AI and Organ-on-a-Chip Technologies Showcased at International Microphysiological Systems Summit
At the technical showcase of the 2026 International Microphysiological Systems Alliance (iMPSS) Asia-Pacific Summit and the 9th Organoid and Organ-on-a-Chip Academic Conference, researchers demonstrated the integration of artificial intelligence with organ-on-a-chip technology. Experts said the combination of AI with organoids and organ-on-a-chip will shape a humanized, data-driven drug development model.
During the conference, Zhang Yujian, CEO of Suzhou Beilun Bio, presented the progress of the world's first personalized mRNA tumor vaccine clinical study integrating organ-on-a-chip and AI. The study was initiated in early 2026 by teams including Suzhou Beilun Bio, the Organ-on-a-Chip Research Institute of Southeast University, and Nanjing Drum Tower Hospital.
Zhang explained that the team constructs organ-on-a-chip models from pathological samples of non-small cell lung cancer patients and feeds experimental data into an AI model to identify tumor-specific antigens and design drug molecules. The molecules are then tested on the chips, with resulting data used to iterate the model. Because each individual's immune complexes differ, AI enables personalized drug molecule design and simulation of its behavior in the body, while organ chips provide high-quality, controlled data for model interpretation.
Gu Zhongze, director of the National Key Laboratory of Digital Medicine Engineering and the Organ-on-a-Chip Research Institute of Southeast University, said organ chips generate data unattainable from clinical trials and animal tests, which are fundamental for developing AI tools. However, integration with cell and animal test data via AI is needed to create a "digital virtual human" to predict drug efficacy. Ai Xiaoni, a professor at Peking University's School of Pharmaceutical Sciences, noted that combining AI with organ chips can accelerate drug screening, target discovery, and toxicity assessment, but the main challenge remains a lack of high-quality data, requiring standardized processes for rapid, scalable data production.
