Assessment of the quality and diagnostic accuracy of patient self-generated photographs used in oculoplastic virtual clinics.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BackgroundA barrier to video consultation in oculoplastics is visualising pathology too small for video resolution. Patient self-generated photographs, which is well-documented in teledermatology but not in oculoplastics, may provide a solution.PurposeTo assess the quality and diagnostic accuracy of patient self-generated photographs of presumed benign eyelid pathology.Methods300 patients with suspected benign lid lesions emailed self-captured photographs that were reviewed by two independent reviewers for quality, focussing, and lighting. The photographs were used in a virtual clinic alongside video consultation. Accuracy of virtual clinic diagnosis was measured by comparison to biopsy or in-person diagnosis where available.Results895 photographs were submitted. 84.5% of these were graded as excellent or adequate quality and there was no statistical difference in image quality based on patient age. Concordance between observers for quality of photographs, measured using Cohen's kappa, was k = 0.77 (P = <0.01). Virtual clinic diagnosis was verified in 236 (78.7%) patients by biopsy or in-person review. Correct diagnosis was made in 80.5%. 99.1% of benign lesions were correctly identified as low-risk, 50% of pre-malignant lesions were identified as suspicious, and 100% of malignant lesions were correctly identified as cancerous.ConclusionsPatients of all ages can provide adequate quality photographs that can aid video consultations for benign lid lesions. We still recommend in-person review where suspicious features are present, or photo quality is suboptimal.