Radiomics Analysis of Lymph Nodes with Esophageal Squamous Cell Carcinoma Based on Deep Learning
Table 3
The predictive performance of multiple radiomic models in the training and test cohorts.
Models
Training cohort
Test cohort
SEN
SPE
ACC
AUC
SEN
SPE
ACC
AUC
Model I-AdaBoost
0.81
0.73
0.77
0.90 (0.86, 0.94)
0.85
0.61
0.73
0.74 (0.64, 0.84)
Model II-AdaBoost
0.78
0.76
0.77
0.90 (0.85, 0.93)
0.78
0.61
0.70
0.76 (0.66, 0.86)
Model III-AdaBoost
0.91
0.86
0.88
0.95 (0.93, 0.98)
0.78
0.74
0.76
0.78 (0.69, 0.88)
Model I-SVM
0.80
0.67
0.73
0.82 (0.77, 0.88)
0.74
0.57
0.65
0.70 (0.60, 0.81)
Model II-SVM
0.63
0.61
0.62
0.66 (0.59, 0.73)
0.67
0.63
0.65
0.71 (0.61, 0.82)
Model III-SVM
0.89
0.81
0.85
0.93 (0.90, 0.96)
0.61
0.72
0.66
0.72 (0.62, 0.82)
Model I-RF
0.69
0.65
0.67
0.77 (0.71, 0.83)
0.70
0.70
0.70
0.74 (0.63, 0.84)
Model II-RF
0.75
0.72
0.74
0.80 (0.74, 0.86)
0.78
0.70
0.74
0.79 (0.70, 0.88)
Model III-RF
0.74
0.78
0.76
0.83 (0.78, 0.89)
0.76
0.76
0.76
0.80 (0.71, 0.89)
SEN, sensitivity; SPE, specificity; ACC, accuracy; AUC, area under the receiver operating characteristic curve; 95% confidence intervals are included in parentheses.
