14 June 2022>: Clinical Research
Fully Automatic Knee Joint Segmentation and Quantitative Analysis for Osteoarthritis from Magnetic Resonance (MR) Images Using a Deep Learning Model
Xiongfeng Tang 1ABCEF , Deming Guo 1BF , Aie Liu 2ACDE , Dijia Wu 2ADE , Jianhua Liu 3BCD , Nannan Xu 3BDF , Yanguo Qin 1ACDG*DOI: 10.12659/MSM.936733
Med Sci Monit 2022; 28:e936733
Table 3 ICC and Spearman results of morphology analysis for the training and test sets.
Biomarkers | Training set | Test set | |||||||
---|---|---|---|---|---|---|---|---|---|
MAD | ICC | R value | P value | MAD | ICC | R value | P value | ||
Volume (mm) | FC | 86.644 | 0.923 | 0.89 | 416.48 | 0.902 | 0.89 | ||
LTC | 97.886 | 0.955 | 0.927 | 56.855 | 0.913 | 0.927 | |||
MTC | 181.832 | 0.938 | 0.933 | 128.818 | 0.889 | 0.933 | |||
LM | 56.978 | 0.973 | 0.956 | 39.507 | 0.944 | 0.956 | |||
MM | 22.976 | 0.976 | 0.957 | 54.108 | 0.932 | 0.9857 | |||
Thickness (mm) | FC | 0.041 | 0.88 | 0.806 | 0.117 | 0.832 | 0.806 | ||
LTC | 0.003 | 0.912 | 0.854 | 0.062 | 0.889 | 0.854 | |||
MTC | 0.059 | 0.91 | 0.858 | 0.131 | 0.841 | 0.858 | |||
LM | 0.041 | 0.96 | 0.908 | 0.081 | 0.899 | 0.908 | |||
MM | 0.036 | 0.959 | 0.927 | 0.139 | 0.918 | 0.927 | |||
JSW (mm) | L_JSW | 0.367 | 0.895 | 0.813 | 0.224 | 0.889 | 0.813 | ||
M_JSW | 0.275 | 0.871 | 0.818 | 0.144 | 0.909 | 0.818 | |||
Coverage (%) | L_Cov | 0.06 | 0.93 | 0.861 | 5.1 | 0.873 | 0.861 | ||
M_Cov | 0.02 | 0.941 | 0.892 | 5.7 | 0.856 | 0.892 | |||
MAD, mean absolute difference; ICC, intraclass correlation coefficient; ICC values 0.90 are indicative of poor, moderate, good, and excellent reliability, respectively. R value calculated by Spearman’s test; When value was less than 0.05, R value 0.8–1.0 was indicative of very strong correlation. FC – femoral cartilage; LM – lateral meniscus; MM – medial meniscus; LTC – lateral tibial cartilage; MTC – medial tibial cartilage; JSW – joint space width. |