Table 2.
Summary of the AI driven models for the pancreatic cancer diagnosis.
| Modality | AI Model | Study Population | Purpose | Sensitivity | Specificity | Accuracy | Reference |
|---|---|---|---|---|---|---|---|
| CT | CNN | 27 | Pancreatic cystic neoplasm malignancy prediction | - | - | 92.9 | Watson et al., 2021 [102] |
| CT | Naïve Bayer classifier | 72 | PDAC identification | - | - | 86 | Ahamed et al., 2022 [103] |
| CT | CNN | 1006 | Pancreas segmentation | - | - | - | Lim et al., 2022 [104] |
| CT | CNN | 68 | Serum tumor marker analysis | 89.31 | 92.31 | - | Qiao et al., 2022 [105] |
| CT | CNN | 513 | Pancreatico enteric Anastomotic Fistulas prediction after a pancreatoduodenectomy | 86.7 | 87.3 | 87.1 | Mu et al., 2020 [106] |
| CT | ANN | 62 | Acute pancreatitis risk prediction | - | - | - | Keogan et al., 2002 [107] |
| CT | Support vector machine | 56 | PDAC histopathological grade discrimination | 78 | 95 | 86 | Qiu et al., 2019 [108] |
| CT | CNN | 370 patients, 320 controls | PC detection | 97.3 (Test set 1) 99 (Test set 2) |
100 (Test set 1) 98.9 (Test set 2) |
98.6(Test set 1) 98.9 (Test set 2) |
Liu et al., 2020 [109] |
| CT | Deep learning | 750 patients 575 controls |
PDAC detection | - | - | 87.8 | Chu et al., 2019 [110] |
| CT | CNN | 222 patients 190 controls |
PC diagnosis | 91.58 | 98.27 | 95.47 | Ma et al., 2020 [89] |
| CT | DCNN | 2890 CT images | Pancreatic cancer detection | 83.76 | 91.79 | 94 | Zhang et al., 2020 [90] |
| CT | Deep learning | 319 | Preoperative pancreatic cancer diagnosis | 86.8 | 69.5 | 87.1 | Si et al., 2021 [111] |
| CT | ANN | 898 | Cancer risk prediction | 80.7 | 80.7 | - | Muhammad et al., 2019 [112] |
| CT | CNN | 669 patients 804 controls |
PC differentiation | 89.7 | 92.8 | - | Chen et al., 2022 [91] |
| MRI | CNN | 139 | Identification of intraductal papillary mucinous neoplasia | 75 | 78 | - | Juan et al., 2019 [78] |
| MRI | CNN | 27 | Automatic image segmentation | - | - | - | Liang et al., 2020 [113] |
| MRI | ANN | 168 | PDAC differentiation | - | - | 96 | Devi et al., 2018 [114] |
| EUS | CNN | 583 | Autoimmune pancreatitis from PDAC | 90 | 85 | - | Marya et al., 2021 [115] |
| EUS | CAD | 920 (Validation) +470 (test) | PDAC detection | - | - | - | Tonozuka et al., 2021 [67] |
| EUS | ANN | 202 (cancerous) & 130 (Non-cancerous) | Computer-aided pancreatic cancer diagnosis using image processing | 83.3 | 93.3 | 87.5 | Ozkan et al., 2019 [65] |
| EUS | ANN | 258 | Pancreatic lesion characterization | - | - | 91 | Saftoiu et al., 2012 [34] |
| EUS | ANN | 388 | PDAC and CP differentiation | 96 | 93 | 94 | Zhu et al., 2013 [63] |
| EUS | ANN | 167 | PDAC and CP differentiation | 94 | 94 | - | Saftoiu et al., 2015 [116] |
| EUS | ANN | 56 | Normal, CP and PDAC differentiation | - | - | 93 | Das et al., 2008 [64] |
| EUS | ANN | 21 | PDAC and CP differentiation | - | - | 89 | Norton et al., 2001 [62] |
| PET/CT | SVM | 80 | Pancreatic cancer segmentation | 95.23 | 97.51 | 96.47 | Li et al., 2018 [100] |