GIST AI-organoid research: current challenges and future directions

This figure focuses on GISTAI-organoid research, and systematically presents the core challenges currently faced at the data level (lack of multimodal data, missing labels, insufficient dimensions and domain deviation), algorithm level (poor generalization ability and weak interpretation) and clinical translation level (insufficient evidence, drug susceptibility testing cycle of 2 to 4 weeks, and unclear regulatory approval path).

Meanwhile, it proposes the future directions such as multi-center collaboration to build high-quality datasets, development of interpretable AI (such as attention mechanism visualization), building a joint platform of "biopsy → organoid culture →AI real-time analysis → clinical decision making", conducting real-world research to accumulate evidence of endpoint indicators, and promoting scientific supervision and product registration.