AUTHOR=Wang Shouchao , Xiao Feng , Sun Wenbo , Yang Chao , Ma Chao , Huang Yong , Xu Dan , Li Lanqing , Chen Jun , Li Huan , Xu Haibo 

TITLE=Radiomics Analysis Based on Magnetic Resonance Imaging for Preoperative Overall Survival Prediction in Isocitrate Dehydrogenase Wild-Type Glioblastoma

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 15 - 2021

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2021.791776

DOI=10.3389/fnins.2021.791776

ISSN=1662-453X

ABSTRACT=Purpose
This study aimed to develop a radiomics signature for the preoperative prognosis prediction of isocitrate dehydrogenase (IDH)-wildtype glioblastoma (GBM) patients, and to provide personalized assistance for clinical decision-making of different patients.
Materials and methods
A total of 142 IDH-wildtype GBM patients classified using new classification criteria of WHO 2021 from two centers were included in the study and randomly divided into training and test sets. First, their clinical characteristics were screened using univariate cox regression. Then the radiomics features were extracted from tumor areas and peritumor edema areas on their CE-T1WI, T2WI and T2-FLAIR magnetic resonance imaging (MRI) images. Subsequently, the Inter- and intra-class correlation coefficients (ICC) analysis, spearman correlation analysis, univariate cox and the least absolute shrinkage and selection operator (LASSO) cox regression were used step by step for feature selection and construction of a radiomics signature. The combined model was established by integrating the selected clinical characteristics. Kaplan-Meier analysis were performed for the validation of model discrimination ability and C-index was used to evaluate their consistency of prediction. Finally, a Radiomics+Clinical nomogram was generated for personalized prognosis analysis, and then validated using the calibration curve.
Results
Clinical characteristics analysis resulted in the screening of two clinical risk factors. The combination of ICC, spearman correlation analysis, univariate and LASSO cox resulted in the selection of 12 radiomics features, which made up the radiomics signature. Both radiomics and combined models can significantly stratify high and low risk patients (p<0.001 and p<0.05 for training and test sets respectively), and obtained a good prediction consistency (C-index=0.75-0.85). Calibration plots exhibited a good agreement in both one-year survival and two-year survival between model prediction and actual observation.
Conclusion
Radiomics is an independent preoperative noninvasive prognostic tool for patients who newly classified as IDH-wildtype GBM. The constructed nomogram, which combined radiomics features with clinical factors, can predict the overall survival (OS) of IDH-wildtype GBM patients, and may be a new supplement to the treatment guidelines.