TruCurve: Mathematical polynomial analysis of eyelid contour for objective assessment in oculoplastic surgery
Author: Kate Reed
Base Hospital / Institution: Moorfields Eye Hospital
ePoster presentation
Abstract ID: 24-495
Purpose
We aimed to evaluate mathematical polynomial lid curvature for the development of an automated assessment tool.
Methods
A custom graphical tool was developed to assess of lid contour. Static images were imported into the custom application, and eyelid landmarks were manually annotated in individuals. Scaling was performed with an estimated white-to-white limbus length to be set at 11mm. The lid margin was outlined using 20 points along the curvature from medial and lateral canthus. The line of best fit along the 20 points was then created using polynomial regression using the TruCurve software to determine the line of best fit to the actual curve of the eyelid.
Results
Fifty subjects (100 eyes; 23 males and 27 females; mean age 29.4 ±8.7) were analysed. The TruCurve software showed good interrater reliability (ICC=0.9). Mean R2 values of polynomial regression analysis for upper lids using x1, x2, x4 and x6 were 0.138, 0.981, 0.993 and 0.992 respectively (p<0.001), post hoc analysis showed significant differences between x1 and other polynomials but no significant difference between x2,x4 or x6. There was a significant difference in r2 values between males and females for x1 and x2 (p=0.05 and 0.03) but no significant difference for x4 or x6 (p=0.82 and 0.35). Ethnicity only affected accuracy for x1 (p=0.003). Mean R2 values for lower lids modelled using x1, x2, x4 and x6 were 0.131, 0.785, 0.975 and 0.980 respectively (P<0.001) post hoc analysis showed no significant differences between x4 and x6. Gender only affected accuracy for x1 (p=0.002). Latin and Mixed Race ethnicities had significantly worse accuracy compared to Asians using x4 polynomial (p=0.035 and 0.013).
Conclusion
This tool aims to establish a new standard for describing eyelid morphology, enhancing the precision and reliability of measurements across different clinical and research settings. By offering a standardised method for describing eyelid shapes and provides a robust platform for future research.
Additional Authors
| First name | Last name | Base Hospital / Institution |
|---|---|---|
| Jimmy | Uddin | Moorfields Eye Hospital |
| Jameel | Mushtaq | Moorfields Eye Hospital |
| Mohsan | Malik | Moorfields Eye Hospital |
| Michael | Milad | University of Oxford |
| Muhammad-Ameen | Kidiya | Queen Mary University of London |
