Rare diseases are characterized by low incidence rates, high heterogeneity, and significant genetic relevance, posing global challenges in clinical diagnosis and treatment, including delayed diagnosis and a scarcity of therapeutic options. Artificial intelligence (AI) technology offers novel solutions to address these challenges in the field of rare diseases. This paper explores the advancements in AI applications for rare diseases from two perspectives: auxiliary diagnosis and treatment decision-making. In terms of auxiliary diagnosis, AI can integrate superficial features, electronic health records, genomic data, and multi-modal data to achieve early and precise diagnosis. Regarding treatment decision-making, AI facilitates drug target discovery, drug repurposing, and the design of gene therapy vectors, thereby promoting the development and application of new treatments. Furthermore, this paper analyzes the challenges of AI in rare disease diagnosis and treatment concerning data, technical algorithms, and clinical application, and proposes future directions, including the construction of a collaborative data ecosystem, enhancement of algorithm interpretability, and improvement of regulatory frameworks.
Under the global background of the accelerated reconstruction of the smart healthcare ecosystem, artificial intelligence technology is deeply driving the transformation of the healthcare paradigm from experience-driven to data-knowledge dual-wheel driven. As a treasure of Chinese civilization, the core value of traditional Chinese medicine lies in the individualized diagnosis and treatment system based on "syndrome differentiation and treatment". The integration of multimodal diagnosis and treatment data and the construction of intelligent decision-making models will become the key path to break through the bottleneck of the modernization of traditional Chinese medicine. This research is based on the strategic orientation of "Healthy China 2030" and relies on the national science and technology major project of the team. It explores the establishment of a "three-stage four-dimensional" model of "data layer - knowledge layer - decision-making layer" and "feature extraction - relationship reasoning - dynamic correction - clinical verification" through a closed-loop verification mechanism of "human-machine collaboration - knowledge iteration", to promote the digital and intelligent transformation of traditional Chinese medicine.
