Table 3.
Summary of studies’ data mining (from newest to older).
| Author | AI Area | AI Sw | Sw Class | Data-Mining Methods | Validated | Validation Test |
|---|---|---|---|---|---|---|
| Chen et al. [44] | ML | Pyton software and Pyradiomics module | OS | RF, SVM, SGD, KNN | Yes | 5-fold cross-validation |
| Umutlu et al. [35] | DL | Matlab | C | SVM | Yes | 5-fold cross-validation |
| Eifer et al. [43] | ML | Pyradiomics, Scikit-learn, TensorFlow libraries | OS | KNN and RF | Yes | 5-fold cross-validation |
| Jo et al. [51] | nd | NA | NA | NA | No | NA |
| Cheng et al. [42] | ML | R-software | OS | Multivariable regression with the Akaike’s information criterion (AIC) | Yes | 10-fold cross-validation |
| Castaldo et al. [62] | ML | R-software | OS | Additive logistic regression (LogitBoost), RF, LDA | Yes | 3-fold cross-validation |
| Araz et al. [45] | ML | WEKA | OS | SVM, Hoeffding tree, J48, and MLP | Yes | 10-fold cross-validation |
| Satoh et al. [56] | DL | Pytorch | OS | CNN based on Xception | No | NA |
| Takahashi et al. [63] | DL | Pytorch | OS | CNN based on Xception | No | NA |
| Moreau et al. [46] | DL | Python and Phytorch | OS | U-Net | Yes | 5-fold cross-validation |
| Chen et al. [71] | ML | Python | OS | MLP, SVM, RF and XGBoost | Yes | 3-fold cross-validation |
| Umutlu et al. [72] | ML | Matlab | C | SVM | Yes | 5-fold cross-validation |
| Krajnc et al. [28] | ML | NA | NA | RF | Yes | 100-fold MC-cross-validation |
| Weber et al. [53] | ML | Matlab | C | CNN | Yes | bootstrap Gauss test |
| Aide et al. [76] | ML | XLSTAT Software | C | RF | Yes | OOB |
| Li et al. [48] | DL | DCNN-based diagnosis method | IH | 3D CNN | Yes | 5-fold cross-validation |
| Song et al. [49] | ML | R | OS | XGBoost | Yes | NA |
| Choi et al. [37] | DL | CNN-based sofware | OS | CNN | Yes | 3-fold cross-validation |
| Satoh et al. [60] | ML | scikit-learn and data mining framework in Pyton | OS | SVM | Yes | 2-fold cross-validation |
| Li et al. [38] | ML | Scikit-learn, numpy, scipy and math packages in Pyton | OS/C | RF | Yes | 10-fold cross-validation |
| Ou et al. [29] | ML | PYTHON and IBM SPSS | OS/C | LDA | Yes | 10-fold cross-validation |
| Antunovic et al. [39] | ML | STATA/R | C/OS | Univariable and multivariable logistic regression | Yes | 10-fold cross-validation |
| Aide et al. [64] | ML | XLSTAT Software | C | RF | Yes | OOB |
| Lee et al. [40] | ML | R | OS | Multivariable logistic regression | Yes | Cross-validation 10-fold, 5-fold, and leave-one-out methods |
| Huang et al. [77] | ML | Python | IH | SVM, RF, logistic regression | Yes | 3-fold cross-validation |
AI: artificial intelligence; BC: breast cancer; CNN: convolutional neural network; DL: deep learning; FTs: features; KNN: k-nearest neighbors; LDA: linear discriminant analysis; IH: in-house; ML: machine learning; MLP: multi-layer perceptron; NA: not applicable; nd: not defined; OOB: Out-Of-Bag; RF: random forest; SGD: stochastic gradient descent; SVM: support vector machine; Sw: software; TA: texture analysis; XGBoost: eXtreme Gradient Boosting.