Research Article
Deep Convolutional Neural Network-Based Brain Magnetic Resonance Imaging Applied in Glioma Diagnosis and Tumor Region Identification
Table 1
The results of analysis of the efficacy of Ktrans, Kep, Ve, and Vp histogram parameters in diagnosing glioma.
| | | | 95% confidence interval (CI) | Sensitivity (%) | Specificity (%) |
| | Ktrans | Mean | 0.88 (0.65,0.99) | 88.6 | 85.1 | | Standard deviation | 0.88 (0.65,0.99) | 84.2 | 85.7 | | Kurtosis | 0.88 (0.65,0.99) | 93.8 | 77.7 | | Skewness | 0.88 (0.65,0.99) | 95.5 | 77.7 |
| | Kep | Standard deviation | 0.88 (0.65,0.99) | 44.3 | 100 | | Skewness | 0.81 (0.77,0.90) | 88.1 | 86.2 |
| | Ve | Mean | 0.83 (0.72,0.96) | 83.9 | 94.1 | | Standard deviation | 0.78 (0.71,0.89) | 77.2 | 68.8 | | Kurtosis | 0.85 (0.69,0.90) | 74.4 | 85.5 | | Skewness | 0.83 (0.78,0.94) | 79.8 | 86.7 |
| | Vp | Mean | 0.88 (0.65,0.99) | 88.2 | 77.3 | | Standard deviation | 0.77 (0.63,0.92) | 95.5 | 70.1 |
|
|
