Deep learning able to identify glaucoma, study shows
As it turned out, the deep learning system performed quite well. The sensitivity and specificity of the system was 96.2 percent and 97.7 percent, respectively.
The most common cause for false positives and false negatives by both the deep learning system and the graders was high or pathologic myopia.
The authors state that these results show a propensity to enhance our current glaucoma screening modalities through time and cost efficiencies.
What this means
The potential implications of this concept are far-reaching.
One is increasingly less likely to have a conversation regarding the diagnosis of any disease without mention of technology. The eye and visual system are no strangers to this notion.
I would argue that the advent and subsequent bettering of technology has made our clinical lives easier—still challenging, but much easier.
The effective use of normative databases and progression algorithms with respect to spectral-domain optical coherence tomography (OCT) has given us highly reliable objective data which ODs can use as metrics for therapeutic efficacy (or to help ensure that no therapy is needed).
Deep learning systems will expand the horizon of how technology can aid in the diagnosis and staging of disease states. What an incredible time it is for ODs to be practicing.
It should be noted, however, that it is likely no mundane detail that the authors of this paper use the word “screening” in their description of potential applications for this deep-learning system.
The definitive diagnosis is the clinician’s to make.
One implication here is that clinicians will continue to make decisions more so than gather data. In the meantime,
I, for one, salute our new computer overlords.
1. Liu H, Li L, Wormstone IM, Qiao C, Zhang C, Liu P, Li S, Wang H, Mou D, Pang R, Yang D, Jiang L, Chen Y, Hu M, Xu Y, Kang H, Ji X, Chang R, Tham C, Cheung C, Ting DSW, Wong TY, Wang Z, Weinreb RN, Xu M, Wang N. Development and Validation of a Deep Learning System to Detect Glaucomatous Optic Neuropathy Using Fundus Photographs. JAMA Ophthalmol. 2019 Sep 12. doi: 10.1001/jamaophthalmol.2019.3501. [Epub ahead of print]