Research Progress on Artificial Intelligence-Assisted Dynamic Evaluation of Psoriasis：A narrative review

This narrative review focuses on psoriasis, a chronic recurrent inflammatory skin disease with multisystem comorbidities that significantly impairs patients' quality of life. The recurrent nature of the disease often imposes substantial psychological burden, accompanied by anxiety, depression, and social avoidance behaviors. Precise dynamic assessment throughout the entire diagnostic and therapeutic cycle is crucial for improving patient prognosis. Artificial intelligence technology, with its powerful image recognition and data analysis capabilities, provides a novel technical pathway for dynamic assessment of psoriasis. This review comprehensively summarizes research progress of AI technology in whole-process assessment of psoriasis before, during, and after treatment, analyzing application characteristics and performance advantages of different AI models in auxiliary disease diagnosis, severity scoring, treatment regimen optimization, relapse risk prediction, comorbidity screening, and psychological status assessment. Furthermore, we examine the core challenges currently faced, including lack of data standardization, insufficient diversity of population and subtype-specific data, unresolved ethical and privacy protection concerns, and barriers to interdisciplinary clinical integration. AI models such as convolutional neural networks and deep learning can effectively improve the accuracy of psoriasis diagnosis and the objectivity of severity assessment, and also exhibit promising application potential in treatment regimen screening, therapeutic effect simulation, prognostic monitoring, and psychological burden identification. Moreover, multimodal fusion and cross-technology integration have emerged as important development directions. To address the existing challenges, future research should focus on advancing the construction of standardized and diverse data systems, fostering interdisciplinary expertise, developing an integrated AI model covering diagnosis, assessment, treatment, prognosis, and psychological status monitoring, and strengthening multicenter clinical validation and the establishment of ethical norms. This review demonstrates that the deep integration of AI technology with dynamic assessment of psoriasis can provide strong support for precise and individualized long-term disease management, and holds important clinical value for optimizing psoriasis diagnosis and treatment regimens, reducing the disease burden, and improving patients' mental health.