Home / Journal / 2026

Smartphone-Based 3D Ocular Reconstruction for Automated Quantitative Blepharoptosis Evaluation

A smartphone app using a 10-second video to build a 3D model of the eye matched specialist hands-on ptosis measurements closely in 952 patients (1,904 eyes) across 5 hospitals, and diagnosed ptosis with 0.97 area-under-the-curve accuracy.

Yang Y, Han Y, Wang R, et al. · JAMA Ophthalmol, 2026 · PMID 42783353 · DOI 10.1001/jamaophthalmol.2026.4120

Overview

Measuring a droopy eyelid (ptosis) by hand is experience-dependent and can vary between examiners. The authors built and tested the Digital Rulers of Ptosis (DRP), a system that reconstructs the eye in 3D from a 10-second smartphone video to automatically measure lid position and muscle function, and compared it against same-day specialist exams across 5 hospitals in 2025.

Findings

  • Prospective study, 5 hospitals, 2025; participants 3 years and older.
  • 952 participants, 1,904 eyes; mean age 27.9 years (range 3-86); 51.8% female, 48.2% male.
  • 1,862 eyes analyzed for eyelid position (MRD1); 1,830 eyes analyzed for levator muscle function (LF), after quality control.
  • Agreement with specialist measurement: mean absolute error 0.52 mm (95% CI, 0.50-0.54 mm) for MRD1; 0.95 mm (95% CI, 0.91-0.98 mm) for levator function.
  • Diagnostic accuracy for ptosis: area under the curve (AUC) 0.97 (95% CI, 0.96-0.97).
  • Accuracy for identifying the surgical approach needed: AUC 0.98 (95% CI, 0.97-0.98).
  • Performance was consistent across sex, age and geographic region.
  • Among 206 surgical participants (248 eyes), 3-month recurrence rate was 7.3% (95% CI, 4.4%-11.2%; 18 of 248 eyes), within the lower range of previously reported rates.

What it means for a patient

  • A smartphone video, rather than only a specialist's hands-on measurement, can measure how droopy an eyelid is and how well its lifting muscle works, closely matching a specialist's numbers.
  • This tool correctly identified ptosis and predicted the likely surgical approach with high accuracy in this study.
  • Among patients who had surgery, about 7% had their ptosis come back within 3 months, a rate on the lower end of what has been reported before.
  • The authors state further studies are needed to show how this changes actual clinical decisions and outcomes, not just measurement accuracy.

Why this paper matters

It validates a smartphone-based measurement tool against specialist ptosis exams in nearly 1,000 patients across 5 hospitals, with high diagnostic accuracy for both detecting ptosis and choosing a surgical approach.

Terms

  • Blepharoptosis (ptosis): an eyelid that sits lower than normal, sometimes covering part of the pupil.
  • Marginal reflex distance 1 (MRD1): the distance from the light reflection in the pupil to the upper eyelid margin, a standard ptosis measurement.
  • Levator function (LF): how far the upper eyelid moves from looking down to looking up, reflecting how well the lifting muscle works.
  • Area under the curve (AUC): a score from 0.5 to 1.0 for how well a test tells two groups apart; 1.0 is perfect.
  • Mean absolute error: the average size of the difference between two measurement methods.

Newest eyelid surgery papers in JAMA Ophthalmol

From the PubMed eyelid surgery record, newest first.

2025
Human Papillomavirus and Thyroid Eye Disease
Garg I, Meyer BI, Gallo RA, Wester ST, Pelaez D · JAMA Ophthalmol · PMID 40272832
2024
Episodic Upper Eyelid Edema in an African American Patient
Akella SS, Srivatsan S, Lin A · JAMA Ophthalmol · PMID 38780928
2024
Older Woman With Proptosis, Ptosis, and Blurred Vision
Gibbons AB, Eberhart C, Li E · JAMA Ophthalmol · PMID 38270967
2023
Punctal Plug Granuloma After Blepharoplasty
Goodyear K, Pirakitikulr N, Rootman DB · JAMA Ophthalmol · PMID 37855838
2022
2022
Bilateral Upper Eyelid Ecchymosis and Edema in a 13-Month-Old
Sriram A, Rosenberg JB, Barmettler A · JAMA Ophthalmol · PMID 34882188