Few objective measures are available for assessing the success of facial rejuvenation after face lift surgery. Convolutional neural networks (CNNs) may be used for this type of measurement. The purpose of this investigation is to use artificial intelligence (AI) via CNNs to objectively classify patient photos by age before and after aesthetic surgery. Uniquely, men and patients undergoing deep plane face lifts were included.
A CNN (FaceX) was used for facial age recognition and age estimation. Patient photos were analyzed preoperatively, and at three (PO1) and 12 months (PO2) postoperatively. The study population included male and female patients who underwent facial rejuvenation at our institution from 2017 to 2021. Patient photos were collected with the same camera, distance, and lighting.
226 patients were analyzed with a mean true age of 62.2 (SD 6.7) years. The AI estimated the mean preoperative age to be 64.7 (SD 10.4) years. The AI was 96.0 % accurate. Across all subjects, a 3.5-year, 5 % reduction in age (p ≤ 0.001) was attributed at PO1, and a 1.7 year, 3 % age reduction (p = 0.034) at PO2. No single ancillary procedure or technique conferred more benefit than others. The 15 males had a 2.0 year, 4 % age reduction (p = 0.06) at PO1.
AI can be used to objectively measure the success of facelift surgery and compare outcomes among rhytidectomy techniques. Additionally, multiple, different approaches were effective with no single approach being superior. As AI continues to rapidly advance, more accurate models may be developed for multiple applications in facial plastic surgery.
Copyright © 2022. Published by Elsevier Inc.
