BLOG: The ‘diabetization’ of glaucoma
Hovanesian is a faculty member at the UCLA Jules Stein Eye Institute and in private practice at Harvard Eye Associates in Laguna Hills, California.
Disclosures: Hovanesian reports he has a financial interest in MDbackline.
In diabetes, the most life-altering consequences can occur decades after diagnosis — like glaucoma.
Premonitory conditions such as insulin resistance and metabolic syndrome (called “prediabetes”) allow primary care providers to modify lifestyle, diet and other factors to stave off the actual disease. The benefits have been life saving for millions. In the cover story of this issue of Ocular Surgery News, we explore the potential for artificial intelligence similarly to transform the diagnosis of glaucoma.
Glaucoma AI imaging will work by identifying patterns among thousands of data points from OCTs of the retinal nerve fiber layer, predicting future loss of visual field. A single scan of an asymptomatic young patient in a primary care office may be enough to flag an at-risk nerve whose appearance would otherwise pass muster even by the most trained glaucoma specialist.
John A. Hovanesian
Predicting future functional impairment is what has been missing in glaucoma therapy since we first recognized the role of pressure in this disease 400 years ago. Basing treatment on risk factors such as IOP, cup-to-disc ratio, family history and corneal thickness puts us far behind the way we treat diabetes.
Another untapped source of valuable information in glaucoma management may be patient-reported outcomes. For several years, we have used MDbackline, a secure web service that annually contacts our patients with glaucoma who take drops, to inquire about their compliance, symptoms, costs of treatment and risk factor history. The information is condensed in a single summary “visual profile report” that goes into our EHR system. We have looked retrospectively at these reports and their ability to predict future glaucoma damage, and we have found some interesting correlations that we are preparing to publish. In other words, collecting structured patient-reported outcomes may be able to predict glaucoma progression in the same way that imaging studies can.
D.H. Lawrence is frequently quoted in medical training with the phrase, “The eye cannot see what the mind does not know.” Whether by artificial intelligence analysis of retinal nerve fiber layer thickness or by rigorous analysis of structured patient history, we have tools at hand that will inexpensively and effectively help us redefine the diagnosis of glaucoma and optimize its treatment. As with our transformed management of diabetes, we are likely to enter a new age of glaucoma management where we can recognize the true incidence of the disease with a single evaluation that can change a patient’s life.