Table 2 The existing research in the area of static ultrasound
From: Application and prospects of AI-based radiomics in ultrasound diagnosis
Reference | Number of patient | Tumor characteristic | Imaging modality | Function and prediction result | Method |
|---|---|---|---|---|---|
Ye et al. [19] | 1844 images | Triple negative breast cancer | BUS | Benign vs TNÂ (AUC): 0.9789, benign vs NTNÂ (AUC): 0.9689, TN vs NTNÂ (AUC): 0.9000 | Resnet50 |
Zhou et al. [20] | 192 | Axillary lymph node metastasis | BUS | Predicting ALN metastasis, AUC = 0.85 | LASSO |
Kwon et al. [21] | 169 | Distant metastasis of follicular thyroid carcinoma | BUS | Distant metastasis classification, AUC = 0.90 | SVM |
Meshram et al. [22] | 101 | Carotid plaque | BUS | Dice coefficients for automatic is 0.55, for semi-automatic is 0.84 | Dilated U-Net |
Wang et al. [23] | 398 | Liver fibrosis | UE | Diagnosing liver fibrosis stages AUC(F4) = 0.97, AUC (≥ F3) = 0.98, AUC (≥ F2) = 0.85 | CNN |
Tahmasebi et al. [24] | 381 | Axillary lymph nodes | UE | Classification of axillary lymph nodes, AuPRC = 0.78 | Google cloud autoML vision, mountain view |
Lu et al. [25] | 807 | Liver fibrosis | UE | Discrimination of significant fibrosis, AUC = 0.91 | CNN |
Zhou et al. [26] | 297 | Liver fibrosis | UE | Assess liver fibrosis stages, AUC (cirrhosis and advanced fibrosis) = 0.98, AUC (significance fibrosis) = 0.76 | CNN |