The search for cognitive solutions in health care is underway. Will optometry choose to utilize innovative technologies such as artificial intelligence (AI) to improve patients’ outcomes, or will it remain fearful and reactive when it comes to meaningful change?
The potential for improving medical services through the use of machine learning has been well documented.1 In November 2016, the online version of the Journal of the American Medical Association (JAMA) featured an article discussing an application of AI in the diagnosis of diabetic retinopathy (DR).2 While the eye care world has been fixated on telemedicine and online eye exams, robots are set to impact our 21st century world.
AI evolving around us
Many have heard Tesla CEO Elon Musk’s warnings about AI at the World Government Summit 2017 in Dubai.3 Living in New York City, I have seen the move from hailing a cab to using your smartphone to call for an Uber, Lyft, or Via. Within a few years, autonomous, self-driving electric cars may transform the taxi and car service industry.
Previously from Dr. Wong: Using social media to improve the way you connect
Once these vehicles are deployed in large numbers, Kyle Vogt, CEO and founder of Cruise Automation, predicts the third generation of self-driving cars will save millions of lives and accelerate the world’s transition to sustainable energy.4
The eventual success of self-driving cars and AI’s success in health care will depend on optimal hardware and software applications. For AI to succeed in eye care, it must be able to improve the current care we provide patients.
AI in optometry
Forum (Zeiss), Synergy (Topcon), and Spectralis (Heidelberg Engineering) software to monitor glaucoma progression has transformed and improved care of glaucoma patients. It provides an early and insightful look into the use of robotics and AI in optometry.
At the 2017 ARVO meeting in Baltimore, surgeons demonstrated the first successful use of a remote-controlled robotic system during retinal surgery in the human eye. A randomized clinical trial recruited six patients who had surgery with the robotic system, and six patients had surgery by traditional methods. Retinal micro-hemorrhage complications were reduced in the robotic system assisted group.5
Additionally, a study out of the Harker School in San Jose, CA, and the Byers Eye Institute at Stanford University Medical School independently developed and evaluated a data-driven, deep-learning algorithm as a diagnostic tool to detect DR.6
The algorithm analyzed fundus images and identified cases for medical referral. Researchers concluded a fully data-driven AI-based grading algorithm has the potential to screen fundus photography in diabetic patients. Diabetes affects more than 415 million people worldwide. The prudent use of AI could reduce the global loss of vision from DR.6,7
In addition to the applications of AI in DR treatment, deep-learning algorithms are being assessed for its value to the treatment of glaucoma progression.
Glaucoma progression analysis software is commonly used in optometric offices. Visulytix has developed a retinal AI platform called Pegasus that autonomously screens for glaucoma via assessment of the optic disc while simultaneously classifying the patient’s stage of DR.8
AI is likely to become commonplace over the next few years helping optometrists and ophthalmologists with clinical decision-making and reducing medical errors and variability in patient care.9
Future of optometry and AI
As technology continues to influence optometry, AI will continue to make transformational changes. The real test for optometry is two-fold. First, we must embrace innovations such as AI. Second, we must be objective in assessing and adopting AI in order for optometry to mature as a profession.
1. Research at Google. Healthcare. Available at: https://research.google.com/teams/brain/healthcare/. Accessed 1/9/18.
2. Gulshan V, Peng L, Coram M, Stumpe MC, Wu D, Narayanaswamy A, Venugopalan S, Widner K, Madams T, Cuadros J, Kim R, Raman R, Nelson P, Mega J, Webster D. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016;316(22):2402-2410.
3. CNBC. Tesla’s Musk warns about artificial intelligence. Available at: https://www.cnbc.com/video/2017/02/13/teslas-musk-warns-about-artificial-intelligence.html. Accessed 1/9/18.
4. Vogt K. How we built the first real self-driving car (really). Available at: https://medium.com/kylevogt/how-we-built-the-first-real-self-driving-car-really-bd17b0dbda55. Accessed 1/9/18.
5. The Association for Research in Vision and Ophthalmology (ARVO). First Use of Surgical Robot Inside the Human Eye. Available at: http://www.newswise.com/articles/view/673836/?sc=sphr&xy=10020710. Accessed 1/9/18.
6. Gargeya R, Leng T. Automated Identification of Diabetic Retinopathy Using Deep Learning. Ophthalmol. 2017 Jul; 124(7): 962-969.
7. International Diabetes Federation (IDF). IDF Diabetes Atlas - 8th edition. Available at: http://www.diabetesatlas.org/. Accessed 1/9/18.
8. Harper R, Reeves B. The sensitivity and specificity of direct ophthalmoloscopic optic disc assessment in screening for glaucoma: a multivariate analysis. Graefes Arch Clin Exp Ophthalmol. 2000 Dec;238(12):949-55.
9. Al-Aswad L. Artificial Intelligence, Big data, and Medical Analytics. EyetubeOD. Available at: http://eyetubeod.com/2017/09/artificial-intelligence-big-data-and-medical-analytics. Accessed 1/9/18.