1Department of Gastroenterology and Hepatology, Changi General Hospital, Singapore Health Services, Singapore
2Academic Medicine Center, Duke-NUS Medical School, Singapore
3Department of Laboratory Medicine, Changi General Hospital, Singapore Health Services, Singapore
4Digestive Disease Center, Showa University Northern Yokohama Hospital, Yokohama, Japan
5Yong Loo Lin School of Medicine, National University of Singapore, Singapore
6Department of General Surgery, Changi General Hospital, Singapore Health Services, Singapore
© 2024 Korean Society of Gastrointestinal Endoscopy
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Conflicts of Interest
The authors have no potential conflicts of interest.
Funding
None.
Author Contributions
Conceptualization: JWL; Data curation: JWL, LMW, KI, TLA; Formal analysis: JWL, LMW, KI, KWL, JCYN; Methodology: JWL, TLA; Supervision: TLA; Writing–original draft: JWL, JCYN; Writing–review & editing: LMW, KI, KWL, JCYN, TLA.
Study | Year published | AI instrument | Data set | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUROC (%) | Accuracy (%) |
---|---|---|---|---|---|---|---|---|---|
Lui et al.46,a) | 2019 | CNN | 8,567 NBI and WLI images | 94.6 (for NBI) | 92.3 (for NBI) | 98.8 (for NBI) | 72.0 (for NBI) | 0.934 | 94.3 (for NBI) |
Luo et al.47 | 2021 | CNN | 9,368 Images (WLI) | 91.2 | 91.0 | 87.6 | 93.7 | 0.970 | 91.1 |
Tokunaga et al.48,b) | 2021 | Single shot multibox detector | 3,442 Images (WLI) | 96.7 | 75 | 90.2 | 90.5 | 0.913 | 90.3 |
Ito et al.49,c) | 2019 | CNN | 190 Conventional WLI images | 67.5 | 89.0 | - | - | 0.871 | 81.2 |
Nakajima et al.50,d) | 2020 | CNN | 1,917 Plain endoscopic images | 81 | 87 | 85 | 83 | 0.888 | 84 |
Lu et al.51,e) | 2022 | CNN | 820,348 WLI and IEE images, 35 videos | 90 | 94.2 | 64.7 | 98.8 | 0.956 | 93.8 |
CAD, computer-aided diagnostic; CRC, colorectal cancer; AI, artificial intelligence; PPV, positive predictive value; NPV, negative predictive value; AUROC, area under the receiver operating characteristic; CNN, Convolutional neural network; NBI, narrow-band imaging; WLI, white-light imaging; IEE, image-enhanced endoscopy.
a)Prediction endoscopically curable lesions (includes sessile serrated adenomas, tubular adenoma with or without villous component, intramucosal adenocarcinoma, and T1a lesions).
b)Differentiation between endoscopically curable lesions (adenomas, high-grade dysplasia, CRC with submucosal invasion <1,000 µm) vs. CRC with submucosal invasion >1,000 µm or advanced CRC.
c)Prediction of Tis/T1a lesions vs T1b lesions.
d)Prediction of T1b CRC.
e)Prediction of lesions with low-grade dysplasia, high-grade dysplasia, intramucosal cancer and CRC with submucosal invasion <1,000 µm vs. CRC with submucosal invasion ≥1,000 µm and advanced CRC.
Study | Year published | AI instrument | Type of data | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUROC | Accuracy (%) | Features used for training |
---|---|---|---|---|---|---|---|---|---|---|
Kwak et al.70 | 2021 | CNN | 164 cases of stage I, II, and III CRCa) | - | - | - | - | 0.677 for PTS score | - | PTS score (consisting of adipose tissue, lymphocytes, mucus, smooth muscle, normal colon mucosa, stroma, colon cancer epithelium) |
Kudo et al.71 | 2021 | ANN | 4073 cases of T1 CRCb) | - | - | - | - | 0.83 | - | Age, sex, tumor size, location, morphology, lymphatic invasion, vascular invasion, histological grade |
0.73c) | ||||||||||
0.57d) | ||||||||||
Kang et al.73 | 2021 | LASSO | 316 cases of T1 CRCa) | 56.1e) | 87.3e) | 39.7e) | 93.0e) | 0.765 | 83.2e) | Histology grade, lymphovascular invasion, tumor budding, background adenoma, CD3_IM, CD3_TC, CD8_IM, CD8_TC, FOXP3_TC |
0.518d) | ||||||||||
Takamatsu et al.75 | 2019 | RFC | 397 cases of T1 CRCb) | 80.0 | 94.5 | - | - | 0.938 | - | Cytokeratin IHC of slides |
0.826d) | ||||||||||
Song et al.76 | 2022 | Deep convolution neural network | 400 cases of T1 CRCb) | 100 | 45 | 32.6 | - | 0.764 | 63.8 | Size of cancer, depth of submucosal invasion, lymphovascular invasion, tumor budding, positive resection margin, microsatellite instability |
100d) | 0d) | 17.5d) | - | - | 17.5d) | |||||
Kasahara et al.77 | 2022 | Support vector machine and random forest | 146 cases of T1b CRCa) | - | - | - | - | - | 91.0 | Cancer cell nuclei and their heterogeneity |
Brockmoeller et al.78 | 2022 | ShuffleNet network model | 203 cases of T1 and T2 CRCa) | - | - | - | - | 0.567 (for T1 CRC), 0.711 (for T2 CRC) | - | Tumor infiltrating lymphocytes, inflamed fat, inflammatory cells at the invasive edge and deeper into the submucosa and into muscularis propria, mesenteric fat, poorly differentiated tumor areas, necrosis, papillary growth pattern |
AI, artificial intelligence; LNM, lymph node metastasis; PPV, positive predictive value; NPV, negative predictive value; AUROC, area under the receiver operating characteristic; CNN, convolutional neural network; CRC, colorectal cancer; PTS, peritumoral stroma; ANN, artificial neural network; LASSO, least absolute shrinkage and selection operator; RFC, random forest classifier; IHC, immunohistochemistry; -, results not reported in study.
a)Study included only surgically resected cases.
b)Study included endoscopically resected, with or without additional surgical resection and lymph node dissection.
c)Prediction of lymph node metastasis based on US guidelines (National Comprehensive Cancer Network).
d)Prediction of lymph node metastasis based on the Japanese Society for Cancer of the Colon and Rectum guidelines.
e)Assuming predicted probability of LNM is 20%.
Study | Year published | AI instrument | Data set | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUROC (%) | Accuracy (%) |
---|---|---|---|---|---|---|---|---|---|
Lui et al.46, |
2019 | CNN | 8,567 NBI and WLI images | 94.6 (for NBI) | 92.3 (for NBI) | 98.8 (for NBI) | 72.0 (for NBI) | 0.934 | 94.3 (for NBI) |
Luo et al.47 | 2021 | CNN | 9,368 Images (WLI) | 91.2 | 91.0 | 87.6 | 93.7 | 0.970 | 91.1 |
Tokunaga et al.48, |
2021 | Single shot multibox detector | 3,442 Images (WLI) | 96.7 | 75 | 90.2 | 90.5 | 0.913 | 90.3 |
Ito et al.49, |
2019 | CNN | 190 Conventional WLI images | 67.5 | 89.0 | - | - | 0.871 | 81.2 |
Nakajima et al.50, |
2020 | CNN | 1,917 Plain endoscopic images | 81 | 87 | 85 | 83 | 0.888 | 84 |
Lu et al.51, |
2022 | CNN | 820,348 WLI and IEE images, 35 videos | 90 | 94.2 | 64.7 | 98.8 | 0.956 | 93.8 |
Study | Year published | AI instrument | Type of data | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | AUROC | Accuracy (%) | Features used for training |
---|---|---|---|---|---|---|---|---|---|---|
Kwak et al.70 | 2021 | CNN | 164 cases of stage I, II, and III CRC |
- | - | - | - | 0.677 for PTS score | - | PTS score (consisting of adipose tissue, lymphocytes, mucus, smooth muscle, normal colon mucosa, stroma, colon cancer epithelium) |
Kudo et al.71 | 2021 | ANN | 4073 cases of T1 CRC |
- | - | - | - | 0.83 | - | Age, sex, tumor size, location, morphology, lymphatic invasion, vascular invasion, histological grade |
0.73 |
||||||||||
0.57 |
||||||||||
Kang et al.73 | 2021 | LASSO | 316 cases of T1 CRC |
56.1 |
87.3 |
39.7 |
93.0 |
0.765 | 83.2 |
Histology grade, lymphovascular invasion, tumor budding, background adenoma, CD3_IM, CD3_TC, CD8_IM, CD8_TC, FOXP3_TC |
0.518 |
||||||||||
Takamatsu et al.75 | 2019 | RFC | 397 cases of T1 CRC |
80.0 | 94.5 | - | - | 0.938 | - | Cytokeratin IHC of slides |
0.826 |
||||||||||
Song et al.76 | 2022 | Deep convolution neural network | 400 cases of T1 CRC |
100 | 45 | 32.6 | - | 0.764 | 63.8 | Size of cancer, depth of submucosal invasion, lymphovascular invasion, tumor budding, positive resection margin, microsatellite instability |
100 |
0 |
17.5 |
- | - | 17.5 |
|||||
Kasahara et al.77 | 2022 | Support vector machine and random forest | 146 cases of T1b CRC |
- | - | - | - | - | 91.0 | Cancer cell nuclei and their heterogeneity |
Brockmoeller et al.78 | 2022 | ShuffleNet network model | 203 cases of T1 and T2 CRC |
- | - | - | - | 0.567 (for T1 CRC), 0.711 (for T2 CRC) | - | Tumor infiltrating lymphocytes, inflamed fat, inflammatory cells at the invasive edge and deeper into the submucosa and into muscularis propria, mesenteric fat, poorly differentiated tumor areas, necrosis, papillary growth pattern |
CAD, computer-aided diagnostic; CRC, colorectal cancer; AI, artificial intelligence; PPV, positive predictive value; NPV, negative predictive value; AUROC, area under the receiver operating characteristic; CNN, Convolutional neural network; NBI, narrow-band imaging; WLI, white-light imaging; IEE, image-enhanced endoscopy. Prediction endoscopically curable lesions (includes sessile serrated adenomas, tubular adenoma with or without villous component, intramucosal adenocarcinoma, and T1a lesions). Differentiation between endoscopically curable lesions (adenomas, high-grade dysplasia, CRC with submucosal invasion <1,000 µm) vs. CRC with submucosal invasion >1,000 µm or advanced CRC. Prediction of Tis/T1a lesions vs T1b lesions. Prediction of T1b CRC. Prediction of lesions with low-grade dysplasia, high-grade dysplasia, intramucosal cancer and CRC with submucosal invasion <1,000 µm vs. CRC with submucosal invasion ≥1,000 µm and advanced CRC.
AI, artificial intelligence; LNM, lymph node metastasis; PPV, positive predictive value; NPV, negative predictive value; AUROC, area under the receiver operating characteristic; CNN, convolutional neural network; CRC, colorectal cancer; PTS, peritumoral stroma; ANN, artificial neural network; LASSO, least absolute shrinkage and selection operator; RFC, random forest classifier; IHC, immunohistochemistry; -, results not reported in study. Study included only surgically resected cases. Study included endoscopically resected, with or without additional surgical resection and lymph node dissection. Prediction of lymph node metastasis based on US guidelines (National Comprehensive Cancer Network). Prediction of lymph node metastasis based on the Japanese Society for Cancer of the Colon and Rectum guidelines. Assuming predicted probability of LNM is 20%.