AI support for colonoscopy quality control using CNN and transformer architectures
BMC Gastroenterology · 09.08.2024 · 24 (1)
DOI: 10.1186/s12876-024-03354-0 ↗Kurz gefasst
Fachartikel in der Zeitschrift BMC Gastroenterology, online erschienen am 09.08.2024. Verfasst von Jian Chen und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Gastroenterology, Changshu Hospital Affiliated to Soochow University, Suzhou, 215500, China. Hauptthemen laut MEDLINE: Colonoscopy, Quality Control, Deep Learning. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Changshu City Medical and Health Science and Technology Plan Project, Health Informatics Key Support Discipline Funding of Suzhou City, Changshu City Science and Technology Plan Project.
Abstract (Originalsprache)
BACKGROUND: Construct deep learning models for colonoscopy quality control using different architectures and explore their decision-making mechanisms. METHODS: A total of 4,189 colonoscopy images were collected from two medical centers, covering different levels of bowel cleanliness, the presence of polyps, and the cecum. Using these data, eight pre-trained models based on CNN and Transformer architectures underwent transfer learning and fine-tuning. The models' performance was evaluated using metrics such as AUC, Precision, and F1 score. Perceptual hash functions were employed to detect image changes, enabling real-time monitoring of colonoscopy withdrawal speed. Model interpretability was analyzed using techniques such as Grad-CAM and SHAP. Finally, the best-performing model was converted to ONNX format and deployed on device terminals. RESULTS: The EfficientNetB2 model outperformed other architectures on the validation set, achieving an accuracy of 0.992. It surpassed models based on other CNN and Transformer architectures. The model's precision, recall, and F1 score were 0.991, 0.989, and 0.990, respectively. On the test set, the EfficientNetB2 model achieved an average AUC of 0.996, with a precision of 0.948 and a recall of 0.952. Interpretability analysis showed the specific image regions the model used for decision-making. The model was converted to ONNX format and deployed on device terminals, achieving an average inference speed of over 60 frames per second. CONCLUSIONS: The AI-assisted quality system, based on the EfficientNetB2 model, integrates four key quality control indicators for colonoscopy. This integration enables medical institutions to comprehensively manage and enhance these indicators using a single model, showcasing promising potential for clinical applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Gastroenterology, Changshu Hospital Affiliated to Soochow University, Suzhou, 215500, China.
- Department of Gastroenterology, Changshu Traditional Chinese Medicine Hospital (New District Hospital), Suzhou, 215500, China.
- Shanghai Haoxiong Education Technology Co., Ltd, Shanghai, 200434, China.
- Department of Information Engineering, Changshu Hospital Affiliated to Soochow University, Suzhou, 215500, China. kjxia@suda.edu.cn.
- Department of Gastroenterology, Changshu Hospital Affiliated to Soochow University, Suzhou, 215500, China. xxddocter@gmail.com.
Förderung
- Changshu City Medical and Health Science and Technology Plan Project
- Health Informatics Key Support Discipline Funding of Suzhou City
- Changshu City Science and Technology Plan Project
Zitieren
APA
Chen, J., Wang, G., Zhou, J., Zhang, Z., Ding, Y., Xia, K., & Xu, X. (2024). AI support for colonoscopy quality control using CNN and transformer architectures. BMC Gastroenterology, 24(1). https://doi.org/10.1186/s12876-024-03354-0
Vancouver
Chen J, Wang G, Zhou J, Zhang Z, Ding Y, Xia K, et al. AI support for colonoscopy quality control using CNN and transformer architectures. BMC Gastroenterology. 2024;24(1). doi:10.1186/s12876-024-03354-0
BibTeX
@article{Chen2024,
author = {Chen, Jian and Wang, Ganhong and Zhou, Jingjie and Zhang, Zihao and Ding, Yu and Xia, Kaijian and Xu, Xiaodan},
title = {AI support for colonoscopy quality control using CNN and transformer architectures},
journal = {BMC Gastroenterology},
year = {2024},
volume = {24},
number = {1},
doi = {10.1186/s12876-024-03354-0}
}
EUVIMED