Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography

Hildur Einarsdottir, Andre Yaroshenko, Astrid Velroyen, Martin Bech, Katharina Hellbach, Sigrid Auweter, Önder Yildirim, Felix G. Meinel, Oliver x Oliver Eickelberg, Maximilian Reiser, Rasmus Larsen, Bjarne Kjær Ersbøll, Franz Pfeiffer

Research output: Contribution to journalJournal articleResearchpeer-review

2 Downloads (Pure)

Abstract

In this work we develop a computer-aided diagnosis (CAD) scheme for classification of pulmonary disease for grating-based x-ray radiography. In addition to conventional transmission radiography, the grating-based technique provides a dark-field imaging modality, which utilizes the scattering properties of the x-rays. This modality has shown great potential for diagnosing early stage emphysema and fibrosis in mouse lungs in vivo. The CAD scheme is developed to assist radiologists and other medical experts to develop new diagnostic methods when evaluating grating-based images. The scheme consists of three stages: (i) automatic lung segmentation; (ii) feature extraction from lung shape and dark-field image intensities; (iii) classification between healthy, emphysema and fibrosis lungs. A study of 102 mice was conducted with 34 healthy, 52 emphysema and 16 fibrosis subjects. Each image was manually annotated to build an experimental dataset. System performance was assessed by: (i) determining the quality of the segmentations; (ii) validating emphysema and fibrosis recognition by a linear support vector machine using leave-one-out cross-validation. In terms of segmentation quality, we obtained an overlap percentage (Ω) 92.63  ±  3.65%, Dice Similarity Coefficient (DSC) 89.74  ±  8.84% and Jaccard Similarity Coefficient 82.39  ±  12.62%. For classification, the accuracy, sensitivity and specificity of diseased lung recognition was 100%. Classification between emphysema and fibrosis resulted in an accuracy of 93%, whilst the sensitivity was 94% and specificity 88%. In addition to the automatic classification of lungs, deviation maps created by the CAD scheme provide a visual aid for medical experts to further assess the severity of pulmonary disease in the lung, and highlights regions affected.
Original languageEnglish
JournalPhysics in Medicine and Biology
Volume60
Pages (from-to)9253-9268
ISSN0031-9155
DOIs
Publication statusPublished - 2015

Keywords

  • X-ray radiography
  • Dark-field imaging
  • Lung segmentation
  • Active appearance model
  • Pulmonary disease
  • Grating based interferometry

Cite this

Einarsdottir, H., Yaroshenko, A., Velroyen, A., Bech, M., Hellbach, K., Auweter, S., ... Pfeiffer, F. (2015). Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography. Physics in Medicine and Biology, 60, 9253-9268. https://doi.org/10.1088/0031-9155/60/24/9253
Einarsdottir, Hildur ; Yaroshenko, Andre ; Velroyen, Astrid ; Bech, Martin ; Hellbach, Katharina ; Auweter, Sigrid ; Yildirim, Önder ; Meinel, Felix G. ; Oliver Eickelberg, Oliver x ; Reiser, Maximilian ; Larsen, Rasmus ; Ersbøll, Bjarne Kjær ; Pfeiffer, Franz . / Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography. In: Physics in Medicine and Biology. 2015 ; Vol. 60. pp. 9253-9268.
@article{8226b2f4877047daa91fc555d772da67,
title = "Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography",
abstract = "In this work we develop a computer-aided diagnosis (CAD) scheme for classification of pulmonary disease for grating-based x-ray radiography. In addition to conventional transmission radiography, the grating-based technique provides a dark-field imaging modality, which utilizes the scattering properties of the x-rays. This modality has shown great potential for diagnosing early stage emphysema and fibrosis in mouse lungs in vivo. The CAD scheme is developed to assist radiologists and other medical experts to develop new diagnostic methods when evaluating grating-based images. The scheme consists of three stages: (i) automatic lung segmentation; (ii) feature extraction from lung shape and dark-field image intensities; (iii) classification between healthy, emphysema and fibrosis lungs. A study of 102 mice was conducted with 34 healthy, 52 emphysema and 16 fibrosis subjects. Each image was manually annotated to build an experimental dataset. System performance was assessed by: (i) determining the quality of the segmentations; (ii) validating emphysema and fibrosis recognition by a linear support vector machine using leave-one-out cross-validation. In terms of segmentation quality, we obtained an overlap percentage (Ω) 92.63  ±  3.65{\%}, Dice Similarity Coefficient (DSC) 89.74  ±  8.84{\%} and Jaccard Similarity Coefficient 82.39  ±  12.62{\%}. For classification, the accuracy, sensitivity and specificity of diseased lung recognition was 100{\%}. Classification between emphysema and fibrosis resulted in an accuracy of 93{\%}, whilst the sensitivity was 94{\%} and specificity 88{\%}. In addition to the automatic classification of lungs, deviation maps created by the CAD scheme provide a visual aid for medical experts to further assess the severity of pulmonary disease in the lung, and highlights regions affected.",
keywords = "X-ray radiography, Dark-field imaging, Lung segmentation, Active appearance model, Pulmonary disease, Grating based interferometry",
author = "Hildur Einarsdottir and Andre Yaroshenko and Astrid Velroyen and Martin Bech and Katharina Hellbach and Sigrid Auweter and {\"O}nder Yildirim and Meinel, {Felix G.} and {Oliver Eickelberg}, {Oliver x} and Maximilian Reiser and Rasmus Larsen and Ersb{\o}ll, {Bjarne Kj{\ae}r} and Franz Pfeiffer",
year = "2015",
doi = "10.1088/0031-9155/60/24/9253",
language = "English",
volume = "60",
pages = "9253--9268",
journal = "Physics in Medicine and Biology",
issn = "0031-9155",
publisher = "IOP Publishing",

}

Einarsdottir, H, Yaroshenko, A, Velroyen, A, Bech, M, Hellbach, K, Auweter, S, Yildirim, Ö, Meinel, FG, Oliver Eickelberg, OX, Reiser, M, Larsen, R, Ersbøll, BK & Pfeiffer, F 2015, 'Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography', Physics in Medicine and Biology, vol. 60, pp. 9253-9268. https://doi.org/10.1088/0031-9155/60/24/9253

Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography. / Einarsdottir, Hildur; Yaroshenko, Andre ; Velroyen, Astrid ; Bech, Martin; Hellbach, Katharina ; Auweter, Sigrid ; Yildirim, Önder ; Meinel, Felix G. ; Oliver Eickelberg, Oliver x; Reiser, Maximilian ; Larsen, Rasmus; Ersbøll, Bjarne Kjær; Pfeiffer, Franz .

In: Physics in Medicine and Biology, Vol. 60, 2015, p. 9253-9268.

Research output: Contribution to journalJournal articleResearchpeer-review

TY - JOUR

T1 - Computer-aided diagnosis of pulmonary diseases using x-ray darkfield radiography

AU - Einarsdottir, Hildur

AU - Yaroshenko, Andre

AU - Velroyen, Astrid

AU - Bech, Martin

AU - Hellbach, Katharina

AU - Auweter, Sigrid

AU - Yildirim, Önder

AU - Meinel, Felix G.

AU - Oliver Eickelberg, Oliver x

AU - Reiser, Maximilian

AU - Larsen, Rasmus

AU - Ersbøll, Bjarne Kjær

AU - Pfeiffer, Franz

PY - 2015

Y1 - 2015

N2 - In this work we develop a computer-aided diagnosis (CAD) scheme for classification of pulmonary disease for grating-based x-ray radiography. In addition to conventional transmission radiography, the grating-based technique provides a dark-field imaging modality, which utilizes the scattering properties of the x-rays. This modality has shown great potential for diagnosing early stage emphysema and fibrosis in mouse lungs in vivo. The CAD scheme is developed to assist radiologists and other medical experts to develop new diagnostic methods when evaluating grating-based images. The scheme consists of three stages: (i) automatic lung segmentation; (ii) feature extraction from lung shape and dark-field image intensities; (iii) classification between healthy, emphysema and fibrosis lungs. A study of 102 mice was conducted with 34 healthy, 52 emphysema and 16 fibrosis subjects. Each image was manually annotated to build an experimental dataset. System performance was assessed by: (i) determining the quality of the segmentations; (ii) validating emphysema and fibrosis recognition by a linear support vector machine using leave-one-out cross-validation. In terms of segmentation quality, we obtained an overlap percentage (Ω) 92.63  ±  3.65%, Dice Similarity Coefficient (DSC) 89.74  ±  8.84% and Jaccard Similarity Coefficient 82.39  ±  12.62%. For classification, the accuracy, sensitivity and specificity of diseased lung recognition was 100%. Classification between emphysema and fibrosis resulted in an accuracy of 93%, whilst the sensitivity was 94% and specificity 88%. In addition to the automatic classification of lungs, deviation maps created by the CAD scheme provide a visual aid for medical experts to further assess the severity of pulmonary disease in the lung, and highlights regions affected.

AB - In this work we develop a computer-aided diagnosis (CAD) scheme for classification of pulmonary disease for grating-based x-ray radiography. In addition to conventional transmission radiography, the grating-based technique provides a dark-field imaging modality, which utilizes the scattering properties of the x-rays. This modality has shown great potential for diagnosing early stage emphysema and fibrosis in mouse lungs in vivo. The CAD scheme is developed to assist radiologists and other medical experts to develop new diagnostic methods when evaluating grating-based images. The scheme consists of three stages: (i) automatic lung segmentation; (ii) feature extraction from lung shape and dark-field image intensities; (iii) classification between healthy, emphysema and fibrosis lungs. A study of 102 mice was conducted with 34 healthy, 52 emphysema and 16 fibrosis subjects. Each image was manually annotated to build an experimental dataset. System performance was assessed by: (i) determining the quality of the segmentations; (ii) validating emphysema and fibrosis recognition by a linear support vector machine using leave-one-out cross-validation. In terms of segmentation quality, we obtained an overlap percentage (Ω) 92.63  ±  3.65%, Dice Similarity Coefficient (DSC) 89.74  ±  8.84% and Jaccard Similarity Coefficient 82.39  ±  12.62%. For classification, the accuracy, sensitivity and specificity of diseased lung recognition was 100%. Classification between emphysema and fibrosis resulted in an accuracy of 93%, whilst the sensitivity was 94% and specificity 88%. In addition to the automatic classification of lungs, deviation maps created by the CAD scheme provide a visual aid for medical experts to further assess the severity of pulmonary disease in the lung, and highlights regions affected.

KW - X-ray radiography

KW - Dark-field imaging

KW - Lung segmentation

KW - Active appearance model

KW - Pulmonary disease

KW - Grating based interferometry

U2 - 10.1088/0031-9155/60/24/9253

DO - 10.1088/0031-9155/60/24/9253

M3 - Journal article

C2 - 26577057

VL - 60

SP - 9253

EP - 9268

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

SN - 0031-9155

ER -