AAFPRS016 - Browlift Efficacy: Correlation of Artificial Intelligence with Excursion Measurements
Wednesday, April 27, 2022
1:30 PM – 1:35 PM CT
Location: Landmark D
Hebel, Nathan B.S.; Callori, Steven M.P.H.; Boonipat, Thanapoom M.D.; Lin, Jason M.D.; Uldis, Bite M.D.
Mayo Clinic Alix School of Medicine (Hebel, Callori)
Mayo Clinic Department of Plastic Surgery (Boonipat, Bite)
St Louis University Department of Plastic Surgery
Medical Student Mayo Clinic Alix School of Medicine Rochester, Minnesota
Introduction: Post-operative analysis is largely subjective. We correlated manual browlift measurements with artificial intelligence emotional analysis to assess browlift efficacy in an objective form. Methods: 52 Mayo Clinic browlift and blepharoplasty patients were identified between January of 2006 and December of 2018. Pre- and post-operative brow elevation was measured at the medial tip, mid pupillary line, and highest lateral point while being standardized off of the iris diameter (12 mm). The measurements were correlated to emotion analysis done using the FaceReader artificial intelligence software. Results: Post-operative browlift measurements showed statistically significant elevation of between 2.3-3.7mm (p<0.05) at each point relative to pre-operative measurements. The artificial intelligence showed decreased perceived sadness (14.3% to 7.4%) and anger (13.6% to 5.7%), with increased happiness (1.77% to 9.9%) (all p<0.05); there was no change in surprise emotion. Amongst those with reduced anger, it was shown that every 1mm change in central brow position of each brow resulted in approximately 7% decrease in anger expression. Conclusion: This study demonstrates validity and proof of concept for the use of artificial intelligence in assessing surgical outcomes. We observed our surgical intervention resulting in ideal emotional alteration of emotions without significantly increasing the appearance of surprise; a common worry expressed by patients. As well, these findings correlated to the manual measurements demonstrated that even modest changes in brow position resulted in dramatic alterations in artificial intelligence observed emotion. Future implementation of this work would be finding the optimal brow elevation for the desired emotional alterations.