Table 4 Details of the datasets and data processing methods used in the examined studies
Authors | Dataset | Patients | Anatomical area | Lesion | Modality | Pre-processing | Objective |
|---|---|---|---|---|---|---|---|
Kitchen [78] | SPIE PROSTATEx challenge 2016 | - | Prostate | - | ADC, T2, \(K^{trans}\) | Normalization [0,1] | Synthesis |
Hu [79] | Private | 104 | Prostate | Gleason Scores (0 to 9) | Diffusion images | Rotation Normalization Flipping | Synthesis |
Wang [81] | Private / PROSTATEx | CS PCs: 134 non-CS PCa: 226 | Prostate | CS & BPH | ADC | - | Classification |
Yang [82] | Private / PROSTATEx | CS PCs: 134 non-CS PCa: 226 | Prostate | CS & BPH | ADC, T2w | Non-rigid registration to ADC & T2w, manually crop ROI | Classification |
Wang [83] | TJPCa / PROSTATEx | CS PCs: 134 non-CS PCa: 226 | Prostate | CS & BPH | ADC, T2w | Crop & align the ROI, Resize pairs | Classification Localization |
F-Quilez [84] | PROMISE 12 | 80 | Prostate | Whole gland | T2w | Linear interpolation Outlier removal Normalize [0,1] CLAHE | Segmentation |
Yu [85] | PROSTATEx’17 | - | Prostate | Benign & malignant | T2w, ADC, DCE | Crop the ROI | Classification |
Yan [88] | LocalPCa/ PROSTATEx | 135/ CS PCa: 64 non-CS PCa: 124 | Prostate | CS PCa non-CS PCa | ADC | - | Privacy Preservation & Classification |
Gao [89] | Private | 96 pNET | Pancreas | Neuroendocrine Tumors | T1ce | - | Classification |
Gao [90] | Private | Train set: 398 Internal val set: 50 External val: 56 | Pancreas | 7 disease Groups | T1ce | - | Classification |
Haarburger [91] | Private | 408 | Breast | Benign Malignant Healthy | T1w, T1ce, T2w | Rescale intensities Resample resolution Crop to patches | Synthesize |
Sun [92] | Liver100/ BRATS17 | 285/ 100 | Liver/ Brain | 7 groups of lesions (23)/ HGG/LGG | T1w,T1ce, T2w, & FLAIR | Normalize to zero mean & unit variance Crop | Segmentation |