Share：
Share this content in WeChat
X
[Chinese][PDF]120976
• Clinical Article •
Automatic segmentation and classification evaluation based on semi-quantitative score of magnetic resonance imaging of knee articular cartilage
SI Li-ping  XUAN Kai  YAO Wei-wu 

DOI:10.12015/issn.1674-8034.2018.12.009.


[Abstract] Objective: Through machine learning a large number of magnetic resonance images diagnosed of knee cartilage with different degrees of lesions, the automatic classification and segmentation of knee cartilage scores is realized, and the accuracy is evaluated. It has the application value of monitoring and detecting OA occurrence and development process.Materials and Methods: This study retrospectively involved 590 knee magnetic resonance imaging 3.0 T data. There were 557 cases of automatic classification data; 33 cases were automatically segmented, of which 27 were used for training and 6 were used for testing. On the basis of deep learning, the migration learning method is used to perform automatic segmentation and classification based on two neural network models, namely V-type network and Inception network. The segmentation of the cartilage is manually labeled by a radiologist for comparison. Classification of cartilage is performed by radiologists based on WORMS zoning method and recht score; quantitative index of automatic segmentation accuracy is calculated using Dice similarity coefficient.Results: The dice similarity coefficient value (DSC) of all knee joints was more than 0.90, and the Dice similarity coefficient of knee articular cartilage was more than 0.70, indicating that the automatic segmentation method proposed in this study can accurately segment the bones and cartilage of knee joint.Conclusions: Methods of automatic segmentation of knee joint and classification of cartilage lesions were established by using deep neural network. The V-type network and inception network demonstrated the efficiency and accuracy of quickly generating accurate segmentation and classification results, which can obtain more accurate results and can be used for extraction of morphological characteristics, with the value of monitoring and diagnosing OA, providing technical support for the automation of imaging diagnosis and treatment process.
[Keywords] Knee joint；Osteoarthritis；Cartilage, articular；Automatic segmentation；Automatic classification；Magnetic resonance imaging

SI Li-ping Department of Radiology, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, Shanghai 200030, China

XUAN Kai Institute for Medical Imaging Technology, School of Biomedical Engineering, Shanghai Jiao Tong University, Shanghai 200030, China

YAO Wei-wu* Department of Radiology, Shanghai Jiao Tong University Affiliated Sixth People’s Hospital, Shanghai 200030, China

*Correspondence to: Yao WW, E-mail: yaoweiwuhuan@163.com

Conflicts of interest   None.

ACKNOWLEDGMENTS  This work was part of National Natural Science Foundation of China No.81771790
Received  2018-07-09
DOI: 10.12015/issn.1674-8034.2018.12.009
DOI:10.12015/issn.1674-8034.2018.12.009.

[1]
Guermazi A, Alizai H, Crema MD, et al. Compositional MRI techniques for evaluation of cartilage degeneration in osteoarthritis. Osteoarthritis Cartilage, 2015, 23(10): 1639-1653.
[2]
Christensen R, Bartels EM, Astrup A, et al. Effect of weight reduction in obese patients diagnosed with knee osteoarthritis: a systematic review and meta-analysis. Ann Rheum Dis, 2007, 66(4): 433-439.
[3]
Roddy E, Zhang W, Doherty M. Aerobic walking or strengthening exercise for osteoarthritis of the knee? A systematic review. Ann Rheum Dis, 2005, 64(4): 544-548.
[4]
Fransen M, McConnell S. Land-based exercise for osteoarthritis of the knee: a metaanalysis of randomized controlled trials. J Rheumatol, 2009, 36(6): 1109-1117.
[5]
Quatman CE, Hettrich CM, Schmitt LC, et al. The clinical utility and diagnostic performance of magnetic resonance imaging for identification of early and advanced knee osteoarthritis: a systematic review. Am J Sports Med, 2011, 39(7): 1557-1568.
[6]
Menashe L, Hirko K, Losina E, et al. The diagnostic performance of MRI in osteoarthritis: a systematic review and meta-analysis. Osteoarthritis Cartilage, 2012, 20(1): 13-21.
[7]
Milletari F, Navab N, Ahmadi SA. V-Net: fully convolutional neural networks for volumetric medical image segmentation. Fourth international conference on 3D vision (3DV), Stanford, CA, USA. 2016: 565-571.
[8]
Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision. (2015-12-11)[2018-09-02]. https://arxiv.org/abs/1512.00567.
[9]
Pan SJ, Yang Q. A survey on transfer learning. IEEE Trans Knowl Data Eng, 2010, 22(10): 1345-1359.
[10]
Deng J, Dong W, Socher R, et al. ImageNet: A large-scale hierarchical image database. IEEE Conference on Computer Vision and Pattern Recognition, Miami, FL, USA. 2009: 248-255.
[11]
Heimann T, Morrison BJ, Styner MA, et al. Segmentation of knee images: a grand challenge. Proc MICCAI Workshop on Medical Image Analysis for the Clinic. 2010: 207-214.
[12]
Paszke A. Automatic differentiation in PyTorch. (2017-10-28)[2018-09-02]. https://openreview.net/forum?id=BJJsrmfCZ.
[13]
Recht MP, Kramer J, Marcelis S, et al. Abnormalities of articular cartilage in the knee: analysis of available MR techniques. Radiology, 1993, 187(2): 473-478.
[14]
陈群，冯阳．关节软骨病理损害的早期评价方法: MRI和关节镜图像对照．中国临床康复, 2005, 9(30): 133-135.
[15]
Peterfy CG, Guermazi A, Zaim S, et al. Whole-organ magnetic resonance imaging score (WORMS) of the knee in osteoarthritis. Osteoarthritis Cartilage, 2004, 12(3): 177-190.
[16]
Hunter DJ, Guermazi A, Lo GH, et al. Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI Osteoarthritis Knee Score). Osteoarthritis Cartilage, 2011, 19(8): 990-1002.
[17]
Schwaiger BJ, Gersing AS, Mbapte Wamba J, et al. Can signal abnormalities detected with MR imaging in knee articular cartilage be used to predict development of morphologic cartilage defects? 48-month data from the osteoarthritis initiative. Radiology, 2016, 281(1): 158-167.
[18]
Markhardt BK, Kijowski R. The clinical significance of dark cartilage lesions identified on MRI. AJR Am J Roentgenol, 2015, 205(6): 1251-1259.
[19]
Bredella MA, Tirman PF, Peterfy CG, et al. Accuracy of T2-weighted fast spin-echo MR imaging with fat saturation in detecting cartilage defects in the knee: comparison with arthroscopy in 130 patients. AJR Am J Roentgenol, 1999, 172(4): 1073-1080.
[20]
Wissman RD, Ingalls J, Nepute J, et al. The trochlear cleft: the "black line" of the trochlear trough. Skeletal Radiol, 2012, 41(9): 1121-1126.
[21]
Nibali A, He Z, Wollersheim D. Pulmonary nodule classification with deep residual networks. Int J Comput Assist Radiol Surg, 2017, 12(10): 1799-1808.
[22]
Lakhani P, Sundaram B. Deep learning at chest radiography: Automated classification of pulmonary tuberculosis by using convolutional neural networks. Radiology, 2017, 284(2): 574-582.
[23]
Mohamed AA, Berg WA, Peng H, et al. A deep learning method for classifying mammographic breast density categories. Med Phys, 2018, 45(1): 314-321.
[24]
Zreik M, Lessmann N, van Hamersvelt RW, et al. Deep learning analysis of the myocardium in coronary CT angiography for identification of patients with functionally significant coronary artery stenosis. Med Image Anal, 2018, 44: 72-85.
[25]
Wang X, Yang W, Weinreb J, et al. Searching for prostate cancer by fully automated magnetic resonance imaging classification: deep learning versus non-deep learning. Sci Rep, 2017, 7(1): 15415.
[26]
Xue Y, Zhang R, Deng Y, et al. A preliminary examination of the diagnostic value of deep learning in hip osteoarthritis. PLoS One, 2017, 12(6): e0178992.
[27]
Artaechevarria X, Munoz-Barrutia A, Ortiz-de-Solorzano C. Combination strategies in multi-atlas image segmentation: application to brain MR data. IEEE Trans Med Imaging, 2009, 28(8): 1266-1277.
[28]
Hafezi-Nejad N, Demehri S, Guermazi A, et al. Osteoarthritis year in review 2017: updates on imaging advancements. Osteoarthritis Cartilage, 2018, 26(3): 341-349.
[29]
Link TM, Stahl R, Woertler K. Cartilage imaging: motivation, techniques, current and future significance. Eur Radiol, 2007, 17(5): 1135-1146.

PREV The diagnostic value of mp-MRI for detecting transition zone prostate cancer based on PI-RADS V2
NEXT The ALFF in the prefrontal lobe of Tai Chi exercisers could predict their emotion control ability: A study of resting state
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn