Table 2. Cutoff values of radiomics features for predicting thymic carcinoma and mean value in the binarized groups.
| Features | Cut-off value | Mean ± SD |
|---|---|---|
| Volume | ||
| Score =0 (n=26) | 10.1±7.9 | |
| Score =1 (n=35) | >23.6 | 114.8±116.7 |
| Ratio (major axis length/minor axis length) | ||
| Score =0 (n=26) | 1.9±0.4 | |
| Score =1 (n=35) | <1.60 | 1.4±0.2 |
| Compactness | ||
| Score =0 (n=26) | 14.6±1.0 | |
| Score =1 (n=35) | >16.1 | 21.5±6.3 |
| Solidity | ||
| Score =0 (n=26) | 0.95±0.01 | |
| Score =1 (n=35) | <0.92 | 0.85±0.06 |
| Minor axis length | ||
| Score =0 (n=39) | 22.8±7.8 | |
| Score =1 (n=22) | >34.4 | 47.6±8.8 |
| GLCM-energy | ||
| Score =0 (n=39) | 0.23±0.04 | |
| Score =1 (n=22) | >0.29 | 0.36±0.04 |
| GLCM-homogeneity | ||
| Score =0 (n=33) | 0.92±0.01 | |
| Score =1 (n=28) | >0.93 | 0.94±0.01 |
| LZHGE | ||
| Score =0 (n=34) | 4.78×105±6.05×105 | |
| Score =1 (n=27) | >1.71×106 | 5.64×106±3.76×106 |
| Kurtosis | ||
| Score =0 (n=37) | 2.1±0.5 | |
| Score =1 (n=24) | >2.83 | 4.6±2.1 |
SD, standard deviation; GLCM, gray-level co-occurrence matrix; LZHGE, long-zone high-grey level emphasis.