Table 4.
Selected features by least absolute shrinkage and selection operator regression in the multi-region radiomics analysis.
| ROI | Imaging filter | Feature | |
|---|---|---|---|
| Lung-GTV | Wavelet-LLH | glcm | JointAverage |
| Lung-GTV | Wavelet-LLH | firstorder | Minimum |
| Lung-GTV | Wavelet-HHL | gldm | Small Dependence Low Gray Level Emphasis |
| Lung-PTVad20 | Wavelet-LLH | ngtdm | Strength |
| 5.00 Gy | Wavelet-LHL | firstorder | Skewness |
| 20.00 Gy | Wavelet-LHL | firstorder | Skewness |
| 30 Gy ∩ Lung-GTV | Wavelet-HLL | gldm | Gray Level Non-Uniformity |
| 50 Gy ∩ Lung-GTV | wavelet-LHL | glszm | Large Area High Gray Level Emphasis |
| 60 Gy ∩ Lung-GTV | Original | Shape | MajorAxis |
| 60 Gy ∩ Lung-GTV | Original | Shape | Flatness |
| 20 Gy ∩ Lung | Original | Shape | MinorAxis |
| 30 Gy ∩ Lung | Original | Shape | MajorAxis |
| 60 Gy ∩ Lung-PTV | Original | Shape | Maximum 3D Diameter |