Back to 2025 Abstracts
Novel Artificial Intelligence Clinical Decision Support System for Aesthetic Surgery
Berk B. Ozmen
*1, Nishant Singh
2, Kavach Shah
2, Ibrahim Berber
3, F Damanjit Singh
2, Eugene Pinsky
2, Nicholas R. Sinclair
1, Graham S. Schwarz
11Department of Plastic Surgery, Cleveland Clinic, Cleveland, OH; 2Department of Computer Science, Metropolitan College, Boston University, Boston, MA; 3Department of Computer Sciences, Case Western Reserve University, Cleveland, MA
Background: Aesthetic surgery demands the integration of specialized anatomical knowledge with refined clinical expertise to achieve optimal patient outcomes. Recent advances in artificial intelligence (AI) have transformed various aspects of healthcare delivery, yet no AI-powered clinical decision support systems have been developed specifically for aesthetic surgery. The aim of this study is to develop a novel AI retrieval-augmented generation system that provides evidence-based clinical decision support for aesthetic surgery.
Methods: We developed AURA (Aesthetic surgery Using Retrieval Augmentation), a retrieval-augmented generation system combining 6,546 full-text open-access papers published between January 2001 and September 2024 from PubMed Central combined with a commercially available large language model. The system was evaluated using 14 complex clinical questions across aesthetic surgery domains. Performance was assessed using multiple metrics: faithfulness to source documents, answer relevancy, G-Eval correctness, semantic quality (SEM score), and confidence level.
Results: The system demonstrated robust performance with average scores of 0.94 for faithfulness, 0.86 for answer relevancy, and 0.77 for correctness. Semantic evaluation showed average scores of 0.73 for SEM Score and 0.80 for SEM Max Similarity, with predominantly moderate confidence ratings. Performance was strongest for established techniques and safety considerations, with more modest results for emerging procedures.
Conclusion: We developed the first AI clinical decision support system for aesthetic surgery, named AURA. Our model effectively provides relevant and accurate information for aesthetic surgery clinical decision support. The model offers an efficient method for accessing evidence-based information for aesthetic surgery.
Back to 2025 Abstracts
