Research Article
Tumor Segmentation in Contrast-Enhanced Magnetic Resonance Imaging for Nasopharyngeal Carcinoma: Deep Learning with Convolutional Neural Network
Table 2
The segmentation performance of all cases.
| Patients number | Volume of current DL method (cm3) | Volume obtained by the two readers (cm3) | Percent match | Corresponding ratio | Dice similarity coefficient | Recall | Jaccard similarity coefficient |
| 1 | 7.8 | 8.0 | 0.84 | 0.74 | 0.83 | 0.81 | 0.70 | 2 | 5.2 | 6.2 | 0.90 | 0.76 | 0.82 | 0.75 | 0.70 | 3 | 5.1 | 6.4 | 0.89 | 0.72 | 0.80 | 0.72 | 0.66 | 4 | 18.5 | 18.4 | 0.88 | 0.83 | 0.89 | 0.89 | 0.79 | 5 | 7.0 | 6.5 | 0.86 | 0.82 | 0.89 | 0.92 | 0.80 | 6 | 11.9 | 10.6 | 0.85 | 0.82 | 0.89 | 0.94 | 0.81 | 7 | 11.9 | 11.9 | 0.83 | 0.74 | 0.83 | 0.82 | 0.71 | 8 | 11.0 | 10.2 | 0.85 | 0.82 | 0.89 | 0.92 | 0.80 | 9 | 5.9 | 6.7 | 0.90 | 0.78 | 0.84 | 0.79 | 0.73 | 10 | 14.8 | 17.1 | 0.86 | 0.71 | 0.80 | 0.75 | 0.66 | 11 | 4.4 | 4.7 | 0.96 | 0.91 | 0.93 | 0.91 | 0.87 | 12 | 17.3 | 17.9 | 0.94 | 0.89 | 0.92 | 0.90 | 0.85 | 13 | 4.5 | 4.3 | 0.88 | 0.84 | 0.90 | 0.92 | 0.82 | 14 | 46.7 | 45.0 | 0.93 | 0.91 | 0.95 | 0.97 | 0.90 | 15 | 5.8 | 5.7 | 0.88 | 0.82 | 0.88 | 0.89 | 0.79 | 16 | 8.8 | 9.2 | 0.90 | 0.83 | 0.88 | 0.87 | 0.79 | 17 | 5.7 | 6.8 | 0.92 | 0.79 | 0.84 | 0.78 | 0.73 | 18 | 10.7 | 11.2 | 0.94 | 0.89 | 0.92 | 0.90 | 0.85 | 19 | 10.6 | 10.9 | 0.95 | 0.91 | 0.94 | 0.93 | 0.88 | 20 | 8.6 | 8.9 | 0.93 | 0.88 | 0.91 | 0.90 | 0.84 | 21 | 13.3 | 13.9 | 0.93 | 0.87 | 0.91 | 0.89 | 0.83 | 22 | 13.7 | 13.6 | 0.95 | 0.92 | 0.95 | 0.95 | 0.90 | 23 | 6.3 | 6.5 | 0.90 | 0.83 | 0.88 | 0.87 | 0.79 | 24 | 16.2 | 16.5 | 0.95 | 0.92 | 0.94 | 0.94 | 0.89 | 25 | 18.2 | 17.7 | 0.94 | 0.92 | 0.95 | 0.96 | 0.91 | 26 | 11.4 | 11.1 | 0.91 | 0.87 | 0.92 | 0.93 | 0.85 | 27 | 10.8 | 11.0 | 0.88 | 0.80 | 0.87 | 0.86 | 0.76 | 28 | 17.1 | 17.9 | 0.95 | 0.90 | 0.93 | 0.91 | 0.86 | 29 | 12.0 | 12.2 | 0.94 | 0.90 | 0.93 | 0.92 | 0.87 | Mean±Std | - | - | 0.90±0.04 | 0.84±0.06 | 0.89±0.05 | 0.88 ±0.07 | 0.81 ±0.07 |
|
|