Journal's Facebook pageDevelopment and Validation of a Few-Shot Rapid Screening Model for Gastrointestinal Cancers Using AGI Large Vision Models
-
School of Automation, Central South University
410083 Changsha, China
{lijiu, tjin}@csu.edu.cn, yinfangjie2023@126.com -
School of Automation, Beijing Information Technology Science and University
102206 Beijing, China
457706420@qq.com, chenwb@bistu.edu.cn (corresponding author) -
Department of Pathology, Beijing Integrated Traditional Chinese and Western Medicine Hospital
100039 Beijing, China
15201232918@163.com -
CAS Blue Bay Cloud Technology (Guangdong) Co., Ltd.
518001 Guangzhou, China
1004297233@qq.com -
Longgang District Maternity & Child Healthcare Hospital of Shenzhen City, Longgang Maternity and Child Institute of Shantou University Medical College
518172 Shenzhen, China
2570758402@qq.com -
Department of Pathology, Beijing Ditan Hospital, Capital Medical University
100015 Beijing, China
liuting1981_2005@126.com, zhouxg1980@126.com (corresponding author) -
Xiangjiang Laboratory
410205 Changsha, China
tjin@csu.edu.cn -
Department of Oncology, Chongqing University Three Gorges Hospital
404010 Chongqing, China
ming_ruijie@qq.edu.cn -
Department of Otolaryngology, Head & Neck Surgery, Peking University First Hospital
100034 Beijing, China
syandf@163.com -
Department of Respiratory and Critical Care Medicine, Tianjin Chest Hospital
300222 Tianjin, China
fengxueku@qq.com -
Richard Dimbleby Laboratory of Cancer Research, Randall Division and Division of Cancer and Pharmaceutical Sciences, King's College London
SE1 1UL London, UK
dan.7.wang@kcl.ac.uk, jinhaideng_kcl@163.com -
Guangzhou Baiyunshan Pharmaceutical Holding Co., Ltd. Baiyunshan Pharmaceutical General Factory/Guangdong Province Key Laboratory for Core Technology of Chemical Raw Materials and Pharmaceutical Formulations
510515 Guangzhou, China
jinhaideng_kcl@163.com (corresponding author)
Abstract
Existing deep learning models in digital pathology typically require extensive labeled data and show limited generalization across organs. In contrast, large vision models exhibit effective feature extraction capabilities, enabling pathological image analysis for gastrointestinal cancer with relatively small sample sizes. In this study, we developed a screening framework leveraging a large vision model for coarse-grained classification of gastric and colorectal tissues. The model was evaluated on multicenter cohorts and under limited-data conditions. Using labeled tiles from only 76 whole-slide images, the model achieved class-averaged sensitivity and precision of 0.9816 and 0.9808 on the internal test set, and 0.9161 and 0.9179 on the external test set. When trained with only 200 tiles per class from 20 whole-slide images, the model maintained comparable performance, achieving sensitivity and precision of 0.9548 and 0.9518. These findings suggest that the model has reliable performance across multicenter cohorts and potential applicability in clinical pathology workflows.
Key words
Deep Learning, Gastrointestinal Cancers, Histopathology, Unified Screening
Digital Object Identifier (DOI)
https://doi.org/10.2298/CSIS251130024L
Publication information
Volume 23, Issue 2 (April 2026)
Year of Publication: 2026
ISSN: 2406-1018 (Online)
Publisher: ComSIS Consortium
Full text
Available in PDF
Portable Document Format
How to cite
Liu, L., Yin, F., Yang, G., Li, Q., Li, S., Pan, T., Liu, T., Tang, J., Ming, R., Song, Y., Feng, X., Wang, D., Zhou, X., Chen, W., Deng, J.: Development and Validation of a Few-Shot Rapid Screening Model for Gastrointestinal Cancers Using AGI Large Vision Models. Computer Science and Information Systems, 23(2), 801–826 (2026). https://doi.org/10.2298/CSIS251130024L
