2. Introduction
• 9.2 million new cases of TB were reported in 2006
resulting into 1.7 million deaths (WHO report)
• Active TB infection occurs when immune system is
compromised
• Existing techniques are time taking and have low
sensitivity
• VOC composition of normal breath is affected
• Electronic gas sensor array is used for detection
3. Proposed Idea
• Exhaled breath is a type of bodily excretion
• To detect the specific pattern of the VOCs in the
exhaled breath using Electronic nose
• Electronic nose – an array of electronic gas sensors
• Signal is analyzed using PCA
• Pattern is generated by reducing the dimensions of
obtained data
• Comparison with the standard
4. VOC composition of normal and
infected breath
A list of 10 most abundant VOCs of M. tuberculosis is given bellow*
• Naphthalene
• 1-methyl-3-Heptanone
• Methylcyclododecane
• Heptanes
• 2, 2, 4, 6, 6-pentamethyl- Benzene,
• 1-methyl-4-(1-methylethyl)-Cyclohexane
• 1,4-dimethyl-3,5-dimethylamphetamine
• Butanal
• 3-methyl- 2-Hexene
• Trans-anti-1-methyl-decahydronaphthalene
5. Electronic Sensors :
3 type of sensors can be used:
• Thick Film:
Paste is screen printed on alumina substrate
Low sensitivity and specificity
• Thin Film:
Technique used – Chemical Vapor Deposition
Uniform layer having higher surface area
• MOS – FET:
Has high sensitivity, specificity
depends on the gate material
6. Signal analysis (PCA)
• To filter noise and reduce the dimensions
• Co-variance matrix is calculated from the obtained
data
• Eigen vectors are calculated for this matrix
• Feature vector is obtained by omitting the less
significant Eigen vectors
• Using this feature vector back calculation is done to
obtain new set of data
• New data varies only along the selected Eigen vectors
7.
8.
9.
10. Pattern Analysis:
• Pattern obtained is unique to that composition
• New samples are compared with a pattern
library using pattern recognition algorithms
• Self learning is needed to build the pattern
library
• Result is obtained in terms of percentage
similarity
• Threshold can be set accordingly
11. Conclusion:
• The test is non invasive hence no contamination
issues
• Rapid and low cost
• High enough sensitivity, can detect TB in early
stages
• Technique can be developed further to detect lung
cancer, asthma, diabetes etc