Table 1.
Summary of clinical studies for artificial intelligence in peritoneal carcinomatosis in recently 6 years.
| Reference | Study type | Number of cases | Tumor type | Methods | Results | Year |
|---|---|---|---|---|---|---|
| Diagnosis | ||||||
| Dong D et al. [10] | Retrospective | 554 advanced gastric cancer patients | GC with PC | Radiomic nomogram | AUC: 0.920 | 2019 |
| Zixing Huang et al. [11] | Retrospective | 554 patients | GC with PC | DCNN | AUC: 0.900 Sensitivity: 81.0 % Specificity: 87.5 %; |
2020 |
| Menglei Li et al. [12] | Retrospective | 779 patients | CRC with PC | Clinical-radiomics model | AUCs: up to 0.855 | 2020 |
| Zixing Huang et al. [13] | Retrospective | At least 109 cases | GC with PC | DCNN | NR | 2020 |
| Chengmao Zhou et al. [14] | Retrospective | 1080 patients with postoperative GC | GC with PC | ML | AUC: up to 0.938 Accuracy：up to 90.9 % |
2020 |
| Ruijiang Li et al. [15] | Multicenter | 1978 patients | GC | DCNN + PMetNet | AUC: 0.920–0.946 Sensitivity: 75.4%–87.5 % Specificity: 92%–98.2 % |
2021 |
| Bin Zheng et al. [16] | Retrospective | 159 patients | GC with and without PC | ML | AUC: 71.2 % Sensitivity: 43.10 % Specificity: 87.12 % |
2021 |
| Xinyu Jin et al. [17] | Retrospective | 11408 images from 131 patients | PC | Meta-learning-based DL | AUC: 0.877 Sensitivity: 73.4 % Specificity: 95.2 % |
2022 |
| Lili Wang et al. [18] | Retrospective | 810 patients | GC | ML | AUCs of clinical models: 0.902–0.969 AUCs of radiomics models: 0.896–0.975 |
2022 |
| Valentin Bejan et al. [19] | Retrospective | NR | CRC | ML | Optimal accuracy: 0.75. | 2022 |
| Zixu Yuan et al. [20] | Retrospective | 19,814 images from 130 patients | CRC with and without PC | DL | AUC: 0.922 Sensitivity: 93.75 % Specificity: 94.44 % Accuracy: 94.11 % |
2022 |
| Dailun Hou et al. [21] | Multicenter | 88 peritoneal tuberculosis and 90 PC patients | PC | ML | AUC: 0.914–0.971 | 2023 |
| Yanyan Chen et al. [22] | Retrospective | 25 patients | GC with and without PC | Proteomic analysis | NR | 2023 |
| Jihong Liu et al. [23] | Retrospective | 98 laboratory tests and clinical feature | Ovarian cancer | AI model | AUC of 0·949 | 2024 |
| Treatment | ||||||
| Milad Shamsi et al. [24] | Retrospective | NR | PC | Computational model | NR | 2018 |
| J M Bereder et al. [25] | Retrospective | 373 cases | PC | ML | Accuracy: close to 98 % | 2019 |
| Alexandros Laios et al. [26] | Retrospective | 154 cases | OC | ML | Accuracy: 66 % | 2020 |
| Nicholas et al. [27] | Retrospective | 60 cases | OC | ML | NR | 2021 |
| Mohsen et al. [28] | Retrospective | NR | PC | Mathematical model | MCDT increased penetration depth more than 13 times | 2022 |
| Mohamed A. et al. [29] | Retrospective | 1959 CRS-HIPEC procedures | PC | ML model | AUC: 0.74 | 2023 |
| Diederick De Jong et al. [30] | Retrospective | 508 patients with ovarian cancer | OC | ML and explainable AI | AUC: 0.91 | 2023 |
| Predicting recurrence | ||||||
| Ruijiang Li et al. [31] | Retrospective | 2320 patients with gastric cancer | GC | Multitask DL model | AUC: 0.843–0.857 C-index: 0.610–0.668 |
2022 |
| Sun, Zepang et al. [32] | Retrospective | 584 quantitative features from 2005 patients | GC | Radiomics | AUC: 0.721–0.732 | 2023 |
Abbreviation: PC: peritoneal carcinomatosis; AUC: area under curve; DCNN: deep convolutional neural network; ML: machine learning; DL: deep learning; NR: not reported; AI: artificial intelligence; GC: gastric cancer; CRC: colorectal cancer; OC: Ovarian cancer; CRS-HIPEC: cytoreductive surgery and hyperthermic intraperitoneal chemotherapy; MCDT: magnetically controlled drug targeting; PMetNet: peritoneal metastasis network.