Table 5. Comparisonof the diagnostic efficacy of IVIM parameters in differentiating TETs based on WHO classification and Masaoka-Koga stage and comparisons of TET ADCs with published data.
| Parameters | AUC | Sensitivity (%) | Specificity (%) | Accuracy (%) | PPV (%) | NPV (%) | Cutoff value |
|---|---|---|---|---|---|---|---|
| LRT vs. HRT+TC | |||||||
| ADCmb (×10−3 mm2/sec) | 0.793 | 56.5 | 90.7 | 79.2 | 70.6 | 81.7 | 1.415 |
| D (×10−3 mm2/sec) | 0.933 | 95.7 | 77.8 | 83.1 | 64.7 | 97.7 | 0.747 |
| D* (×10−3 mm2/sec) | 0.919 | 95.7 | 77.8 | 83.1 | 64.7 | 97.7 | 5.256 |
| Logistic (D, D*) # | 0.959 | 95.7 | 87.0 | 89.6 | 75.9 | 97.9 | 0.193 |
| ADC (×10−3 mm2/sec) (16)* | 0.851 | 87.0 | 85.0 | 86.0 | 87.0 | 85.0 | 1.22 |
| ADC (×10−3 mm2/sec) (17)§ | 0.864 | 94.7 | 63.6 | 78.1 | − | − | 1.309 |
| Early vs. advanced stage | |||||||
| ADCmb (×10−3 mm2/sec) | 0.711 | 78.8 | 56.8 | 64.9 | 56.5 | 77.4 | 1.095 |
| D (×10−3 mm2/sec) | 0.793 | 78.8 | 72.7 | 75.3 | 68.4 | 82.1 | 0.694 |
| D* (×10−3 mm2/sec) | 0.789 | 78.8 | 77.3 | 77.9 | 72.2 | 82.9 | 4.88 |
| ADC (×10−3 mm2/sec) (17)§ | 0.730 | 91.7 | 58.8 | 73.2 | − | − | 1.243 |
AUC = area under curve; PPV = positive predictive value; NPV = negative predictive value; LRT = low risk thymoma; HRT = high risk thymoma; TC = thymic carcinoma; ADCmb = ADC calculated using mono-exponential model DWI (multi b-values: 0 - 1200 sec/mm2); D = ADCslow or pure diffusion coefficient; D* = ADCfast or pseudo-diffusion coefficient.
#The results of Logistic (D, D*) were acquired using group as the dependent variable and the D and D* parameters as covariates to generate a binary logistic regression and P values for each patient. This P value was analyzed using ROC. The actual Logistic (D, D*) for LRT vs. HRT+TC model was as follows: ln[P/(1-P)] = 9.364 - 6.763D - 0.446D*.
*Results from Razek et al. for differentiating LRT from HRT+TC (16).
§Results from Priola et al. for differentiating LRT from HRT (17).