Table 2. Summary of characteristics of key studies.
| References | Application | Number of patients | Machine learning algorithm | Feature type | Optimal results |
|---|---|---|---|---|---|
| Diagnosis | |||||
| Wang 2020 (19) | T staging | 244 | RF | Radiomics | AUC, 0.899 |
| Sun 2020 (20) | T staging | 572 | SVM, ANN | Radiomics, DL | AUC, 0.900 |
| Ma 2017 (21) | Differentiating Borrmann type IV GC from PGL | 70 | LASSO | Radiomics | AUC, 0.903 |
| Feng 2022 (22) | Differentiating Borrmann type IV GC from PGL | 438 | Transfer learning | DL | AUC, 0.990 |
| Wang 2021 (23) | Differentiating gastric neuroendocrine carcinomas from adenocarcinomas | 126 | LASSO | Radiomics | AUC, 0.821 |
| Chen 2022 (24) | Differential diffuse-type from signet ring cell GC | 693 | SVM | Radiomics | AUC, 0.918 |
| Metastasis prediction | |||||
| Gao 2020 (15) | Lymph node metastasis | 768 | LASSO | Radiomics | AUC, 0.920 |
| Chen 2020 (18) | Lymphovascular invasion | 160 | LASSO | Radiomics | AUC, 0.856 |
| Dong 2020 (25) | Lymph node metastasis | 730 | SVM, ANN, RF | Radiomics, DL | AUC, 0.822 |
| Wang 2020 (26) | Lymph node metastasis | 247 | RF | Radiomics | AUC, 0.886 |
| Li 2020 (27) | Lymph node metastasis | 204 | SVM, ANN | Radiomics | AUC, 0.840 |
| Jin 2021 (28) | Lymph node metastasis | 1,699 | CNN | DL | AUC, 0.876 |
| Fan 2022 (29) | Lymphovascular invasion | 101 | Adaptive boosting, linear discriminant analysis, logistic regression | Radiomics | AUC, 0.944 |
| Liu 2020 (30) | Peritoneal metastasis | 233 | SVM | Radiomics | AUC, 0.762 |
| Dong 2019 (31) | Peritoneal metastasis | 554 | SVM, ANN, LASSO | Radiomics | AUC, 0.958 |
| Huang 2020 (32) | Peritoneal metastasis | 955 | LASSO | Radiomics | AUC, 0.870 |
| Mirniaharikandehei 2021 (33) | Peritoneal metastasis | 159 | Gradients boosting machine | Radiomics | AUC, 0.69 |
| Chen 2021 (34) | Peritoneal metastasis | 239 | RF | Radiomics | AUC, 0.981 |
| Liu 2021 (35) | Peritoneal metastasis | 599 | LR | Radiomics | AUC, 0.873 |
| Huang 2020 (36) | Peritoneal metastasis | 544 | CNN | DL | AUC, 0.900 |
| Jiang 2021 (37) | Peritoneal metastasis | 1,225 | CNN | DL | AUC, 0.946 |
| Genetic status and molecular subtypes | |||||
| Zhao 2021 (38) | Epstein-Barr virus status | 133 | LASSO | Radiomics | AUC, 0.955 |
| Zhang 2022 (39) | Epstein-Barr virus status | 54 | Decision tree | Radiomics | AUC, 0.870 |
| Wang 2021 (40) | Human epidermal growth factor 2 | 132 | RF | Radiomics | AUC, 0.830 |
| Prognosis prediction | |||||
| Li 2019 (41) | OS | 181 | LASSO | Radiomics | HR, 2.72 |
| Jiang 2018 (42) | OS, DFS | 1,591 | LASSO | Radiomics | HR, 3.308 (OS); HR, 1.742 (DFS) |
| Jin 2021 (43) | OS, DFS | 428 | LASSO | Radiomics | AUC, 0.965 (OS); AUC, 0.824 (DFS) |
| Shin 2021 (44) | RFS | 410 | LASSO | Radiomics | AUC, 0.719 |
| Jiang 2021 (45) | OS, DFS | 1,615 | S-Net | DL | HR, 0.159 (OS); HR, 0.318 (DFS) |
| Zhang 2021 (46) | OS | 640 | Multi-focus and multi-level fusion feature pyramid network | DL | HR, 9.46 |
| Treatment response prediction | |||||
| Jiang 2020 (47) | Chemotherapy response | 1,778 | LASSO | Radiomics | HR, 0.591 |
| Li 2020 (48) | Chemotherapy response | 739 | SVM | Radiomics | HR, 1.526 |
| Li 2022 (49) | Chemotherapy response | 855 | U-net | Radiomics, DL | AUC, 0.797 |
| Xu 2021 (50) | Neoadjuvant chemotherapy | 292 | SVM | Radiomics | AUC, 0.922 |
| Liu 2021 (51) | Neoadjuvant chemotherapy | 69 | LASSO | Radiomics | AUC, 0.934 |
| Wang 2021 (52) | Neoadjuvant chemotherapy | 155 | LASSO | Radiomics | AUC, 0.953 |
| Tan 2020 (53) | Chemotherapy response | 86 | RF | Delta-radiomics | AUC, 0.828 |
| Liang 2022 (54) | PD-1 inhibitor | 87 | Logistic regression, SVM | Radiomics | AUC, 0.865 |
RF, random forest; AUC, area under the curve; SVM, support vector machine; ANN, artificial neural network; DL, deep learning; GC, gastric cancer; PGL, primary gastric lymphoma; LASSO, least absolute shrinkage and selection operator; CNN, convolutional neural network; LR, logistic regression; HR, hazard ratio; OS, overall survival; DFS, disease-free survival; RFS, recurrence-free survival; PD-1, programmed cell death-1.