Table 1.
Normalization methods and grey level discretization applied in recent radiomics studies dedicated to brain tumors.
| References | Multicenter | Number of patients | MRI sequences | Normalization technique | Grey-level discretization | Radiomics software | Features | Objective |
|---|---|---|---|---|---|---|---|---|
| Su et al.15 | No | 100 | T2w-flair | – | – | Pyradiomics | 18 first-order, 13 shape, 54 texture | Investigate the feasibility of predicting H3 K27M mutation status by applying an automated machine learning approach to the MR radiomics features of patients with midline gliomas |
| Liu et al.16 | Yes | 130 | T1w, T2w-fl1air | ComBat | – | Artificial Intelligence Kit (GE) | First-order, texture | Develop and validate a model that can be used to predict the individualized treatment response in children with cerebral palsy |
| Bologna et al.17 | – | Phantom | T1w, T2w | Z-Score | 32 FBN | Pyradiomics | 18 first-order, 14 shape, 75 texture | Analysis of virtual phantom for preprocessing evaluation and detection of a robust feature set for MRI-radiomics of the brain |
| Elsheikh et al.18 | Yes | 135 | T1w, T1w-gd, T2w, T2w-flair | – | – | – | First-order, texture | Analysis of multi-stage association of glioblastoma gene expressions with texture and spatial patterns |
| Tixier et al.19 | Yes | 90 | T1w-gd, T2w-flair | – | 128 FBN | CERR | 72 features (first-order, texture, shape) | Study the impact of tumor segmentation variability on the robustness of MRI radiomics features |
| Ortiz-Ramón et al.20 | No | 200 | T1w, T2w, T2w-flair | – | 32 FBN | MATLAB | 114 textures | Identify the presence of ischaemic stroke lesions by means of texture analysis on brain MRI |
| Vamvakas et al.21 | No | 40 | T1w, T1w-gd, T2w, T2w-flair | – | – | MATLAB | 11 first-order, 16 texture | Investigate the value of advanced multiparametric MRI biomarker analysis based on radiomics features and machine learning classification for glioma grading |
| Tixier et al.22 | Yes | 159 | T1w, T1w-gd, T2w-flair | – | 128 FBN | CERR | 286 features (first-order, shape, texture) | Evaluate the capacity of radiomics features to add complementary information to MGMT status, to improve the ability to predict prognosis |
| Wu et al.23 | Yes | 126 | T1w, T1w-gd, T2w, T2w-flair | – | – | – | 704 features (first-order, shape, texture) | Identify the optimal radiomics-based machine learning method for isocitrate dehydrogenase genotype prediction in diffuse gliomas |
| Artzi et al.24 | No | 439 | T1w-gd | WhiteStripe | – | MATLAB | 757 features (first-order, shape, texture) | Differentiate between glioblastoma and brain metastasis subtypes using radiomics analysis |
| Kniep et al.25 | No | 189 | T1w, T1w-gd, T2w-flair | WhiteStripe | – | Pyradiomics | 18 first-order, 17 shape, 56 texture | Investigate the feasibility of tumor type prediction with MRI radiomics image features of different brain metastases in a multiclass machine learning approach for patients with unknown primary lesion at the time of diagnosis |
| Sanghani et al.26 | Yes | 163 | T1w, T1w-gd, T2w, T2w-flair | – | – | Pyradiomics | 2200 features (first-order, shape, texture) | Predict overall survival in glioblastoma multiforme patients from volumetric, shape and texture features using machine learning |
| Liu et al.27 | Yes | 84 | T2w | Z-Score | – | MATLAB | 131 features (first-order, shape, texture) | Develop a radiomics signature for prediction of progression-free survival (PFS) in lower-grade gliomas and investigate the genetic background behind the radiomics signature |
| Peng et al.28 | No | 66 | T1w-gd, T2w-flair | – | 64 FBN | MATLAB | 51 features (first-order, shape, texture) | Distinguish true progression from radionecrosis after stereotactic radiation therapy for brain metastases with machine learning and radiomics |
| Bae et al.29 | No | 217 | T1w-gd, T2w-flair | WhiteStripe | – | Pyradiomics | 796 features (first-order, shape, texture) | Investigate whether radiomics features based on MRI improve survival prediction in patients with glioblastoma multiforme (GBM) when they are integrated with clinical and genetic profiles |
| Chen et al.30 | Yes | 220 | T1w, T1w-gd, T2w, T2w-flair | Nyul | – | Pyradiomics | 420 features (first-order, shape, texture) | Classify gliomas combining automatic segmentation and radiomics |