Limits...
Detecting Lung Diseases from Exhaled Aerosols: Non-Invasive Lung Diagnosis Using Fractal Analysis and SVM Classification.

Xi J, Zhao W, Yuan JE, Kim J, Si X, Xu X - PLoS ONE (2015)

Bottom Line: By employing the 10-fold cross-validation method, we achieved 100% classification accuracy among four asthmatic models using an ideal 108-sample dataset and 99.1% accuracy using a more realistic 324-sample dataset.The fractal-SVM classifier has been shown to be robust, highly sensitive to structural variations, and inherently suitable for investigating aerosol-disease correlations.For the first time, this study quantitatively linked the exhaled aerosol patterns with their underlying diseases and set the stage for the development of a computer-aided diagnostic system for non-invasive detection of obstructive respiratory diseases.

View Article: PubMed Central - PubMed

Affiliation: School of Engineering and Technology, Central Michigan University, Mount Pleasant, Michigan, United States of America.

ABSTRACT

Background: Each lung structure exhales a unique pattern of aerosols, which can be used to detect and monitor lung diseases non-invasively. The challenges are accurately interpreting the exhaled aerosol fingerprints and quantitatively correlating them to the lung diseases.

Objective and methods: In this study, we presented a paradigm of an exhaled aerosol test that addresses the above two challenges and is promising to detect the site and severity of lung diseases. This paradigm consists of two steps: image feature extraction using sub-regional fractal analysis and data classification using a support vector machine (SVM). Numerical experiments were conducted to evaluate the feasibility of the breath test in four asthmatic lung models. A high-fidelity image-CFD approach was employed to compute the exhaled aerosol patterns under different disease conditions.

Findings: By employing the 10-fold cross-validation method, we achieved 100% classification accuracy among four asthmatic models using an ideal 108-sample dataset and 99.1% accuracy using a more realistic 324-sample dataset. The fractal-SVM classifier has been shown to be robust, highly sensitive to structural variations, and inherently suitable for investigating aerosol-disease correlations.

Conclusion: For the first time, this study quantitatively linked the exhaled aerosol patterns with their underlying diseases and set the stage for the development of a computer-aided diagnostic system for non-invasive detection of obstructive respiratory diseases.

No MeSH data available.


Related in: MedlinePlus

Distance matrix and dendrogram for the 108-sample-experiment which has a respiration range of 30±3 L/min with no upper airway variation.Detailed dendrograms showing hierarchical clustering of particle sizes have been cut off for visual clarity in (a). The color histogram is shown in (b) and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in (c). The prediction accuracy is 100% in this idealized condition. Q30: inhalation flow rate = 30L/min.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4589383&req=5

pone.0139511.g007: Distance matrix and dendrogram for the 108-sample-experiment which has a respiration range of 30±3 L/min with no upper airway variation.Detailed dendrograms showing hierarchical clustering of particle sizes have been cut off for visual clarity in (a). The color histogram is shown in (b) and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in (c). The prediction accuracy is 100% in this idealized condition. Q30: inhalation flow rate = 30L/min.

Mentions: The distance matrix for the 108-sample dataset is shown in Fig 7A, with the main feature dendrogram tree shown on the right side. Detailed dendrograms showing hierarchical clustering for particle sizes have been cut off for visual clarity. The color histogram is shown in Fig 7B and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in Fig 7C. From Fig 7A, the 108 samples were divided into 12 subgroups; each subgroup shares a given combination of D- and Q-features, as displayed on the left side of the heat-map. As a result, the selected 32-dimenisonal feature vectors were adequate in capturing the differences between the D- and Q-feature samples. On the other hand, it was also observed in Fig 7A that the distributions of D- or Q-feature subgroups were mixed in the heat-map. This mixed distribution indicated that in the original space (32-dimensional space), samples with the same D- or Q-feature could not be readily distinguished. Rather, a new transformed space was needed to classify these specimens. Our classification analysis results showed that by mapping samples into high dimensional space, the SVM algorithm could accurately distinguish the samples with different D- or Q-features. The feature classification accuracies were 100% (10-fold cross-validation) in this 108-sample dataset (Table 2).


Detecting Lung Diseases from Exhaled Aerosols: Non-Invasive Lung Diagnosis Using Fractal Analysis and SVM Classification.

Xi J, Zhao W, Yuan JE, Kim J, Si X, Xu X - PLoS ONE (2015)

Distance matrix and dendrogram for the 108-sample-experiment which has a respiration range of 30±3 L/min with no upper airway variation.Detailed dendrograms showing hierarchical clustering of particle sizes have been cut off for visual clarity in (a). The color histogram is shown in (b) and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in (c). The prediction accuracy is 100% in this idealized condition. Q30: inhalation flow rate = 30L/min.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4589383&req=5

pone.0139511.g007: Distance matrix and dendrogram for the 108-sample-experiment which has a respiration range of 30±3 L/min with no upper airway variation.Detailed dendrograms showing hierarchical clustering of particle sizes have been cut off for visual clarity in (a). The color histogram is shown in (b) and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in (c). The prediction accuracy is 100% in this idealized condition. Q30: inhalation flow rate = 30L/min.
Mentions: The distance matrix for the 108-sample dataset is shown in Fig 7A, with the main feature dendrogram tree shown on the right side. Detailed dendrograms showing hierarchical clustering for particle sizes have been cut off for visual clarity. The color histogram is shown in Fig 7B and an example of the particle dendrogram at the location ([12, 12]: Q30, D2) is shown in Fig 7C. From Fig 7A, the 108 samples were divided into 12 subgroups; each subgroup shares a given combination of D- and Q-features, as displayed on the left side of the heat-map. As a result, the selected 32-dimenisonal feature vectors were adequate in capturing the differences between the D- and Q-feature samples. On the other hand, it was also observed in Fig 7A that the distributions of D- or Q-feature subgroups were mixed in the heat-map. This mixed distribution indicated that in the original space (32-dimensional space), samples with the same D- or Q-feature could not be readily distinguished. Rather, a new transformed space was needed to classify these specimens. Our classification analysis results showed that by mapping samples into high dimensional space, the SVM algorithm could accurately distinguish the samples with different D- or Q-features. The feature classification accuracies were 100% (10-fold cross-validation) in this 108-sample dataset (Table 2).

Bottom Line: By employing the 10-fold cross-validation method, we achieved 100% classification accuracy among four asthmatic models using an ideal 108-sample dataset and 99.1% accuracy using a more realistic 324-sample dataset.The fractal-SVM classifier has been shown to be robust, highly sensitive to structural variations, and inherently suitable for investigating aerosol-disease correlations.For the first time, this study quantitatively linked the exhaled aerosol patterns with their underlying diseases and set the stage for the development of a computer-aided diagnostic system for non-invasive detection of obstructive respiratory diseases.

View Article: PubMed Central - PubMed

Affiliation: School of Engineering and Technology, Central Michigan University, Mount Pleasant, Michigan, United States of America.

ABSTRACT

Background: Each lung structure exhales a unique pattern of aerosols, which can be used to detect and monitor lung diseases non-invasively. The challenges are accurately interpreting the exhaled aerosol fingerprints and quantitatively correlating them to the lung diseases.

Objective and methods: In this study, we presented a paradigm of an exhaled aerosol test that addresses the above two challenges and is promising to detect the site and severity of lung diseases. This paradigm consists of two steps: image feature extraction using sub-regional fractal analysis and data classification using a support vector machine (SVM). Numerical experiments were conducted to evaluate the feasibility of the breath test in four asthmatic lung models. A high-fidelity image-CFD approach was employed to compute the exhaled aerosol patterns under different disease conditions.

Findings: By employing the 10-fold cross-validation method, we achieved 100% classification accuracy among four asthmatic models using an ideal 108-sample dataset and 99.1% accuracy using a more realistic 324-sample dataset. The fractal-SVM classifier has been shown to be robust, highly sensitive to structural variations, and inherently suitable for investigating aerosol-disease correlations.

Conclusion: For the first time, this study quantitatively linked the exhaled aerosol patterns with their underlying diseases and set the stage for the development of a computer-aided diagnostic system for non-invasive detection of obstructive respiratory diseases.

No MeSH data available.


Related in: MedlinePlus