Retrospective cohort study reveals upper eyelid blepharoplasty reduces perceived age by nearly five years, highlighting artificial intelligence as an objective aesthetic evaluation tool.
Upper eyelid blepharoplasty is a commonly performed procedure for periorbital rejuvenation; however, objective evaluation of its aesthetic outcomes remains limited. This study aimed to assess changes in perceived age following upper blepharoplasty using an artificial intelligence (AI)-based age estimation model. In this retrospective cohort study, patients who underwent isolated upper blepharoplasty between January 2022 and January 2025 were included. Standardized preoperative and postoperative (3-6 mo) facial photographs were analyzed using a multimodal AI system. Each image was evaluated 3 times with a standardized prompt, and mean values were recorded. The primary outcome was the change in perceived age. A total of 47 patients were included (mean chronological age: 48.77±7.74 y). The mean AI-estimated age decreased from 48.57±7.30 years preoperatively to 43.74±7.00 years postoperatively, corresponding to a mean reduction of 4.83 years (P<0.001; Cohen d=3.60). Preoperative AI estimates showed close agreement with chronological age (mean difference: -0.19 y), whereas postoperative estimates demonstrated a significant underestimation (-5.02 y; P<0.001). The mean absolute error increased from 1.77 years preoperatively to 5.15 years postoperatively. Bland-Altman analysis confirmed minimal bias before surgery and a systematic negative bias after surgery, indicating a shift toward younger age estimation. These findings suggest that upper eyelid blepharoplasty results in a measurable reduction in perceived age. AI-based facial analysis may provide an objective and reproducible tool for evaluating outcomes in aesthetic facial surgery.
No takes yet. Share an insight, caveat, or question.
Çetin et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: