An Integrated AI-Enabled System Using One Class Twin Cross Learning (OCT-X) for Early Gastric Cancer Detection

Xian-Xian LIU¹Yuanyuan WEI²Mingkun Xu^3*Yongze GUO⁴Hongwei ZHANG⁴Huicong DONG⁴Qun SONG⁵Qi Zhao⁶Wei Luo⁷Feng Tian⁸Juntao Gao⁹Simon FONG^1*

• ¹Department of Computer and Information Science, Faculty of Science and Technology, University of Macau, Taipa, China
• ²Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, California,, United States
• ³Guangdong Institute of Intelligence Science and Technology, Zhuhai, China
• ⁴Department of Gastroenterology, Affiliated Hospital of Hebei University of Engineering, Handan, China
• ⁵Institute of Artificial Intelligence, Chongqing Technology and Business University, Chongqing, China
• ⁶MoE Frontiers Science Center for Precision Oncology, University of Macau, Taipa, Macao, SAR China
• ⁷The director of the Institute of Clinical Medicine, The First People’s Hospital of Foshan, Guangzhou, China
• ⁸Hebei Key Laboratory of Medical Data Science, Institute of Biomedical Informatics, School of Medicine, Hebei University of Engineering, Handan, Hebei, China
• ⁹The Beijing National Research Center for Information Science and Technology (BNRist), Tsinghua University, Beijing, China

Early detection of gastric cancer, a leading cause of cancer-related mortality worldwide, remains significantly hampered by the limitations of current diagnostic technologies, resulting in high rates of misdiagnosis and missed diagnoses. To address these clinical challenges, we propose an integrated AI-enabled imaging system that synergizes advanced hardware and software technologies to optimize both speed and diagnostic accuracy. Central to this system is our newly developed One Class Twin Cross Learning (OCT-X) algorithm, which leverages a fast double-threshold grid search strategy (FDT-GS) and a patch-based deep fully convolutional network for precise lesion surveillance and classification in real-time. The hardware platform incorporates an all-in-one point-of-care testing (POCT) device, equipped with high-resolution imaging sensors, real-time data processing capabilities, and wireless connectivity, supported by the NI CompactDAQ system and LabVIEW software for seamless data acquisition and control. This integrated system achieved a diagnostic accuracy of 99.70%, outperforming existing state-of-the-art models by up to 4.47%, and demonstrated a 10% improvement in multirate adaptability, ensuring robust performance across varied imaging conditions and patients profiles. These results highlight the potential of the OCT-X algorithm and the integrated platform to enable more accurate, efficient, and non-invasive early detection of gastric cancer in point-of-care settings.