Research Article
A CT-Based Radiomics Nomogram Model for Differentiating Primary Malignant Melanoma of the Esophagus from Esophageal Squamous Cell Carcinoma
Table 2
Parameters of radiomics analysis.
| | Feature name | Image | AUC |
| | Original_shape_sphericity | Plain CT | 0.668 (0.551-0.785) | | Wavelet-HLH_glrlm_low gray level run emphasis | Plain CT | 0.647 (0.536-0.759) | | Wavelet-HLH_glszm_size zone nonuniformity normalized | Plain CT | 0.685 (0.572-0.799) | | Wavelet-HLL_glcm_maximum probability | Plain CT | 0.705 (0.597-0.813) | | Wavelet-LHH_firstorder_minimum | Plain CT | 0.619 (0.497-0.740) | | Wavelet-HLH_firstorder_maximum | Enhanced CT | 0.639 (0.513-0.765) | | Wavelet-HLL_ngtdm_busyness | Enhanced CT | 0.707 (0.584-0.829) | | Wavelet-LHH_glcm_autocorrelation | Enhanced CT | 0.676 (0.558-0.793) | | Wavelet-LHL_glszm_gray level nonuniformity normalized | Enhanced CT | 0.676 (0.562-0.789) |
|
|
CT: computed tomography; AUC: area under curve. Data in parentheses are 95% confidence intervals.
|
