Grading Trachoma with a Novel Modified Mobile Phone Camera

Abstract

Purpose: : Trachoma is the leading infectious cause of blindness worldwide and is a World Health Organization "priority eye disease". An important component of trachoma treatment is mass azithromycin distributions, which are delivered based on the prevalence of active trachoma. Normally, clinically active trachoma is diagnosed by non-ophthalmic personnel via conjunctival examination. Trachoma could also be diagnosed by conjunctival photography, though this is not routinely performed due to the cost and complexity of current cameras. In this study, we evaluated a simple-to-use, relatively inexpensive modified mobile phone camera. We report the agreement of trachoma grades assessed by clinical examination in the field with those assessed by review of photographs.

Methods: : During trachoma surveillance in 15 communities from an area of Ethiopia with hyperendemic trachoma, 2 Ophthalmologists each examined the right upper tarsal conjunctiva of a separate group of 99 individuals, and then photographed the conjunctiva with an iPhone® 4G S camera coupled to a custom attachment with a +25 diopter lens and an external light-emitting diode (LED) light source (Ocular CellScope, University of California, Berkeley). Grades of TF (follicular trachoma) and TI (intense inflammatory trachoma) were assigned according to the World Health Organization simplified grading system in the field. Each Ophthalmologist, masked to the original in-field grades, then re-graded the highest quality photograph from each of the 99 participants he originally graded in the field. To assess agreement between the in-field and photo grades, we calculated the kappa statistic for each Ophthalmologist.

Results: : In-field grades and photographs were performed for all 198 study participants. 196 (99%) participants had photographs that were of adequate quality for grading trachoma. When grading the presence of follicular trachoma (TF), agreement was good between the in-field and photo grades for both ophthalmologists (kappa = 0.79 [95% CI: 0.65-0.93] and 0.68 [95% CI: 0.50-0.83], respectively). Agreement was less robust for the grade of inflammatory trachoma (TI) (kappa = 0.59 [95% CI: 0.21-0.95] and 0.46 [95% CI: 0.24-0.66]).

Conclusions: : Trachoma diagnostic grades derived from photographs taken with a novel modified iPhone® camera seem to have adequate agreement with in-field trachoma grading. These results suggest that this relatively inexpensive technology may hold promise as a diagnostic and telemedicine tool for trachoma surveillance.

Keywords: trachoma • imaging/image analysis: clinical