Table 4 Prediction of Lymph Node Status Results and Performance Comparison of Different Quantitative Dual-Energy Computed Tomography Parameters
Parameters | AUC | cut-off point | 95%CI | Threshold | Sensitivity (%) | Specificity (%) | PPV(%) | NPV (%) | Accuracy (%) |
|---|---|---|---|---|---|---|---|---|---|
Short diameter | 0.8140 | 0.4971 | 0.751~0.870 | 0.52 | 70.71 | 79.00 | 69.70 | 80.00 | 74.87 |
SD-40 keV | 0.8550 | 0.6068 | 0.798~0.902 | 70.80 | 85.86 | 75.00 | 75.76 | 81.00 | 78.39 |
SD-70 keV | 0.8200 | 0.5387 | 0.756~0.874 | 29.85 | 86.90 | 67.00 | 69.70 | 77.00 | 73.37 |
Venous phase IC | 0.9500 | 0.7580 | 0.915~0.972 | 3.95 | 79.80 | 96.00 | 88.89 | 86.00 | 87.44 |
Venous phase NIC (%) | 0.9320 | 0.7383 | 0.888~0.961 | 54.00 | 82.83 | 91.00 | 83.84 | 85.00 | 84.42 |
Venous phase Rho | 0.6620 | 0.2876 | 0.584~0.741 | 26.85 | 75.76 | 53.00 | 65.66 | 57.00 | 61.31 |
Venous phase Z | 0.9220 | 0.6982 | 0.881~0.952 | 9.26 | 81.82 | 88.00 | 81.82 | 85.00 | 83.42 |
Venous phase DEI | 0.9430 | 0.7394 | 0.905~0.966 | 0.02 | 93.94 | 80.00 | 82.83 | 85.00 | 83.92 |
Venous phase λHU | 0.8540 | 0.5978 | 0.794~0.902 | 6.5483 | 77.78 | 82.00 | 78.79 | 77.00 | 77.89 |
Multivariable logistic regression Model | 0.9910 | 0.9196 | 0.975~0.997 | 0.66 | 95.96 | 96.00 | 94.95 | 95.00 | 94.97 |