The following is a summary of “Application of a deep learning system to detect papilledema on nonmydriatic ocular fundus photographs in an emergency department,” published in the November 2023 issue of Ophthalmology by Biousse et al. 

Researchers performed a retrospective study to test if a deep learning system could have helped emergency room(ER) doctors detect papilledema more accurately. 

In this retrospective secondary analysis, a testing dataset comprising 1608 photographs from 828 patients in the Fundus photography vs. Ophthalmoscopy Trial Outcomes in the Emergency Department(FOTO-ED) studies was used. These photographs were reclassified based on the deep learning system’s optic disc classification, which included “normal optic discs,” “papilledema,” and “other optic disc abnormalities.” The system’s performance was assessed by calculating the area under the curve (AUC), sensitivity, and specificity using a one-vs-rest approach, with comparisons made to expert neuro-ophthalmologists’ evaluations.

The  Brain and Optic Nerve Study Artificial Intelligence (BONSAI)-deep learning system exhibited successful discrimination between normal and abnormal optic discs, with an [(AUC 0.92 (95%CI, 0.90-0.93); sensitivity 75.6% (73.7%-77.5%) and specificity 89.6% (86.3%-92.8%)]. The system effectively differentiated papilledema from normal and other optic disc abnormalities, achieving an [(AUC 0.97 (0.95-0.99); sensitivity 84.0% (75.0%-92.6%) and specificity 98.9% (98.5%-99.4%)]. The deep learning system correctly identified six patients with papilledema in one eye despite missing it in the other eye.

The study found that a deep learning system accurately identified papilledema and normal optic discs in eye photos taken in the ER.

Source: ajo.com/article/S0002-9394(23)00456-7/fulltext