A classification of articular cartilage grafting and regeneration developed using an artificial intelligence narrative review

This is the authors' abstract. We don't add key points for this paper.

Introduction Cartilage grafting, in contrast to bone grafting, lacks a standardized classification system. We hypothesized that a robust classification system for cartilage grafts could be developed effectively with artificial intelligence (AI)-assistance. Objectives To synthesize properties identified through an AI-augmented narrative review and thereby generate a classification of cartilage grafting techniques. Methods Twenty-four cartilage-related search terms were identified by the authors based on their clinical experience. These terms were used to query 6 large language models. Each AI was prompted to list properties relevant to cartilage grafts, based on its determination of what published researchers have reported. The resulting properties were consolidated by the authors and then submitted to the AIs for suggestions to name the properties, using Greek-based roots to describe each property. The output was synthesized by the authors into a final 5-category system. Results A classification with 5 primary categories of cartilage graft properties was devised. The categories were as follows: providing load-bearing and force-dissipating capacity (termed "chondrodynamic"); providing a smooth and resilient surface ("chondrolissic"); filling the defect volume ("chondroplerotic"); integrating with defect margins ("chondrosyndetic"); and supporting cellular activity ("chondrobiotic"). Conclusions We present a novel 5-category classification system for cartilage grafting and regeneration developed through an AI-assisted narrative review and synthesis. This study may fill important gaps. While the risks of AI are real (especially source fabrication, also known as "hallucination"), with further validation and consensus-building, the proposed framework can be used in clinical and research settings and guide the future use of AI in knowledge aggregation and application.