Local Time : 06:13 CET

C-0257 - A Preliminary Examination of the Diagnostic Value of Deep Learning in Hip Osteoarthritis

Y. Xue1, T. Jiang2, K. Chen2; 1 BEIJING/CN 2 Beijing/CN Type: Scientific Exhibit
Area of Interest: Computer applications
Imaging Technique: Neural networks
Procedure: Computer Applications-Detection, diagnosis
Special Focus: Arthritides
Add poster to my schedule In your schedule (remove)

Aims and objectives: Osteoarthritis (OA) is the most common form of arthritis that involves inflammation and major structural changes of the joint, resulting in pain and functional disability [1]. Its main symptoms,pain and stiffness, are major reasons undermining the ability of performing daily living activities,especi[...]

Methods and materials: 1.420 anteroposterior (AP) view hip X-ray images were   assessed by 2 chief physicians with more than 20 years    of experience. 219 normal and 201 samples that had OA.1/5 of the images were randomly selected as the test   samples  (43 normal, 40 abnormal), and th[...]

Results: 1.The CNN model : sensitivity 38/40 = 95.0% ;                                    specificity 39/43 = 90.7%. 2.The DAR(Diagnosis A[...]

Conclusion: 1.The CNN model has high accuracy 92.8%, a balance  between high sensitivity of 95.0% and high specificity of  90.7%.         2.The CNN model achieves higher DAR with more experienced  physicians. The CNN model has a similar diagnostic accuracy  to the attendi[...]

Personal information:

References: 1.Cross M, Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: estimates from the Global Burden of Disease 2010 study. Ann Rheum Dis [Internet]. 2014;73(7):1323–30. 2.Nho SJ, Kymes SM, Callaghan JJ, et al. The burden of hip osteoarthritis in the United States: epidemiologic and [...]

This website uses cookies. Learn more