Screen 1
Click the “View PDF” button or the Image for a high quality PDF file.
Click BACK in browser to return after viewing PDF.
ePosters can only be viewed by delegates with a ticket or BOPSS members with a paid sibscription
| Title | Artificial Intelligence on a Smartphone Versus the Ruler: Objective Measurement of Ptosis Severity |
| Number | 302 |
| Author | Christine Safieh |
| Principal Unit where research was undertaken | MID Yorkshire Teaching NHS Trust |
| Purpose | Accurate measurement of marginal reflex distance (MRD1) is essential in the assessment and management of ptosis. Traditional ruler-based measurements are widely used but may be subject to inter-observer variability and measurement error. Smartphone-based artificial intelligence (AI) tools may offer a rapid and objective alternative. This study aimed to compare AI-assisted smartphone measurement of MRD1 with conventional ruler-based clinical assessment. |
| Methods | A prospective observational study was conducted in patients presenting with upper eyelid ptosis. Standardised frontal photographs were captured using a smartphone camera under consistent clinical conditions. MRD1 was measured manually using a millimetre ruler by two independent clinicians. The same images were analysed using a smartphone-based AI facial landmark detection algorithm to automatically calculate MRD1. Agreement between AI-derived and clinician-measured MRD1 values was evaluated using Pearson correlation and Bland–Altman analysis. Measurement time and inter-observer variability were also recorded. |
| Results | Sixty-two patients (124 eyelids) were included in the analysis. AI-derived MRD1 measurements showed strong correlation with manual ruler measurements (r = 0.90, p < 0.001). Bland–Altman analysis demonstrated a mean difference of 0.06 mm with 95% limits of agreement between −0.41 mm and 0.53 mm. Inter-observer variability was reduced using the AI method compared with manual measurement. Mean assessment time was significantly shorter using the smartphone AI tool (1.8 ± 0.4 minutes) compared with ruler-based assessment (4.5 ± 0.8 minutes). |
| Conclusion | Smartphone-based AI measurement provides accurate and rapid quantification of ptosis severity compared with traditional ruler assessment. This approach may offer a simple and accessible tool for objective eyelid measurement in routine oculoplastic practice. |
Additional Authors
| First name | Last name | Hospital / Institution |
|---|
Screen 2
No ePosters found. Please select a year or use the search form
Screen 3
No ePosters found. Please select a year or use the search form
