Objective and automated facial palsy grading and outcome assessment after facial palsy reanimation surgery - A prospective observational study

Abstract

Background: Facial palsy (FP) is a widespread condition affecting over 3 million people annually, with a complex etiology requiring tailored, multidisciplinary management. Despite advancements, there remains a lack of reliable, automated tools for objective pre- and postoperative assessment, limiting progress in treatment optimization. This study introduces the AI Research Metrics Model (CAARISMA ® ARMM) to evaluate FP severity and outcomes following microsurgical gracilis muscle transfer.

Methods: We analyzed pre- and postoperative images of 20 FP patients using CAARISMA ® ARMM, which identifies 17 facial landmarks and evaluates 1,030 parameters. CAARISMA ® ARMM calculates three indices: Facial Youthfulness Index (FYI), Facial Aesthetic Index (FAI), and Skin Quality Index (SQI). All surgical procedures were performed by the senior author. Statistical analysis compared preoperative and postoperative scores using independent t-tests and Wilcoxon-Mann-Whitney tests, with significance set at p < 0.05.

Results: Significant improvements were observed in the FAI scores post-surgery (p < 0.001). In contrast, FYI and SQI scores did not show significant postoperative changes (p = 0.39 and p = 0.60, respectively). Significant gender differences emerged: females showed increased FYI scores postoperatively, while males exhibited a decline (p = 0.0065). Age-related variations were also significant, with younger patients showing improved SQI and older patients experiencing declines (p = 0.040).

Conclusion: The CAARISMA ® ARMM effectively captures aesthetic improvements post-reanimation. Gender and age significantly influence outcomes, underscoring the key role of personalized and adaptable assessment tools. Future studies should integrate dynamic assessments and validate the CAARISMA ® ARMM across additional patient populations. CAARISMA ® ARMM holds promise as a standardized tool in FP outcome evaluation.

Keywords: Artificial intelligence; Bell's palsy; Facial palsy; Facial paralysis; Facial reanimation; Smile restoration.