1
|
Zhong Z, Yang W, Zhu C, Wang Z. Role and progress of artificial intelligence in radiodiagnosing vascular calcification: a narrative review. ANNALS OF TRANSLATIONAL MEDICINE 2023; 11:131. [PMID: 36819510 PMCID: PMC9929846 DOI: 10.21037/atm-22-6333] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 11/17/2022] [Accepted: 01/10/2023] [Indexed: 01/15/2023]
Abstract
Background and Objective Vascular calcification has important clinical significance due to its vital prognostic value for cardiovascular diseases, chronic kidney disease (CKD), diabetes, fracture, and other multisystem diseases. Radiology is the main diagnostic method of it, but facing great pressure such as the increasing workload and decreasing working accuracy rate. Therefore, radiology needs to find a way out to better realize the clinical value of vascular calcification. Artificial intelligence (AI) encompasses any algorithm imitating human intelligence. AI has shown great potential in image analysis, such as its high speed and accuracy, becoming the savior of the current situation. In order to promote more rational utilization, the role and progress of AI in this field were reviewed. Methods A search was conducted in PubMed and Web of Science. The key words included "artificial intelligence", "machine learning", "deep learning", and "vascular calcification". The qualitative analysis of literature was achieved through repeated deliberation after refining valuable content. The theme is the role and progress of AI in the diagnostic radiology of vascular calcification. Key Content and Findings Sixty-two articles were included. AI has been applied to the diagnostic radiology of 5 types of vascular calcification, including coronary artery calcification (CAC), thoracic aortic calcification (TAC), abdominal aortic calcification (AAC), carotid artery calcification, and breast artery calcification (BAC). Deep learning (DL), the latest technology in this field has been well applied and satisfactorily performed. Radiologists have been able to achieve efficient diagnosis of 5 types of vascular calcification through AI, with reliable accuracy. Conclusions Increasingly, advanced AI has achieved an accuracy comparable to that of human experts, with a faster speed. Moreover, the ability to reduce noise and artifacts enables more imaging equipment to obtain reliable quantification. AI has acquired the ability to cooperate with radiology departments in future work. However, the research in AAC and carotid artery calcification can be more in-depth, and more types of vascular calcification and more fields of radiology should be expanded to. The interpretation of results made by AI and the promotion of existing achievements to the development of other disciplines are also the focus in future.
Collapse
Affiliation(s)
- Zhiqi Zhong
- Department of Cardiology, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Wenjun Yang
- Department of Cardiology, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Chengcheng Zhu
- Digestive Endoscopy Center, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| | - Zhongqun Wang
- Department of Cardiology, Affiliated Hospital of Jiangsu University, Zhenjiang, China
| |
Collapse
|
2
|
Bortsova G, Bos D, Dubost F, Vernooij MW, Ikram MK, van Tulder G, de Bruijne M. Automated Segmentation and Volume Measurement of Intracranial Internal Carotid Artery Calcification at Noncontrast CT. Radiol Artif Intell 2021; 3:e200226. [PMID: 34617024 DOI: 10.1148/ryai.2021200226] [Citation(s) in RCA: 6] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/18/2020] [Revised: 05/31/2021] [Accepted: 06/07/2020] [Indexed: 01/22/2023]
Abstract
Purpose To develop and evaluate a fully-automated deep learning-based method for assessment of intracranial internal carotid artery calcification (ICAC). Materials and Methods This was a secondary analysis of prospectively collected data from the Rotterdam study (2003-2006) to develop and validate a deep learning-based method for automated ICAC delineation and volume measurement. Two observers manually delineated ICAC on noncontrast CT scans of 2319 participants (mean age, 69 years ± 7 [standard deviation]; 1154 women [53.2%]), and a deep learning model was trained to segment ICAC and quantify its volume. Model performance was assessed by comparing manual and automated segmentations and volume measurements to those produced by an independent observer (available on 47 scans), comparing the segmentation accuracy in a blinded qualitative visual comparison by an expert observer, and comparing the association with first stroke incidence from the scan date until 2016. All method performance metrics were computed using 10-fold cross-validation. Results The automated delineation of ICAC reached a sensitivity of 83.8% and positive predictive value (PPV) of 88%. The intraclass correlation between automatic and manual ICAC volume measures was 0.98 (95% CI: 0.97, 0.98; computed in the entire dataset). Measured between the assessments of independent observers, sensitivity was 73.9%, PPV was 89.5%, and intraclass correlation coefficient was 0.91 (95% CI: 0.84, 0.95; computed in the 47-scan subset). In the blinded visual comparisons of 294 regions, automated delineations were judged as more accurate than manual delineations in 131 regions, less accurate in 94 regions, and equally accurate in the rest of the regions (131 of 225, 58.2%; P = .01). The association of ICAC volume with incident stroke was similarly strong for both automated (hazard ratio, 1.38 [95% CI: 1.12, 1.75]) and manually measured volumes (hazard ratio, 1.48 [95% CI: 1.20, 1.87]). Conclusion The developed model was capable of automated segmentation and volume quantification of ICAC with accuracy comparable to human experts.Keywords CT, Neural Networks, Carotid Arteries, Calcifications/Calculi, Arteriosclerosis, Segmentation, Vision Application Domain, Stroke Supplemental material is available for this article. © RSNA, 2021.
Collapse
Affiliation(s)
- Gerda Bortsova
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - Daniel Bos
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - Florian Dubost
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - Meike W Vernooij
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - M Kamran Ikram
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - Gijs van Tulder
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| | - Marleen de Bruijne
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine (G.B., M.d.B.), Department of Epidemiology (D.B., M.W.V., M.K.I.), and Department of Radiology and Nuclear Medicine (M.W.V.), Erasmus MC, PO Box 2040, 3000 CA Rotterdam, the Netherlands; Department of Biomedical Data Science, Stanford University, Stanford, Calif (F.D.); Faculty of Science, Radboud University, Nijmegen, the Netherlands (G.v.T.); and Machine Learning Section, Department of Computer Science, University of Copenhagen, Copenhagen, Denmark (M.d.B.)
| |
Collapse
|
3
|
Aizaz M, Moonen RPM, van der Pol JAJ, Prieto C, Botnar RM, Kooi ME. PET/MRI of atherosclerosis. Cardiovasc Diagn Ther 2020; 10:1120-1139. [PMID: 32968664 DOI: 10.21037/cdt.2020.02.09] [Citation(s) in RCA: 11] [Impact Index Per Article: 2.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/17/2022]
Abstract
Myocardial infarction and stroke are the most prevalent global causes of death. Each year 15 million people worldwide die due to myocardial infarction or stroke. Rupture of a vulnerable atherosclerotic plaque is the main underlying cause of stroke and myocardial infarction. Key features of a vulnerable plaque are inflammation, a large lipid-rich necrotic core (LRNC) with a thin or ruptured overlying fibrous cap, and intraplaque hemorrhage (IPH). Noninvasive imaging of these features could have a role in risk stratification of myocardial infarction and stroke and can potentially be utilized for treatment guidance and monitoring. The recent development of hybrid PET/MRI combining the superior soft tissue contrast of MRI with the opportunity to visualize specific plaque features using various radioactive tracers, paves the way for comprehensive plaque imaging. In this review, the use of hybrid PET/MRI for atherosclerotic plaque imaging in carotid and coronary arteries is discussed. The pros and cons of different hybrid PET/MRI systems are reviewed. The challenges in the development of PET/MRI and potential solutions are described. An overview of PET and MRI acquisition techniques for imaging of atherosclerosis including motion correction is provided, followed by a summary of vessel wall imaging PET/MRI studies in patients with carotid and coronary artery disease. Finally, the future of imaging of atherosclerosis with PET/MRI is discussed.
Collapse
Affiliation(s)
- Mueez Aizaz
- Department of Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, The Netherlands.,CARIM School for Cardiovascular Diseases, Maastricht University, Maastricht, The Netherlands
| | - Rik P M Moonen
- Department of Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, The Netherlands.,CARIM School for Cardiovascular Diseases, Maastricht University, Maastricht, The Netherlands
| | - Jochem A J van der Pol
- Department of Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, The Netherlands
| | - Claudia Prieto
- School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK.,Escuela de Ingenieria, Pontificia Universidad Catolica de Chile, Santiago, Chile
| | - René M Botnar
- School of Biomedical Engineering and Imaging Sciences, King's College London, London, UK.,Escuela de Ingenieria, Pontificia Universidad Catolica de Chile, Santiago, Chile
| | - M Eline Kooi
- Department of Radiology and Nuclear Medicine, Maastricht University Medical Center, Maastricht, The Netherlands.,CARIM School for Cardiovascular Diseases, Maastricht University, Maastricht, The Netherlands
| |
Collapse
|
4
|
Jodas DS, da Costa MFM, Parreira TAA, Pereira AS, Tavares JMRS. Using a distance map and an active contour model to segment the carotid artery boundary from the lumen contour in proton density weighted magnetic resonance images. Comput Biol Med 2020; 123:103901. [PMID: 32658794 DOI: 10.1016/j.compbiomed.2020.103901] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/27/2020] [Revised: 06/20/2020] [Accepted: 06/28/2020] [Indexed: 10/23/2022]
Abstract
Segmentation methods have assumed an important role in image-based diagnosis of several cardiovascular diseases. Particularly, the segmentation of the boundary of the carotid artery is demanded in the detection and characterization of atherosclerosis and assessment of the disease progression. In this article, a fully automatic approach for the segmentation of the carotid artery boundary in Proton Density Weighted Magnetic Resonance Images is presented. The approach relies on the expansion of the lumen contour based on a distance map built using the gray-weighted distance relative to the center of the identified lumen region in the image under analysis. Then, a Snake model with a modified weighted external energy based on the combination of a balloon force along with a Gradient Vector Flow-based external energy is applied to the expanded contour towards the correct boundary of the carotid artery. The average values of the Dice coefficient, Polyline distance, mean contour distance and centroid distance found in the segmentation of 139 carotid arteries were 0.83 ± 0.11, 2.70 ± 1.69 pixels, 2.79 ± 1.89 pixels and 3.44 ± 2.82 pixels, respectively. The segmentation results of the proposed approach were also compared against the ones obtained by related approaches found in the literature, which confirmed the outstanding performance of the new approach. Additionally, the proposed weighted external energy for the Snake model was shown to be also robust to carotid arteries with large thickness and weak boundary image edges.
Collapse
Affiliation(s)
- Danilo Samuel Jodas
- CAPES Foundation, Ministry of Education of Brazil, Brasília - DF, 70040-020, Brazil; Instituto de Ciência e Inovação em Engenharia Mecânica e Engenharia Industrial, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, s/n, 4200-465, Porto, Portugal.
| | - Maria Francisca Monteiro da Costa
- IFE Neurorradiologia, Serviço de Neurorradiologia, Centro Hospitalar São João, Alameda Professor Hernâni Monteiro, 4200-319 Porto, Portugal.
| | - Tiago A A Parreira
- AH Neurorradiologia, Serviço de Neurorradiologia, Centro Hospitalar São João, Alameda Professor Hernâni Monteiro, 4200-319 Porto, Portugal.
| | - Aledir Silveira Pereira
- Universidade Estadual Paulista Júlio de Mesquita Filho, Rua Cristóvão Colombo, 2265, 15054-000, S. J. do Rio Preto, Brazil.
| | - João Manuel R S Tavares
- Instituto de Ciência e Inovação em Engenharia Mecânica e Engenharia Industrial, Faculdade de Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, s/n, 4200-465, Porto, Portugal.
| |
Collapse
|
5
|
Moerman AM, Dilba K, Korteland S, Poot DHJ, Klein S, van der Lugt A, Rouwet EV, van Gaalen K, Wentzel JJ, van der Steen AFW, Gijsen FJH, Van der Heiden K. An MRI-based method to register patient-specific wall shear stress data to histology. PLoS One 2019; 14:e0217271. [PMID: 31170183 PMCID: PMC6553699 DOI: 10.1371/journal.pone.0217271] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/09/2019] [Accepted: 04/27/2019] [Indexed: 11/18/2022] Open
Abstract
Wall shear stress (WSS), the frictional force exerted on endothelial cells by blood flow, is hypothesised to influence atherosclerotic plaque growth and composition. We developed a methodology for image registration of MR and histology images of advanced human carotid plaques and corresponding WSS data, obtained by MRI and computational fluid dynamics. The image registration method requires four types of input images, in vivo MRI, ex vivo MRI, photographs of transversally sectioned plaque tissue and histology images. These images are transformed to a shared 3D image domain by applying a combination of rigid and non-rigid registration algorithms. Transformation matrices obtained from registration of these images are used to transform subject-specific WSS data to the shared 3D image domain as well. WSS values originating from the 3D WSS map are visualised in 2D on the corresponding lumen locations in the histological sections and divided into eight radial segments. In each radial segment, the correlation between WSS values and plaque composition based on histological parameters can be assessed. The registration method was successfully applied to two carotid endarterectomy specimens. The resulting matched contours from the imaging modalities had Hausdorff distances between 0.57 and 0.70 mm, which is in the order of magnitude of the in vivo MRI resolution. We simulated the effect of a mismatch in the rigid registration of imaging modalities on WSS results by relocating the WSS data with respect to the stack of histology images. A 0.6 mm relocation altered the mean WSS values projected on radial bins on average by 0.59 Pa, compared to the output of original registration. This mismatch of one image slice did not change the correlation between WSS and plaque thickness. In conclusion, we created a method to investigate correlations between WSS and plaque composition.
Collapse
Affiliation(s)
- A. M. Moerman
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | - K. Dilba
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
- Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands
| | - S. Korteland
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | - D. H. J. Poot
- Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands
- Department of Medical Informatics, Erasmus MC, Rotterdam, The Netherlands
| | - S. Klein
- Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands
- Department of Medical Informatics, Erasmus MC, Rotterdam, The Netherlands
| | - A. van der Lugt
- Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, The Netherlands
| | - E. V. Rouwet
- Department of Surgery, Erasmus MC, Rotterdam, The Netherlands
| | - K. van Gaalen
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | - J. J. Wentzel
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | | | - F. J. H. Gijsen
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | - K. Van der Heiden
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
- * E-mail:
| |
Collapse
|
6
|
Jodas DS, Pereira AS, Tavares JMRS. Classification of calcified regions in atherosclerotic lesions of the carotid artery in computed tomography angiography images. Neural Comput Appl 2019. [DOI: 10.1007/s00521-019-04183-z] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.4] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/27/2022]
|
7
|
Chrencik MT, Khan AA, Luther L, Anthony L, Yokemick J, Patel J, Sorkin JD, Sikdar S, Lal BK. Quantitative assessment of carotid plaque morphology (geometry and tissue composition) using computed tomography angiography. J Vasc Surg 2019; 70:858-868. [PMID: 30850296 DOI: 10.1016/j.jvs.2018.11.050] [Citation(s) in RCA: 28] [Impact Index Per Article: 5.6] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/28/2018] [Accepted: 11/26/2018] [Indexed: 10/27/2022]
Abstract
OBJECTIVE Quantification of carotid plaque morphology (geometry and tissue composition) may help stratify risk for future stroke and assess plaque progression or regression in response to medical risk factor modification. We assessed the feasibility and reliability of morphologic measurements of carotid plaques using computed tomography angiography (CTA) and determined the minimum detectable change in plaque features by this approach. METHODS CTA images of both carotid arteries in 50 patients were analyzed by two observers using a semiautomatic image analysis program, yielding 93 observations per user (seven arteries were excluded because of prior stenting). One observer repeated the analyses 4 weeks later. Measurements included total plaque volume; percentage stenosis (by diameter and area); and tissue composition for calcium, lipid-rich necrotic core (LRNC), and intraplaque hemorrhage (IPH). Reliability of measurements was assessed by intraclass and interclass correlation and Bland-Altman plots. Dice similarity coefficient (DSC) and modified Hausdorff distance (MHD) assessed reliability of geometric shape measurements. We additionally computed the minimum amount of change in these features detectable by our approach. RESULTS The cohort was 51% male (mean age, 70.1 years), and 56% had a prior stroke. The mean (± standard deviation) plaque volume was 837.3 ± 431.3 mm3, stenosis diameter was 44.5% ± 25.6%, and stenosis area was 58.1% ± 29.0%. These measurements showed high reliability. Intraclass correlation coefficients for plaque volume, percentage stenosis by diameter, and percentage stenosis by area were 0.96, 0.87, and 0.83, respectively; interclass correlation coefficients were 0.88, 0.84, and 0.78. Intraclass correlations for tissue composition were 0.99, 0.96, and 0.86 (calcium, LRNC, and IPH, respectively), and interclass correlations were 0.99, 0.92, and 0.92. Shape measurements showed high intraobserver (DSC, 0.95 ± 0.04; MHD, 0.16 ± 0.10 mm) and interobserver (DSC, 0.94 ± 0.05; MHD, 0.19 ± 0.12 mm) luminal agreement. This approach can detect a change of at least 3.9% in total plaque volume, 1.2 mm3 in calcium, 4.3 mm3 in LRNC, and 8.6 mm3 in IPH with the same observer repeating measurements and 9.9% in plaque volume, 1.9 mm3 in calcium, 7.9 mm3 in LRNC, and 6.8 mm3 in IPH for two different observers. CONCLUSIONS Carotid plaque geometry (total volume, diameter stenosis, and area stenosis) and tissue composition (calcium, LRNC, and IPH) are measured reliably from clinical CTA images using a semiautomatic image analysis program. The minimum change in plaque volume detectable is ∼4% if the same observer makes both measurements and ∼10% for different observers. Small changes in plaque composition can also be detected reliably. This approach can facilitate longitudinal studies for identifying high-risk plaque features and for quantifying plaque progression or regression after treatment.
Collapse
Affiliation(s)
- Matthew T Chrencik
- Department of Vascular Surgery, University of Maryland School of Medicine, Baltimore, Md; Vascular Service, Veterans Affairs Medical Center, Baltimore, Md
| | - Amir A Khan
- Department of Bioengineering, George Mason University, Fairfax, Va
| | - Lauren Luther
- Department of Vascular Surgery, University of Maryland School of Medicine, Baltimore, Md; Vascular Service, Veterans Affairs Medical Center, Baltimore, Md
| | - Laila Anthony
- Department of Vascular Surgery, University of Maryland School of Medicine, Baltimore, Md; Vascular Service, Veterans Affairs Medical Center, Baltimore, Md
| | - John Yokemick
- Department of Vascular Surgery, University of Maryland School of Medicine, Baltimore, Md; Vascular Service, Veterans Affairs Medical Center, Baltimore, Md
| | - Jigar Patel
- Imaging Service, VA Maryland Health Care System, Baltimore, Md
| | - John D Sorkin
- Baltimore VA Medical Center Geriatric Research, Education, and Clinical Center, Baltimore Veterans Affairs Medical Center, Baltimore, Md; Claude D. Pepper Older Americans Independence Center, University of Maryland School of Medicine, Baltimore, Md
| | | | - Brajesh K Lal
- Department of Vascular Surgery, University of Maryland School of Medicine, Baltimore, Md; Vascular Service, Veterans Affairs Medical Center, Baltimore, Md.
| |
Collapse
|
8
|
Sheahan M, Ma X, Paik D, Obuchowski NA, St. Pierre S, Newman WP, Rae G, Perlman ES, Rosol M, Keith JC, Buckler AJ. Atherosclerotic Plaque Tissue: Noninvasive Quantitative Assessment of Characteristics with Software-aided Measurements from Conventional CT Angiography. Radiology 2018; 286:622-631. [PMID: 28858564 PMCID: PMC5790306 DOI: 10.1148/radiol.2017170127] [Citation(s) in RCA: 56] [Impact Index Per Article: 9.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/26/2023]
Abstract
Purpose To (a) evaluate whether plaque tissue characteristics determined with conventional computed tomographic (CT) angiography could be quantitated at higher levels of accuracy by using image processing algorithms that take characteristics of the image formation process coupled with biologic insights on tissue distributions into account by comparing in vivo results and ex vivo histologic findings and (b) assess reader variability. Materials and Methods Thirty-one consecutive patients aged 43-85 years (average age, 64 years) known to have or suspected of having atherosclerosis who underwent CT angiography and were referred for endarterectomy were enrolled. Surgical specimens were evaluated with histopathologic examination to serve as standard of reference. Two readers used lumen boundary to determine scanner blur and then optimized component densities and subvoxel boundaries to best fit the observed image by using semiautomatic software. The accuracy of the resulting in vivo quantitation of calcification, lipid-rich necrotic core (LRNC), and matrix was assessed with statistical estimates of bias and linearity relative to ex vivo histologic findings. Reader variability was assessed with statistical estimates of repeatability and reproducibility. Results A total of 239 cross sections obtained with CT angiography and histologic examination were matched. Performance on held-out data showed low levels of bias and high Pearson correlation coefficients for calcification (-0.096 mm2 and 0.973, respectively), LRNC (1.26 mm2 and 0.856), and matrix (-2.44 mm2 and 0.885). Intrareader variability was low (repeatability coefficient ranged from 1.50 mm2 to 1.83 mm2 among tissue characteristics), as was interreader variability (reproducibility coefficient ranged from 2.09 mm2 to 4.43 mm2). Conclusion There was high correlation and low bias between the in vivo software image analysis and ex vivo histopathologic quantitative measures of atherosclerotic plaque tissue characteristics, as well as low reader variability. Software algorithms can mitigate the blurring and partial volume effects of routine CT angiography acquisitions to produce accurate quantification to enhance current clinical practice. Clinical trial registration no. NCT02143102 © RSNA, 2017 Online supplemental material is available for this article. An earlier incorrect version of this article appeared online. This article was corrected on September 15, 2017.
Collapse
Affiliation(s)
- Malachi Sheahan
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Xiaonan Ma
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - David Paik
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Nancy A. Obuchowski
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Samantha St. Pierre
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - William P. Newman
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Guenevere Rae
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Eric S. Perlman
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Michael Rosol
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - James C. Keith
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| | - Andrew J. Buckler
- From the Louisiana State University Health Sciences Center, New Orleans, La (M.S., W.P.N., G.R.); Elucid Bioimaging, 225 Main St, Wenham, MA 01984 (X.M., D.P., S.S.P., M.R., J.C.K., A.J.B.); Department of Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio (N.A.O.); and Perlman Advisory Group, Boynton Beach, Fla (E.S.P.)
| |
Collapse
|
9
|
Jodas DS, Pereira AS, R.S. Tavares JM. Lumen segmentation in magnetic resonance images of the carotid artery. Comput Biol Med 2016; 79:233-242. [DOI: 10.1016/j.compbiomed.2016.10.021] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/22/2016] [Revised: 10/24/2016] [Accepted: 10/24/2016] [Indexed: 11/15/2022]
|
10
|
Lekadir K, Galimzianova A, Betriu A, Del Mar Vila M, Igual L, Rubin DL, Fernandez E, Radeva P, Napel S. A Convolutional Neural Network for Automatic Characterization of Plaque Composition in Carotid Ultrasound. IEEE J Biomed Health Inform 2016; 21:48-55. [PMID: 27893402 DOI: 10.1109/jbhi.2016.2631401] [Citation(s) in RCA: 93] [Impact Index Per Article: 11.6] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/15/2022]
Abstract
Characterization of carotid plaque composition, more specifically the amount of lipid core, fibrous tissue, and calcified tissue, is an important task for the identification of plaques that are prone to rupture, and thus for early risk estimation of cardiovascular and cerebrovascular events. Due to its low costs and wide availability, carotid ultrasound has the potential to become the modality of choice for plaque characterization in clinical practice. However, its significant image noise, coupled with the small size of the plaques and their complex appearance, makes it difficult for automated techniques to discriminate between the different plaque constituents. In this paper, we propose to address this challenging problem by exploiting the unique capabilities of the emerging deep learning framework. More specifically, and unlike existing works which require a priori definition of specific imaging features or thresholding values, we propose to build a convolutional neural network (CNN) that will automatically extract from the images the information that is optimal for the identification of the different plaque constituents. We used approximately 90 000 patches extracted from a database of images and corresponding expert plaque characterizations to train and to validate the proposed CNN. The results of cross-validation experiments show a correlation of about 0.90 with the clinical assessment for the estimation of lipid core, fibrous cap, and calcified tissue areas, indicating the potential of deep learning for the challenging task of automatic characterization of plaque composition in carotid ultrasound.
Collapse
|
11
|
Koppal S, Warntjes M, Swann J, Dyverfeldt P, Kihlberg J, Moreno R, Magee D, Roberts N, Zachrisson H, Forssell C, Länne T, Treanor D, de Muinck ED. Quantitative fat and R2* mapping in vivo to measure lipid-rich necrotic core and intraplaque hemorrhage in carotid atherosclerosis. Magn Reson Med 2016; 78:285-296. [PMID: 27510300 DOI: 10.1002/mrm.26359] [Citation(s) in RCA: 9] [Impact Index Per Article: 1.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/24/2016] [Revised: 06/29/2016] [Accepted: 07/07/2016] [Indexed: 11/08/2022]
Abstract
PURPOSE The aim of this work was to quantify the extent of lipid-rich necrotic core (LRNC) and intraplaque hemorrhage (IPH) in atherosclerotic plaques. METHODS Patients scheduled for carotid endarterectomy underwent four-point Dixon and T1-weighted magnetic resonance imaging (MRI) at 3 Tesla. Fat and R2* maps were generated from the Dixon sequence at the acquired spatial resolution of 0.60 × 0.60 × 0.70 mm voxel size. MRI and three-dimensional (3D) histology volumes of plaques were registered. The registration matrix was applied to segmentations denoting LRNC and IPH in 3D histology to split plaque volumes in regions with and without LRNC and IPH. RESULTS Five patients were included. Regarding volumes of LRNC identified by 3D histology, the average fat fraction by MRI was significantly higher inside LRNC than outside: 12.64 ± 0.2737% versus 9.294 ± 0.1762% (mean ± standard error of the mean [SEM]; P < 0.001). The same was true for IPH identified by 3D histology, R2* inside versus outside IPH was: 71.81 ± 1.276 s-1 versus 56.94 ± 0.9095 s-1 (mean ± SEM; P < 0.001). There was a strong correlation between the cumulative fat and the volume of LRNC from 3D histology (R2 = 0.92) as well as between cumulative R2* and IPH (R2 = 0.94). CONCLUSION Quantitative mapping of fat and R2* from Dixon MRI reliably quantifies the extent of LRNC and IPH. Magn Reson Med 78:285-296, 2017. © 2016 International Society for Magnetic Resonance in Medicine.
Collapse
Affiliation(s)
- Sandeep Koppal
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Division of Cardiovascular Medicine, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Marcel Warntjes
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Division of Cardiovascular Medicine, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden.,SyntheticMR AB, Linköping, Sweden
| | - Jeremy Swann
- School of Computing, University of Leeds, Leeds, United Kingdom
| | - Petter Dyverfeldt
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Division of Cardiovascular Medicine, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Johan Kihlberg
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Department of Radiology, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Rodrigo Moreno
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,KTH, Royal Institute of Technology, Stockholm, Sweden
| | - Derek Magee
- School of Computing, University of Leeds, Leeds, United Kingdom
| | - Nicholas Roberts
- Division of Brain Sciences, Department of Medicine, Institute of Neurology, Imperial College, London, United Kingdom
| | - Helene Zachrisson
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Division of Cardiovascular Medicine, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Claes Forssell
- Department of Thoracic and Vascular Surgery, and Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Toste Länne
- Department of Thoracic and Vascular Surgery, and Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| | - Darren Treanor
- Department of Pathology and Tumor Biology, Leeds Institute of Molecular Medicine, University of Leeds, Leeds, United Kingdom.,Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden
| | - Ebo D de Muinck
- Center for Medical Image Science and Visualization (CMIV), Linköping University, Linköping, Sweden.,Division of Cardiovascular Medicine, Department of Medical and Health Sciences, Linköping University, Linköping, Sweden
| |
Collapse
|
12
|
Abstract
Plaque imaging by MR imaging provides a wealth of information on the characteristics of individual plaque that may reveal vulnerability to rupture, likelihood of progression, or optimal treatment strategy. T1-weighted and T2-weighted images among other options reveal plaque morphology and composition. Dynamic contrast-enhanced-MR imaging reveals plaque activity. To extract this information, image processing tools are needed. Numerous approaches for analyzing such images have been developed, validated against histologic gold standards, and used in clinical studies. These efforts are summarized in this article.
Collapse
Affiliation(s)
- Huijun Chen
- Department of Biomedical Engineering, Center for Biomedical Imaging Research, School of Medicine, Tsinghua University, Room No. 109, Haidian District, Beijing, China
| | - Qiang Zhang
- Department of Biomedical Engineering, Center for Biomedical Imaging Research, School of Medicine, Tsinghua University, Room No. 120, Haidian District, Beijing, China
| | - William Kerwin
- Department of Radiology, School of Medicine, University of Washington, 850 Republican Street, Seattle, WA 98109, USA.
| |
Collapse
|
13
|
Gao S, van 't Klooster R, van Wijk DF, Nederveen AJ, Lelieveldt BPF, van der Geest RJ. Repeatability of in vivo quantification of atherosclerotic carotid artery plaque components by supervised multispectral classification. MAGNETIC RESONANCE MATERIALS IN PHYSICS BIOLOGY AND MEDICINE 2015; 28:535-45. [PMID: 26162931 PMCID: PMC4651977 DOI: 10.1007/s10334-015-0495-2] [Citation(s) in RCA: 14] [Impact Index Per Article: 1.6] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 01/06/2015] [Revised: 06/24/2015] [Accepted: 06/29/2015] [Indexed: 12/17/2022]
Abstract
Objective
To evaluate the agreement and scan–rescan repeatability of automated and manual plaque segmentation for the quantification of in vivo carotid artery plaque components from multi-contrast MRI. Materials and methods Twenty-three patients with 30–70 % stenosis underwent two 3T MR carotid vessel wall exams within a 1 month interval. T1w, T2w, PDw and TOF images were acquired around the region of maximum vessel narrowing. Manual delineation of the vessel wall and plaque components (lipid, calcification, loose matrix) by an experienced observer provided the reference standard for training and evaluation of an automated plaque classifier. Areas of different plaque components and fibrous tissue were quantified and compared between segmentation methods and scan sessions. Results In total, 304 slices from 23 patients were included in the segmentation experiment, in which 144 aligned slice pairs were available for repeatability analysis. The correlation between manual and automated segmented areas was 0.35 for lipid, 0.66 for calcification, 0.50 for loose matrix and 0.82 for fibrous tissue. For the comparison between scan sessions, the coefficient of repeatability of area measurement obtained by automated segmentation was lower than by manual delineation for lipid (9.9 vs. 17.1 mm2), loose matrix (13.8 vs. 21.2 mm2) and fibrous tissue (24.6 vs. 35.0 mm2), and was similar for calcification (20.0 vs. 17.6 mm2). Conclusion Application of an automated classifier for segmentation of carotid vessel wall plaque components from in vivo MRI results in improved scan–rescan repeatability compared to manual analysis.
Collapse
Affiliation(s)
- Shan Gao
- Division of Image Processing, Department of Radiology, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands
| | - Ronald van 't Klooster
- Division of Image Processing, Department of Radiology, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands
| | - Diederik F van Wijk
- Department of Vascular Medicine, Academic Medical Center, Amsterdam, The Netherlands
| | - Aart J Nederveen
- Department of Radiology, Academic Medical Center, Amsterdam, The Netherlands
| | - Boudewijn P F Lelieveldt
- Division of Image Processing, Department of Radiology, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands
| | - Rob J van der Geest
- Division of Image Processing, Department of Radiology, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands.
| |
Collapse
|
14
|
Nieuwstadt HA, Fekkes S, Hansen HHG, de Korte CL, van der Lugt A, Wentzel JJ, van der Steen AFW, Gijsen FJH. Carotid plaque elasticity estimation using ultrasound elastography, MRI, and inverse FEA - A numerical feasibility study. Med Eng Phys 2015; 37:801-7. [PMID: 26130603 DOI: 10.1016/j.medengphy.2015.06.003] [Citation(s) in RCA: 11] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/10/2014] [Revised: 06/02/2015] [Accepted: 06/07/2015] [Indexed: 12/13/2022]
Abstract
The material properties of atherosclerotic plaques govern the biomechanical environment, which is associated with rupture-risk. We investigated the feasibility of noninvasively estimating carotid plaque component material properties through simulating ultrasound (US) elastography and in vivo magnetic resonance imaging (MRI), and solving the inverse problem with finite element analysis. 2D plaque models were derived from endarterectomy specimens of nine patients. Nonlinear neo-Hookean models (tissue elasticity C1) were assigned to fibrous intima, wall (i.e., media/adventitia), and lipid-rich necrotic core. Finite element analysis was used to simulate clinical cross-sectional US strain imaging. Computer-simulated, single-slice in vivo MR images were segmented by two MR readers. We investigated multiple scenarios for plaque model elasticity, and consistently found clear separations between estimated tissue elasticity values. The intima C1 (160 kPa scenario) was estimated as 125.8 ± 19.4 kPa (reader 1) and 128.9 ± 24.8 kPa (reader 2). The lipid-rich necrotic core C1 (5 kPa) was estimated as 5.6 ± 2.0 kPa (reader 1) and 8.5 ± 4.5 kPa (reader 2). A scenario with a stiffer wall yielded similar results, while realistic US strain noise and rotating the models had little influence, thus demonstrating robustness of the procedure. The promising findings of this computer-simulation study stimulate applying the proposed methodology in a clinical setting.
Collapse
Affiliation(s)
- H A Nieuwstadt
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands.
| | - S Fekkes
- Department of Radiology and Nuclear Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
| | - H H G Hansen
- Department of Radiology and Nuclear Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
| | - C L de Korte
- Department of Radiology and Nuclear Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands
| | - A van der Lugt
- Department of Radiology, Erasmus MC, Rotterdam, The Netherlands
| | - J J Wentzel
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands
| | - A F W van der Steen
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands; Department of Imaging Science and Technology, Delft University of Technology, Delft, The Netherlands
| | - F J H Gijsen
- Department of Biomedical Engineering, Erasmus MC, Rotterdam, The Netherlands.
| |
Collapse
|
15
|
van Engelen A, van Dijk AC, Truijman MTB, Van't Klooster R, van Opbroek A, van der Lugt A, Niessen WJ, Kooi ME, de Bruijne M. Multi-Center MRI Carotid Plaque Component Segmentation Using Feature Normalization and Transfer Learning. IEEE TRANSACTIONS ON MEDICAL IMAGING 2015; 34:1294-1305. [PMID: 25532205 DOI: 10.1109/tmi.2014.2384733] [Citation(s) in RCA: 13] [Impact Index Per Article: 1.4] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/04/2023]
Abstract
Automated segmentation of plaque components in carotid artery magnetic resonance imaging (MRI) is important to enable large studies on plaque vulnerability, and for incorporating plaque composition as an imaging biomarker in clinical practice. Especially supervised classification techniques, which learn from labeled examples, have shown good performance. However, a disadvantage of supervised methods is their reduced performance on data different from the training data, for example on images acquired with different scanners. Reducing the amount of manual annotations required for each new dataset will facilitate widespread implementation of supervised methods. In this paper we segment carotid plaque components of clinical interest (fibrous tissue, lipid tissue, calcification and intraplaque hemorrhage) in a multi-center MRI study. We perform voxelwise tissue classification by traditional same-center training, and compare results with two approaches that use little or no annotated same-center data. These approaches additionally use an annotated set of different-center data. We evaluate 1) a nonlinear feature normalization approach, and 2) two transfer-learning algorithms that use same and different-center data with different weights. Results showed that the best results were obtained for a combination of feature normalization and transfer learning. While for the other approaches significant differences in voxelwise or mean volume errors were found compared with the reference same-center training, the proposed approach did not yield significant differences from that reference. We conclude that both extensive feature normalization and transfer learning can be valuable for the development of supervised methods that perform well on different types of datasets.
Collapse
|