Home Computer-aided Diagnosis of Pulmonary Infections Using Texture Analysis and Support Vector Machine Classification
Post
Cancel

Computer-aided Diagnosis of Pulmonary Infections Using Texture Analysis and Support Vector Machine Classification

Posted Feb 28, 2011 Updated Apr 10, 2023
By Jianhua Yao PhD , Andrew Dwyer MD , Ronald M. Summers MD PhD , Daniel J. Mollura MD
9 min read

Rationale and Objectives

The purpose of this study was to develop and test a computer-assisted detection method for the identification and measurement of pulmonary abnormalities on chest computed tomographic (CT) imaging in cases of infection, such as novel H1N1 influenza. The method developed could be a potentially useful tool for classifying and quantifying pulmonary infectious disease on CT imaging.

Materials and Methods

Forty chest CT examinations were studied using texture analysis and support vector machine classification to differentiate normal from abnormal lung regions on CT imaging, including 10 patients with immunohistochemistry-proven infection, 10 normal controls, and 20 patients with fibrosis.

Results

Statistically significant differences in the receiver-operating characteristic curves for detecting abnormal regions in H1N1 infection were obtained between normal lung and regions of fibrosis, with significant differences in texture features of different infections. These differences enabled the quantification of abnormal lung volumes on CT imaging.

Conclusion

Texture analysis and support vector machine classification can distinguish between areas of abnormality in acute infection and areas of chronic fibrosis, differentiate lesions having consolidative and ground-glass appearances, and quantify those texture features to increase the precision of CT scoring as a potential tool for measuring disease progression and severity.

In trying to increase the clinical utility of infectious disease imaging, researchers currently face several challenges, including the relatively low specificity for diagnosing pathogens and the limited quantification of disease burden for assessing severity and predicting outcomes. The low specificity of infectious disease imaging stems from the similarity between visual appearances of infectious and inflammatory diseases . The second major challenge, the quantification of severity through radiologic techniques, requires standardized methods for measuring lesions and translating those measurements into validated clinical implications. A third unsolved problem is that the detection of subtle pulmonary parenchymal changes may not be visually apparent, and traditional visual scoring methods for pulmonary disease on computed tomographic (CT) imaging are often limited by interobserver biases and lack of validation. These limitations came to light during the outbreak of novel swine-origin influenza A/H1N1 in 2009 and 2010 . Reports indicated that severe infection with novel H1N1 demonstrate patchy ground-glass opacities with consolidations on thoracic CT scans . Unfortunately, this visual appearance is so similar to other infectious and inflammatory etiologies that it is difficult to unequivocally diagnose H1N1 on the basis of CT findings alone . Although most cases of H1N1 were predominately mild in severity, with a mortality of <1%, the severe cases often rapidly led to respiratory impairment and death, and it was clinically and radiologically challenging to prognosticate and identify these severe cases for earlier treatment .

In this report, we present a pilot method for detecting and quantifying H1N1 pulmonary infection using computer-assisted texture analysis and support vector machines (SVMs). To our knowledge, H1N1 pulmonary infection and associated inflammation have not been characterized using texture analysis to date. Simply defined, texture analysis quantifies an image by identifying statistical relationships among the pixels’ densities, which can be used to identify lesions and quantify the volume of an organ manifesting those patterns associated with lesions. It has been established that specific tissues and even specific pathologies yield unique texture patterns on chest CT images. Therefore, these textures could be important attributes for characterizing and distinguishing objects, lesions, and regions .

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Materials and methods

Data Sets

Get Radiology Tree app to read full this article<

Table 1

Summary of Patient Population

Cohort Number of Patients (Male/Female) Age (y), Range (Mean ± SD) H1N1 4 (3/1) 31–59 (50 ± 13) Fibrosis 20 (11/9) 30–79 (55 ± 12) Normal 10 (4/6) 38–75 (53 ± 12) Infection other than H1N1 Pneumonia 3 (3/0) 17–44 (33 ± 14) Parainfluenza 2 (1/1) 59–60 MAC 1 (1/0) 82 All 40 (17/23) 17–82 (53 ± 14)

MAC, mycobacterium avium complex; SD, standard deviation.

Get Radiology Tree app to read full this article<

Method Overview

Get Radiology Tree app to read full this article<

Figure 1, Method diagram for image analysis.

Get Radiology Tree app to read full this article<

Lung Segmentation

Get Radiology Tree app to read full this article<

Texture Analysis

Get Radiology Tree app to read full this article<

Figure 2, Texture examples. Top row: normal lung; middle row: H1N1-associated lung opacities; bottom row: lung fibrotic tissues. Each block contains 16 × 16 pixels.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Table 2

List of Texture Features

Histogram Statistics Co-occurrence Matrix Run-length Matrix Mean Energy Long run emphasis Skewness Inertia difference Run length non-uniformity Deviation Correlation Low gray-level run emphasis Variance Average difference Short run low gray-level emphasis Kurtosis Entropy difference Long run low gray-level emphasis Inertia Short run high gray-level emphasis Entropy Long run high gray-level emphasis Average sum Short run emphasis Run gray-level non-uniformity Run percentage

Get Radiology Tree app to read full this article<

Training and SVMs

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Pixelwise Classification

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Statistical Analysis

Get Radiology Tree app to read full this article<

Results

Get Radiology Tree app to read full this article<

Figure 3, Comparison of six texture features for seven classes of patterns. Descriptive statistics (mean and standard deviation of texture values) are plotted for six texture features (density mean, density deviation, correlation, average sum, gray-level nonuniformity, and high gray-level run emphasis). Seven different texture patterns (H1N1, fibrosis, mycobacterium avium complex [MAC], parainfluenza, normal, and normal in H1N1) are compared. The features show statistically significant difference ( P < .001) between abnormal H1N1 region and fibrosis and between H1N1 region and normal lung regions.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 4, Comparison of texture patterns in pneumonia and H1N1 cases. The H1N1 cases with pneumonia have texture values intermediate between those of pneumonia and H1N1 cases without pneumonia.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 5, Pixelwise classification. The left column (a,c,e) from one patient and the right (b,d,f) from the other patient with pathology and reverse transcriptase polymerase chain reaction confirmed swine-origin influenza A/H1N1 infection. (a.b) Raw grayscale computed tomographic images. (c,d) Binary depiction of pixel texture support vector machine (SVM) values based on a cutoff value of 0.5 in which colorized regions are detected as abnormal by the software. (e,f) Graded maps of the pixel texture SVM values. Graded classification presents subtle areas of abnormality as green and yellow, corresponding to pathology-proven regions of bronchitis.

Table 3

Pixelwise Classification of Four Patients with H1N1

Patient Right Lung Volume (cm 3 ) Right Lung Abnormal Volume (cm 3 ) Left Lung Volume (cm 3 ) Left Lung Abnormal Volume (cm 3 ) Total Lung Volume (cm 3 ) Total Lung Abnormal Volume (cm 3 ) Right Lung Ratio Left Lung Ratio Total Lung Ratio 1 1617 1210 1662 262 3278 1472 0.75 0.16 0.45 2 1125 739 1211 195 2336 934 0.66 0.16 0.40 3 1809 623 1814 638 3624 1261 0.34 0.35 0.35 4 1659 405 1478 625 3138 1030 0.24 0.42 0.33

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 6, Receiver-operating characteristic analysis of H1N1 versus normal lungs.

Figure 7, Receiver-operating characteristic analysis of H1N1 versus fibrosis regions.

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Figure 8, Multipattern classification. (a) Computed tomographic (CT) scan of a patient on August 11, 2009; (b) CT scan of the same patient on October 6, 2009; (c) classification results superimposed on (a) ; (d) classification result superimposed on (b) . Blue indicates normal lung or lung not having consolidation or fibrosis, green indicates fibrosis, and red indicates consolidation due to pneumonia (sputum culture–proven aspergillums). Comparison shows that the computer-aided diagnosis system differentiated fibrosis from new consolidation that developed on subsequent scan (b) in comparison to baseline scan (a) .

Get Radiology Tree app to read full this article<

Discussion

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Future Work

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

Conclusions

Get Radiology Tree app to read full this article<

Acknowledgments

Get Radiology Tree app to read full this article<

Get Radiology Tree app to read full this article<

References

  • 1. Kim E.A., Lee K.S., Primack S.L., et. al.: Viral pneumonias in adults: radiologic and pathologic findings. Radiographics 2002; 22: pp. S137-S149.

  • 2. Mollura D., Asnis E., Conetta R.: Imaging findings in a fatal case of pandemic swine-origin influenza A (H1N1). AJR Am J Roentgenol 2009; 193: pp. 1500-1503.

  • 3. Agarwal P., Cinti S., Kazerooni E.: Chest radiographic and CT findings in novel swine-origin influenza A (H1N1) virus (S-OIV) infection. AJR Am J Roentgenol 2009; 193: pp. 1488-1493.

  • 4. Ajlan A.M., Quiney B., Nicolaou S., et. al.: Swine-origin influenza A (H1N1) viral infection: radiographic and CT findings. AJR Am J Roentgenol 2009; 193: pp. 1494-1499.

  • 5. Chabat F., Yang G.-Z., Hansell D.: Obstructive lung diseases: texture classification for differentiation at CT. Radiology 2003; 228: pp. 871-877.

  • 6. Barequet G., Shapiro D., Tal A.: Multilevel sensitive reconstruction of polyhedral surfaces from parallel slices. Vis Comput 2000; 16: pp. 116-133.

  • 7. Yao J, Avila N, Dwyer A, et al. Computer-aided grading of lymphangioleiomyomatosis (LAM) using HRCT. Presented at: International Conference of Pattern Recognition; Tampa, FL; 2008.

  • 8. Sluimer I., Schilham A., Prokop M., et. al.: Computer analysis of computed tomography scans of the lung: a survey. IEEE Trans Med Imaging 2006; 25: pp. 385-405.

  • 9. van Rikxoort E.M., de Hoop B., van de Vorst S., et. al.: Automatic segmentation of pulmonary segments from volumetric chest CT scans. IEEE Trans Med Imaging 2009; 28: pp. 621-630.

  • 10. Uppaluri R., Hoffman E.A., Sonka M., et. al.: Computer recognition of regional lung disease patterns. Am J Respir Crit Care Med 1999; 160: pp. 648-654.

  • 11. Zavaletta VA, Bartholmai BJ, Robb RA. 3D morphological analysis of lung pathology. Presented at: IEEE International Symposium on Biomedical Imaging; Arlington, VA; 2007.

  • 12. Caban J, Yao J, Avila NA, et al. Texture-based computer-aided diagnosis system for lung fibrosis. Presented at: SPIE Medical Imaging (CAD); San Diego, CA; 2007.

  • 13. Lee Y, Seo JB, Bokyoung K, et al. Performance comparison of classifiers for differentiation among obstructive lung diseases based on features of texture analysis at HRCT. In: SPIE Medical Imaging: Image Processing. San Diego, CA, 2007. Available at: http://spie.org/x648.html?product_id=710394 .

  • 14. Cristianini N., Taylor J.S.: An introduction to support vector machines.2000.Cambridge University PressCambridge, UK

  • 15. Yao J., Summers R.M., Hara A.K.: Optimizing the support vector machines (SVM) committee configuration in colonic polyp CAD system. Presented at2005.SPIE Medical ImagingSan Diego, CA

  • 16. Metz C., Herman B., Roe C.: Statistical comparison of two ROC curve estimates obtained from partially-paired datasets. Med Decis Making 1998; 18: pp. 110-121.

  • 17. Morens D.M., Taubenberger J.K., Fauci A.S.: Predominant role of bacterial pneumonia as a cause of death in pandemic influenza: implications for pandemic influenza preparedness. J Infect Dis 2008; 198: pp. 962-970.

  • 18. Mueller-Mang C., Grosse C., Schmid K., et. al.: What every radiologist should know about idiopathic interstitial pneumonias. Radiographics 2007; 27: pp. 595-615.

  • 19. Sumikawa H., Johkoh T., Ichikado K., et. al.: Usual interstitial pneumonia and chronic idiopathic interstitial pneumonia: analysis of CT appearance in 92 patients. Radiology 2006; 241: pp. 258-266.

  • 20. van den Brand J., Stittelaar K.J., van Amerongen G., et. al.: Severity of pneumonia due to new H1N1 influenza virus in ferrets is intermediate between that due to seasonal H1N1 virus and highly pathogenic avian influenza H5N1 virus. J Infect Dis 2010; 201: pp. 993-999.

  • 21. Martinez S., McAdams H.P., Batchu C.S.: The many faces of pulmonary nontuberculous mycobacterial infection. AJR Am J Roentgenol 2007; 189: pp. 177-186.

Academic Radiology, Volume 18, Issue 3
Journals General Radiology
This post is licensed under CC BY 4.0 by the author.
Recently Updated
Contents