very important position in the structure of the gas-collection device is shown in Figure 2.