EyeWorld/ASCRS reporting from AAO 2019 in San Francisco, Tuesday, October 15, 2019
Tuesday’s programming at AAO included symposia on a number of topics, including retina, endophthalmitis, interesting cases, and more.
Big data, technology, and gender equality in ophthalmology
A symposium Tuesday morning took a look at the “Millennial Movement: How Gender Equality, Big Data and Technology Will Be Embraced by the Young Ophthalmologist.”
Peter Karth, MD, Eugene, Oregon, said that there are currently 46 FDA approvals for artificial intelligence (AI) technologies in healthcare—two of which are in ophthalmology. Most AI work in ophthalmology is currently focused on diabetic retinopathy, glaucoma, and macular degeneration, 1) because of the ability to get large data sets from these conditions and 2) because of the impact that can be made, Dr. Karth said.
Dr. Karth provided an overview of several papers that show the utility of AI in these categories and how it is also allowing for analysis of things that even surpass human detection. Dr. Karth said he believes we will soon live in a world where algorithms detect, classify, and predict progression and treatment response of major eye conditions as well as or better than highly trained physicians—and more economically. When these systems get FDA approval, that’s a different timeline, Dr. Karth said.
There are pitfalls in AI, such as the black box problem. Physicians, Dr. Karth said, need to demand that algorithms they’re going to use are interrogated; the algorithm needs to tell us what it is looking at. This can be costly and timely, but it’s important that it be required from the FDA and technology partners, Dr. Karth said.
Input data is another pitfall. There needs to be large amounts of data with multiple grades for each image. What is the population of the training data? Is it diverse? Is there enough? These are questions we need to ask, Dr. Karth said. He also said the gold standard for algorithm quality and equality are those trained with board-certified ophthalmologists with ground truth disagreement adjudication by those who are fellowship trained.
Brian Stagg, MD, Salt Lake City, provided the audience with advice on how to get to the truth in the world of big data. Big data does not equal or guarantee truth in all cases, he said, nor does it ensure clinical relevance.
How do we figure out the truth? Dr. Stagg said to look at the sample, variables, and analysis. “This means we have to look at the methods,” he said. For the sample, evaluate the study population, the data source, inclusion/exclusion criteria, and how the study sample influences the findings. For variables, he said to ask yourself what the outcome variable and key predictor variable are and how they are measured. What are the variable definitions? How did the variables themselves influence the findings? Finally, Dr. Stagg acknowledged that analysis can get complicated, but it can be simplified by asking what question the investigators are asking and what is being adjusted for. “It is the responsibility of the reader to understand limitations and ask how they influences findings,” Dr. Stagg said. “How does it influence this patient sitting in front of me?”
Andrea Tooley, MD, New York, and Marie Louise Rasmussen, MD, PhD, Copenhagen, Denmark, discussed different perspectives on millennials—those in the West vs. East (Europe). Dr. Tooley said she thinks millennials in the U.S. get a bad reputation. “I think millennials are asking for a sense of purpose,” she said, adding that they want to feel engaged, part of a bigger community. “We are also hustlers. It’s common for millennials to have an alternate way of making money.” Dr. Tooley covered the influence technology has on millennials and how they use it to enhance their communication and work. She also said that millennials seek balance, prioritizing work-life balance.
Dr. Rasmussen said the discussion about millennials in Europe is quite different. “Millennials is not a topic in northern Europe,” she said, explaining that she had to Google the term when broached with this topic. In Scandinavia where she lives, Dr. Rasmussen said paternity leave is more of a discussion point today because maternity leave has decades-long acceptance. Young ophthalmologists in Europe, to Dr. Rasmussen, are empowered, creating new resources, and are doers. “They don’t have an idea, they execute it,” she said. There is also increasing subspecialization in young European ophthalmologists, Dr. Rasmussen said.
Other presentations included topics on tele-ophthalmology/tele-education and gender bias/the importance of diversity in the field of ophthalmology.
Editors’ note: Dr. Karth has financial interests with Google and Verily. Drs. Stagg, Tooley, and Rasmussen do not have financial interests related to their comments.
‘Delivery Therapeutics to the Posterior Ocular Segment’
A Tuesday morning symposium, co-sponsored by ARVO focused on a number of new technologies in development.
First, Marco Zarbin, MD, PhD, Newark, New Jersey, presented on treating blindness at the nanoscale.
Dr. Zarbin highlighted three different nanotechnologies: optogenetics, photoswitch, and quantum dots. These are aimed at bypassing degenerate photoreceptors to restore sight, he said.
The rationale behind ophthalmic nanoscale technology, Dr. Zarbin said, is to have an alternative to retinal prosthesis for ganglion, bipolar, or residual cone cell stimulation. Potential advantages are the higher spatial resolution and minimally invasive surgery, he said.
Optogenetics transfect retinal ganglion cells or bipolar cells with DNA coding light-activated ion channels, Dr. Zarbin said. Transfected cells initiate a synaptic signal when exposed to light and restore vision in patients with photoreceptor-related blindness. Dr. Zarbin noted channelrhodopsin, which he said is in clinical trial now.
Meanwhile, photoswitch features light-induced trans-cis isomerization, which causes a shape change in the photoswitch molecule. The change in molecule shape causes activation of native voltage-gated ion channels in neurons. The activation of the voltage-gated ion channels then cause a synaptic transmission, Dr. Zarbin said. The DAD molecule has been extensively studied with this, he added, which selectively photosensitizes retinae with photoreceptor degeneration.
Quantum dots were the third technology Dr. Zarbin highlighted. A quantum dot is a nanoparticle that behaves as a semiconductor, he said. Quantum dots exposed to light generate electric current or dipole moment and then voltage-gated ion channel activation in adjacent retinal neurons which then causes synaptic transmission.
All three of these nanotechnologies are delivered by intravitreal injection, with optogenetics also having a subretinal injection delivery system.
Also during the session, Debra Goldstein, MD, Chicago, presented on a microneedle drug delivery for uveitis. This is essentially using a tiny needle to go through the sclera to get drug into the suprachoroidal space, she said specifically mentioning the Clearside microinjector.
Animal data shows that drug localizes better to the choroid and outer retina with suprachoroidal injection compared to intravitreal injection, Dr. Goldstein said.
She also added that animal data shows that triamcinolone remains in ocular tissues for at least 90 days in rabbit eyes. A porcine model also showed interesting results with triamcinolone, comparing suprachoroidal and intravitreal injections.
Dr. Goldstein discussed the Phase 1/2 open-label trial of suprachoroidal triamcinolone in patients with uveitis, which was a small, three-site study, and the PEACHTREE Phase 3, randomized, controlled, double-masked, multicenter trial, which was a larger study.
She said that data suggests that microneedle delivery for the suprachoroidal space is a reasonable approach for treatment of uveitis, and at this time, a NDA has been submitted to the FDA. Further data regarding triamcinolone stability has been requested, she said, and there is still an unanswered question of whether a lower dose of triamcinolone would be as effective (in the trials, 4 mg doses of triamcinolone was used).
Editors’ note: Dr. Zarbin has no relevant financial interests. Dr. Goldstein has relevant financial interests with Clearside.
Endophthalmitis diagnosis and management
Russell Van Gelder, MD, PhD, Seattle, began a Tuesday morning symposium by discussing the current status of diagnostic modalities for endophthalmitis. He first described endophthalmitis as a purulent inflammation of the intraocular fluids, usually due to infection. It’s most commonly post-surgical but may also be endogenous or sterile, he said. The incidence in post-cataract endophthalmitis varies between 0.03% and 0.1% in large studies, he added.
Dr. Van Gelder said that molecular diagnostic techniques can offer improved sensitivity for detection of pathogenic organisms in postoperative endophthalmitis. He added that deep sequencing techniques offer further advantages, including strain-type information and antibiotic resistance profiles. Additionally, deep sequencing techniques have revealed unexpected viruses in culture-negative and culture-positive cases of endophthalmitis, he said, adding that technologic advances will make next generation sequencing more accessible in the near future.
Editors’ note: Dr. Van Gelder has no relevant financial interests.
Complex cases in ‘Grand Rounds’
Several interesting, complex cases were presented during the “Grand Rounds” symposium Tuesday morning.
One by Amy Yuan, MD, Boston, was of a 23-year-old who had chronic, progressive vision loss. He was born at 7 months gestation and was seemingly healthy with good visual acuity until 7 years old. At 7, he started having relapsing fevers with episodes that would last 2–3 hours; they were self-resolving. At age 15, he started losing vision. He was diagnosed with bilateral panuveitis and showed no improvement with systemic intravitreal steroid implants. Vision loss progressed. His history also reported an inability to produce tooth enamel, headaches, and inability to produce sweat. His parents were first cousins and there was no ocular or autoimmune family history, but there were cases of renal failure on the maternal side. His workup revealed posterior uveitis, optic nerve edema, systemic amyloidosis, splenomegaly, but there were no autoimmune, infectious, or neoplastic etiologies.
Genetic testing for this patient was repeated for known autoimmune diseases panels and for an ectodermal dysplasia gene panel; these were normal. A whole genetic workup revealed de novo heterozygous ALPK1. ALPK1’s function is poorly understood, but it is expressed in most tissues and highly expressed in retinal pigment epithelium, photoreceptors, and myoepithelial cells of sweat glands. It has a possible role in cilia formation, cell polarity, and cell cycle regulation.
Earlier this year, Dr. Yuan said, a paper was published describing 15 patients from five unrelated families who exhibited the same symptoms as this patient and heterozygous ALPK1. The paper coined the term ROSAH (Retinal dystrophy, Optic nerve edema, Splenomegaly, Anhidrosis, and migraine Headache) syndrome.
While the mystery of this patient’s diagnosis was solved, Dr. Yuan said the question of the pathophysiology and treatment options remains to be elucidated.
Editors’ note: Dr. Yuan does not have financial interests related to her comments.
EyeWorld Onsite is a digital publication of the American Society of Cataract and Refractive Surgery and the American Society of Ophthalmic Administrators.
Medical editors: Eric Donnenfeld, MD, chief medical editor; Rosa Braga-Mele, MD, cataract editor; Clara Chan, MD, cornea editor; Nathan Radcliffe, MD, glaucoma editor; Vance Thompson, MD, refractive editor
For sponsorship opportunities or membership information, contact:
ASCRS•ASOA • 4000 Legato Rd., Suite 700 • Fairfax, VA 22033 • Phone: 703-591-2220 • Fax: 703-591-0614
Opinions expressed in EyeWorld Onsite do not necessarily reflect those of ASCRS•ASOA. Mention of products or services does not constitute an endorsement by ASCRS•ASOA.
Copyright 2019, EyeWorld News Service, a division of ASCRSMedia. All rights reserved. unsubscribe