AUTHOR=Zhang Liqiang , Yao Rui , Gao Jueni , Tan Duo , Yang Xinyi , Wen Ming , Wang Jie , Xie Xiangxian , Liao Ruikun , Tang Yao , Chen Shanxiong , Li Yongmei 

TITLE=An Integrated Radiomics Model Incorporating Diffusion-Weighted Imaging and 18F-FDG PET Imaging Improves the Performance of Differentiating Glioblastoma From Solitary Brain Metastases

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.732704

DOI=10.3389/fonc.2021.732704

ISSN=2234-943X

ABSTRACT=Background: We aimed to develop an integrated radiomics model to improve the performance of differentiating GBM from SBM.   
Methods: One hundred patients with solitary brain tumors (50 with GBM, 50 with SBM) were retrospectively enrolled and randomly assigned to the training set (n= 80) or validation set (n= 20). A total of 4424 radiomic features were obtained from CE-T1WI with the contrast-enhancing and peri-enhancing oedema region, T2WI, DWI-derived ADC, and  18F-FDG PET images. The partial least squares (PLS) regression with 5-fold cross-validation is used to analyze the correlation between different radiomic features and between different modalities. The cross validity analysis was performed to judge whether a new principal component or a new feature dimension can significantly improve the final prediction effect. The principal components with effective interpretation in all radiomic features are projected to a low dimensional space (this paper is 2D). Then, the effective features of the new projection mapping are sent to the random forest classifier to predict the results. The performance of differentiating GBM from SBM was compared between the integrated radiomics model and other radiomics models or non-radiomics methods using the area under the receiver operating characteristics curve (AUC). 
Results: Through the cross-validity analysis of partial least squares, hundreds of radiomic features were projected into a new 2-dimensional space to complete the construction of radiomics model. Compared with the combined radiomics model using DWI+18F-FDG PET (AUC=0.93, P = 0.014), conventional MRI(cMRI)+ DWI (AUC=0.89, P = 0.011), cMRI + 8F-FDG PET (AUC= 0.91, P = 0.015) and single radiomics model using cMRI (AUC=0.85, P = 0.018), DWI (AUC=0.84, P = 0.017) and 18F-FDG PET (AUC=0.85, P = 0.421), the integrated radiomics model (AUC, 0.98) showed more efficient diagnostic performance. The integrated radiomics model (AUC= 0.98) also showed significantly better performance than any single ADC, SUV or TBR parameter (AUC=0.57–0.71, P < 0.05). The integrated radiomics model showed better performance in the training (AUC=0.98) and validation (AUC=0.93) set than any other models and methods, thus demonstrating robustness.
Conclusions: We developed an integrated radiomics model incorporating DWI and 18F-FDG PET, which improved the performance of differentiating GBM from SBM.