Differentiation of Recurrence from Radiation Necrosis in Gliomas Based on the Radiomics of Combinational Features and Multimodality MRI Images

<div>Estimation of the classification performance of multivariable models constructed from T1C, T2, T1, FLAIR, multimodality, AlexNet, Inception v3, fusion AlexNet, and fusion Inception v3 images using optimal features in the training set (a) and validation set (b) for the model orders 1–10. The optimal degrees of freedom were separately found in terms of the maximum 0.632 + bootstrap AUC for each model order. Error bars represent the standard error of the mean at the 95% confidence interval.</div>

Computational and Mathematical Methods in Medicine

fig3

Figure 3

Figure 3: Differentiation of Recurrence from Radiation Necrosis in Gliomas Based on the Radiomics of Combinational Features and Multimodality MRI Images 

Figure 3 | Differentiation of Recurrence from Radiation Necrosis in Gliomas Based on the Radiomics of Combinational Features and Multimodality MRI Images 