EXAMINATION OF POSSIBLE LINKS BETWEEN DRUG RESISTANCE AND MORPHOLOGY OF LUNG IMAGES OF TUBERCULOSIS PATIENTS
Abstract
The purpose of this paper is to present the results of an exploratory study of possible correlations between the drug resistance and the structural features of CT and X-ray images of lungtuberculosis patients. A multi-step procedure is suggested which includes calculation of textural image features, extracting their principal components, correlating them to patients’ clinical data and mapping the significant principal components back to image descriptor elements and then to the corresponding image structures they found to be linked with. The results of a detailed statistical analysis of the revealed links between the drug resistance and the image features are presented. The analysis includes finding
one-factor correlations, performing multivariate regression analysis and cross-validation.
one-factor correlations, performing multivariate regression analysis and cross-validation.
About the Authors
V. A. Kovalev
Объединенный институт проблем информатики НАН Беларуси
Russian Federation
Russian Federation
V. A. Liauchuk
Объединенный институт проблем информатики НАН Беларуси
Russian Federation
Russian Federation
I. U. Safonau
Объединенный институт проблем информатики НАН Беларуси
Russian Federation
Russian Federation
A. U. Tarasau
Республиканский научно-практический центр пульмонологии и фтизиатрии
Russian Federation
Russian Federation
References
1. Ferguson, L.A. Multidrug-resistant and extensively drug-resistant tuberculosis : The new face of an old disease / L.A. Ferguson, J. Rhoads // Journal of American Academy Nurse Practitioners. – 2009. – Vol. 21, № 11. – P. 603–609.
2. Chiang, C.Y. Drug-resistant tuberculosis: Past, present, future / C.Y. Chiang, R. Centis, G.B. Migliori // Respirology. – 2010. – Vol. 15, № 3. – P. 413–432.
3. Multidrug-resistant tuberculosis in Belarus: the size of the problem and associated risk factors / A. Skrahina [et al.] // Bulletin of the World Health Organization. – 2013. – Vol. 91. – P. 36–45.
4. Belarus Tuberculosis Portal [Электронный ресурс]. – Режим доступа :http://tuberculosis.by. – Дата доступа : 20.02.2013.
5. Radiological Findings of Extensively Drug-Resistant Pulmonary Tuberculosis in Non-AIDS Adults : Comparisons with Findings of Multidrug-Resistant and Drug-Sensitive Tuberculosis / J. Cha [et al.] // Korean Journal of Radiology. – 2009. – Vol. 10. – P. 207–216.
6. Computed Tomography Features of Extensively Drug-Resistant Pulmonary Tuberculosis in Non-HIV-Infected Patients / E.S. Lee [et al.] // Journal of Computer Assisted Tomography. – 2010. – Vol. 34. – P. 559–563.
7. Three-dimensional texture analysis of MRI brain datasets / V.A. Kovalev [et al.] // IEEE Transactions on Medical Imaging. – 2001. – Vol. 20, № 5. – P. 424–433.
8. A method for identification and visualization of histological image structures relevant to the cancer patient conditions / V.A. Kovalev [et al.] // Proc. of the 27-th Int. congress on Computer Analysis of Images and Patterns (CAIP-2011). – Spain, 2011. – Vol. 6854, № 1. – P. 460–468.
9. Kovalev, V.A. Gender and age effects in structural brain asymmetry as measured by MRI texture analysis / V.A. Kovalev, F. Kruggel, D.Y. von Cramon // NeuroImage. – 2003. – Vol. 19. – P. 896–905.
10. Kaiser, H.F. The application of electronic computers to factor analysis / H.F. Kaiser // Educational and Psychological Measurement. – 1960. – Vol. 20. – P. 141–151.
11. Sapsford, R. Data Collection and Analysis / R. Sapsford, V. Jupp. – London : Sage, 2006. – 332 p.
12. Kohavi, R. A study of cross-validation and bootstrap for accuracy estimation and model selection / R. Kohavi // Proc. of the Fourteenth Int. Joint Conf. on Artificial Intelligence (IJCAI'95). – USA, 1995. – Vol. 2. – P. 1137–1143.
Review
For citations:
Kovalev V.A., Liauchuk V.A., Safonau I.U., Tarasau A.U. EXAMINATION OF POSSIBLE LINKS BETWEEN DRUG RESISTANCE AND MORPHOLOGY OF LUNG IMAGES OF TUBERCULOSIS PATIENTS. Informatics. 2013;(4):13-22. (In Russ.)
Views: 813
Email this article 