Pediatric cataract surgery presents challenges for surgeons due to the small size of the child’s eye. Low scleral rigidity also makes it difficult to maintain the chamber during surgery.

Opacification of the crystalline lens can lead to vision loss and an impaired quality of life for any patient. In 

, this can be a frightening experience for the child and treatment will involve both patient and parents.

“Pediatric cataract surgery is different from adult cataract surgery, mainly due to the small size of the child’s eye, with an axial length of 16.4 mm instead of 24 mm on average in adults, and the anterior chamber is only about 2 mm deep,” said Hee-Jung Park, MD, MPH, assistant professor of ophthalmology, Zanvyl Krieger Children’s Eye Center, Wilmer Eye Institute, Johns Hopkins University School of Medicine, Wilmer Eye Institute, Baltimore.

The low scleral rigidity makes it difficult to maintain the chamber during surgery. 

 the sclera is about four times more pliable and has only half the tensile strength of the adult sclera.”

To date, congenital cataract has been one of the most challenging surgeries I have performed. Two of my cases discussed here were both challenging and difficult.

However, I have found that microinstruments (MicroSurgical Technology [MST]) deliver a confident ability to perform a smoother, more-controlled surgery.

David Yorston, FRCOphth, Moorfield’s Eye Hospital, London, suggests that the exact cause may be unknown for many cases of childhood cataract, but there can often be associated systemic condition, such as:

Prenatal (intrauterine) infection, e.g., rubella, cytomegalovirus, syphilis.

Prenatal (intrauterine) drug exposure, e.g., corticosteroids, vitamin A.

Prenatal (intrauterine) ionizing radiation, e.g., x-rays.

Prenatal/perlnatal metabolic disorder, e.g., maternal diabetes.

Hereditary (isolated - without associated eye or systemic disorder), e.g., autosomal dominant inheritance.

Hereditary with associated systemic disorder or multisystem syndrome.

Chromosomal, e.g., Down’s syndrome (trisomy 21), Turner’s syndrome.

Skeletal disease or muscle disorder, e.g., Stickler syndrome, Myotonic dystrophy.

Central nervous system disorder, e.g., Norrie’s disease.

Renal disease, e.g., Lowe’s syndrome, Alport’s syndrome.

Mandibulo-facial disorder, e.g., Nance-Horan cataract-dental syndrome.

Dermatological disorder, e.g., congenital icthyosis, Incontinentia pigmenti.

Where unilateral cataracts are not inherited or associated with a systemic disease, they are usually the result of local dysgenesis and may be associated with other ocular dysgenesis, such as persistent fetal vasculature (PFV), posterior lenticonus, or lentiglobus.

A 6-year-old boy presented to me with unilateral congenital cataract with a refraction of + 8 D and a preoperative best-corrected distance visual acuity (CDVA) of 0.5 logMAR.

After having been sent away from two other clinics, the child’s parents approached me looking for a solution to their child’s vision problems. Cataract surgery 

 is challenging because of added issues, such as increased scleral elasticity, thicker corneas, and eye rubbing.

Adding to the challenge, during surgery, children placed under general anesthetic tend to turn up their eye globe which makes the anterior segment of the eye squashed for space. With limited space, the surgery csn be difficult.

Postoperatively, other issues such as reduced compliance with activity restriction, and the effect of postoperative astigmatism on amblyopia also must be considered.

IOL selection was also a challenge because the child was hyperopic. The choice of lens power is paramount to visual rehabilitation and a lens that leaves a blurred retinal image should be avoided.

The choice of IOL power should be individualized based on the child’s need and refractive status of the other eye in unilateral cases.

In recent years, acrylic IOLs have gained popularity over polymethyl methacrylate (PMMA) IOLs, which had remained the IOL of choice for many years.

In children, hydrophobic acrylic IOLs are considered better than PMMA IOLs in terms of greater biocompatibility and smaller incision size with use of foldable design, with late onset and a lower rate of posterior capsule opacification (PCO) formation.

Hydrophobic acrylic IOLs are used by 93% of pediatric cataract surgeons.

In children with uveitic cataracts, decreased postoperative inflammation has been reported with the use of heparin-surface-coated PMMA IOLs.

I targeted this case as slightly hyperopic, matching the refraction of the fellow eye. The IOL (IC-8, AcuFocus) had a power of + 27.5 D, and a target refraction of +1.32 D, using the Hoffer Q formula. An iris hook was used to gently open the iris due to posterior synechia and pupil dysgenisia.

Microforceps had to be used through different side ports. A forceps (23-gauge Seibel Capsulorhexis Forceps, MicroSurgical Technology) was used for opening of the capsule in the shallow anterior chamber (Figure 1).

This instrument is designed to minimize ophthalmic viscosurgical device (OVD) loss to maintain stable anterior chamber, as well as provide greater visibility and control of capsulorhexis, especially in eyes with a shallow anterior chamber.

As the shaft is stable, the branches can be moved to allow optimal control over the capsulorhexis.

Surgery was successful, and postoperative results are illustrated in Table 1.
	At the 12-week postoperative visit, manifest refraction was measured as +4.00/–3.00 × 43, with a corrected visual acuity of 0.15 logMAR at distance and 0.1 logMAR at near.

With the last visit in October, the patient and his parents are satisfied with the results and refraction has decreased to +3.00/–2.50 × 52 and a corrected visual acuity of 0.2 logMAR at distance.

The surgery was performed in 2016, and the follow-up was 18 months postoperatively.

A 5-year-old boy had a unilateral congenital cataract. Like the first child, he had been turned away from a clinic and a university for treatment. He had uveitis with pupillary membrane. The whole pupil was clogged with a dense fibrin membrane.

The boy’s vision had been decreased to hand movements from 0.5 logMAR over the previous 24 months.

During surgery (Figure 2), I used the 23-gauge, curved scissors (Hoffman/Ahmed horizontal curved scissors, MicroSurgical Technology) (Figure 3) to cut the pupillary membrane and inserted a Malyugin Ring 2.0, which was easy to use, expanding the pupil up to 6 mm and protecting the iris from damage.

I proceeded with the irrigation port and performed the anterior capsulotomy with the capsulotomy system (Zepto, Mynosys) to open the lens capsule and perform the cataract surgery.

The IOL (CT Lucia 211P, Carl Zeiss Meditec) with a power of + 28 D was implanted, with a target refraction of +0.35 D using Barrett formula.

For this surgery, I chose this IOL type because it is hydrophobic acrylic IOL with imbedded heparin surface coating to prevent further inflammation. With the capsulotomy system, an element surrounded by a silicone plastic shell is attached to a suction tube, and introduced into the eye via a tiny incision (2.2 to 2.4 mm).

As Mark Packer, MD, suggests, the silicone shell squeezes down and there is a push rod inside the silicone sleeve used to expand the ring to a circle, and nitinol retains it back to a circle. Pulling the push rod out, you retract it and turn on the suction and the whole thing sucks itself down onto the anterior capsule.

The capsulotomy system is fast and easy to use and offers a precise capsulotomy opening with a higher stability compared with manual capsulotomy.
	In the anterior chamber, I carefully grabbed the pupillary membrane in the center and lifted it.

Then I made a hole at the iris membrane margin and filled viscoelastic behind the pupillary membrane and went in with the instrument to loosen any deposits (synechiae) behind the membrane, then filled it up again with viscoelastic so the membrane comes up in the anterior chamber.

From the left side port, I entered again with 23-gauge, micro-holding forceps (Figure 4, Page 14), lifted up the membrane, and on the other side, I came in with the curved scissors (Hoffman/Ahmed), which are exchangeable.

When dealing with complicated cases, I find it much easier to use the MST handle, which is compatible with all exchangeable single-use heads that can be opened from the packaging when needed.

With the exchangeable 360 handle (MST), the direction of the instrument can be turned, so I was able to cut around the whole pupil just by turning the scissors in the right direction.
	This benefit allowed flexibility and ease of use intraprocedure. The development of fine-gauge instruments, especially the scissors and forceps, helps in chamber stability.

At one day postoperatively, uncorrected vision was measured as 0.5 logMAR. At postoperative day four, best-corrected visual acuity was 0.2 logMAR. The child has chronic uveitis, for which he is prescribed topical steroid treatment, but overall his vision has vastly improved from hand movements before surgery. Therefore, his quality of life has improved.

At three months after surgery, the uveitis was under control with topical dexamethasone drops, uncorrected visual acuity is 0.4 logMAR which is improved to 0.2 logMAR with correction.

The child’s mother has reported being very happy with the surgical outcomes.

Possible complications or risks postoperatively include glaucoma, retinal detachment, infection and the need for more surgeries. In my opinion, the use of single-use instruments may offer surgeons the potential to reduce some of these risks, especially in the pediatric cases.

Both cases are similar in that both were very small eyes in children who were close in age.

Pediatric cataract surgery is a challenge because the eyes are small and there is less space to maneuver. This is where precise, fine instruments are required to access the anterior chamber.

For instance, in cases with uveitis with pupillary membrane you need a forceps intraprocedure to hold the membrane on one side and then the curved scissors in the other hand to open the membrane. It is easy to turn the head of the scissors instead of having to turn the whole instrument in your hand.

Such pediatric cases with a complete membrane in the pupil area are rare, and consequently, I do not hold a set of reusable surgical instruments. As a result, disposable instruments were used for these two surgeries.

With disposable instruments, I am certain to have a sterile instrument and scissors with a sharp blade as well as single-use forceps which will hold tight and not be damaged. Being able to easily exchange just the head of the instruments is ideal as I do not require a whole set of instruments.

There is a higher risk of damage with traditional instruments as they are frequently re-sterilized and handled more often than disposable instrumentation.

With a wider selection of disposable tips for microinstruments available, there is a greater choice. It also helps to know you can carry out what you are planning with the new instruments. Single-use promises they will not be broken or damaged from re-use.

So, I find disposable instruments to be more practical and safer

Dr. Kretz is affiliated with Augentagesklinik Rheine, Germany. Dr. Kretz did not indicate proprietary interests in the subject matter.

1. Meszaros L. Congenital cataract still a challenge. 

. 2012. http://www.ophthalmologytimes.com/modern-medicine-now/congenital-cataract-still-challenge

2. Yorston D, Wood M, Foster A. Results of cataract surgery in young children in East Africa. 

3. Heidar K, et al. Cataracts in Children, Congenital and Acquired. American Academy of Ophthalmology. 2017. http://eyewiki.aao.org/Cataracts_in_Children,_Congenital_and_Acquired

4. Medsinge A, Nischal KK. Pediatric cataract: challenges and future directions. 

5. Wilson ME, Trivedi RH. Choice of intraocular lens for pediatric cataract surgery: survey of AAPOS members. 

6. Ram J, Jain VK, Agarwal A, Kumar J. Hydrophobic acrylic versus polymethyl methacrylate intraocular lens implantation following cataract surgery in the first year of life. 

7. Rowe NA, Biswas S, Lloyd IC. Primary IOL implantation in children: a risk analysis of foldable acrylic v PMMA lenses. 

8. Basti S, Aasuri MK, Reddy MK, Preetam P, Reddy S, Gupta S, et al. Heparin-surface-modified intraocular lenses in pediatric cataract surgery: prospective randomized study. 

9. Malyugin B. Managing small pupils: A surgeon’s step-wise approach to treatment. 

10. Stephenson M. Devices Focus, New Devices in Cataract Surgery. Eye World. 2018. https://www.eyeworld.org/new-devices-cataract-surgery

11. Kumar Khokhar S, Pillay G, Agarwal E, Mahabir M. Innovations in pediatric cataract surgery. 

Articles

Microinstrumentation provides surgeons increased control

Managing unique challenges of pediatric congenital cataract

With new agents approved and more new drugs moving through the pipeline, there is plenty of innovation occurring for the treatment of glaucoma.

The long drought in new glaucoma medications is over. After more than 20 years without a single new glaucoma eye drop, several therapeutic agents have been approved in recent months, with other novel drugs moving through clinical trials.

“There has been a lot of innovation in glaucoma surgery with MIGS, but also a lot of innovation with glaucoma medical therapy,” said Richard L. Lindstrom, MD, founder and attending surgeon, 

, and adjunct clinical professor emeritus of ophthalmology, University of Minnesota.

The most recent approval in March 2019 is a combination of netarsudil and latanoprost (Rocklatan) from Aerie Pharmaceuticals. Pooled data from the phase III Mercury trials showed the once-daily combination is more effective at lowering IOP than either of its ingredients used as a single agent. The combination is being pitched as a single-product alternative to multiple eye drops, which could improve adherence as well as therapeutic effect.

 had its first FDA approval in 2018 with netarsudil (Rhopressa), a Rho kinase inhibitor, or ROCK inhibiter, that works differently from other currently approved classes of glaucoma agents. Netarsudil reduces aqueous humor production, increases uveoscleral outflow and increases trabecular meshwork outflow.

Another recent introduction is latanoprostene bunod ophthalmic solution (Vyzulta, 

), a prostaglandin analogue that also releases nitric oxide. The single molecule is metabolized into two active moieties, latanoprost acid and nitric oxide. The combination facilitates aqueous humor outflow through both the uveoscleral and trabecular meshwork pathways.

Nitric oxide is an endogenous signaling molecule that increases the permeability of the trabecular meshwork to enhance aqueous humor outflow. It activates the soluble guanylate cyclase-cGMP cascade to inhibit Rho kinase and lower intracellular calcium levels to mediate cell relaxation in the trabecular meshwork to improve outflow.

Also new is the first (benzalkonium chloride) BAK-free formulation of latanoprost drops. The once-daily, BAK-free formulation was developed by Sun Pharma Advanced Research Co., Sun Pharma’s R&D division. The formulation is stable at room temperature and could reduce the risk of BAK-associated ocular surface disease and will be marketed in the United States by Sun Ophthalmics.


	Eye drops work well in glaucoma, but only when patients are compliant. New drug-delivery systems offer the hope of improved therapeutic outcomes by taking patient adherence out of the loop.

Punctal plugs that elute travoprost and latanoprost agents are in development. 

has phase I trials following preclinical data showing sustained zero-order drug release and a marked reduction in IOP. Clinical data to date show sustained IOP reduction for up to 150 days using a hydrogel formulation.

 uses a harder plug material that is easy to insert and remove. Phase II data from the Evolute plug shows sustained drug release over 12 weeks.

 has two delivery devices in development, a bimatoprost eluting ring and an injectable formulation. Both are showing very good efficacy in reducing intraocular tension, and both offer the opportunity to improve outcomes in patients with adherence problems.

A variety of innovators big and small are working on everything from printing latanoprost microdots (Enovia) to new Rho kinase inhibitors or more familiar prostaglandin analogs delivered via ocular implants (

) and even gene therapy using adenoviral vectors to deliver therapeutic genes into retinal ganglion cells to enhance survival (

BioMed/Manner Research and Aerpro are both developing a Tier 2 activating molecule targeting glaucoma. BioMed is working on POAG while Aerpro is working on POAG associated with diabetes.

Neuroprotection is another active field, and 

 is working on a human placental cell extract using intranasal delivery for neuroprotection. Disarm Therapeutics is developing a neuroprotective agent using a selective androgen receptor modulator (SARM) inhibitor.

Allergan had worked on oral memantine as a potential neuroprotective agent. Unfortunately, trials in nearly 2,300 patients comparing two doses of memantine and placebo found no effect on glaucoma visual field progression or optic nerve changes.

 has shown neuroprotection for retinal cells even after IOP elevation.

, InMed Pharmaceuticals, and Axim Biotechnologies, are working on a POAG agent based on cannabinoids. 

“There is plenty of innovation still coming for the treatment of glaucoma,” Dr. Lindstrom concluded.

This article was adapted from Dr. Lindstrom’s presentation at the 2018 meeting of the American Academy of Ophthalmology. Dr. Lindstrom has no financial interests to report. 

Eye drops, drug-delivery system innovations aim to improve outcomes, adherence

Pharmacologic pipeline makes waves in glaucoma

Artificial intelligence can provide a knowledge base that can be a foundation for the interpretation of data. The importance of the humanistic elements of medicine remain vital.

Creative disruption of an existing model can be the path forward to a new technology, such as telehealth, a market that has grown rapidly since 2012, and is expected to continue to do so over the next several years.

 defines telemedicine as the transfer of medical information via telecommunication (both synchronous and asynchronous) technology or specially designed medical devices for the purpose of delivering healthcare services and clinical information.

 is located, telemedicine is described as the use of interactive audio, video, or other telecommunications or electronic technology by a physician in the practice of medicine outside the physical presence of the patient. Multiple layers make up a telemedicine system-tools and equipment, modes or types of telemedicine, specialty areas of care, individual patients and providers, and the locations where virtual care will be delivered.


	Both the patient and the clinician need certain tools to perform telemedicine. For the patient these can include a laptop, smartphone, or other personal smart device, equipped with a camera, and the appropriate app or software. In some cases remote monitoring equipment at home can be useful as well. The physician will need similar equipment, plus additional software and peripherals for use in diagnosing disease or progression and reporting results.


	There are three main types of telemedicine, each good for certain uses. Only one of them (synchronous) requires the physician and patient to be available at the same time. Telehealth makes individualized, patient-centric care possible for patients across the continuum of care.


	At Johns Hopkins, telemedicine connects the entire healthcare system including community physician practices, all Johns Hopkins’ hospitals, affiliated hospitals, the home-care group, skilled-nursing facilities, and other collaborative facilities.


	Many challenges remain in using telemedicine, most significant of which include policies and reimbursement.
	While the telehealth market continues to grow, telemedicine volumes expand at the speed of reimbursement. Commercial payors have largely embraced telemedicine but government payors (Medicare and Medicaid) have been more reluctant. There has been an uptick in reimbursement for telemedicine services from government payors.
	Without appropriate reimbursement, providers will not utilize new technology tools for health care delivery. Direct-to-consumer online virtual care has been more successful with self-pay options. Other notable concerns include obtaining licensing to practice medicine across state lines (a medical license is required in the state where the patient is located), and credentialing and privileging at the originating site.

Johns Hopkins Medicine (JHM) Office of Telemedicine,

 established July 1, 2016, coordinates all telemedicine efforts across the Hopkins healthcare enterprise.
	The goals of the JHM system are to:

Simplify and streamline the process for physicians

Embed reporting within the electronic health record

Telemedicine reduces costs by avoiding unnecessary patient visits and face-to-face consultations with physicians and by replacing clinic visits with less-expensive virtual care. It increases revenue by giving patients greater access to healthcare, expanding patient share capture, and by increasing a facility’s capacity to accommodate referrals and transfers for more appropriate higher-acuity patients.

Johns Hopkins Medicine has launched 44 telemedicine programs, in 31 specialties, for more than 16,700 patient encounters. Another 38 programs are being prepared to launch, with 30 more in the early pipeline stages.
	
	


	Ingrid Zimmer-Galler, MD, associate professor of ophthalmology at the 

, points out that teleophthalmology has the potential to be a radical transformer of care delivery.

Current teleophthalmology programs include diabetic retinopathy screening; retinopathy of prematurity (ROP) screening; glaucoma screening and disease management; age-related macular degeneration screening and management; anterior segment and ocular adnexal disease’ and remote consultation and remote emergency department (ED) evaluations.


	Diabetic retinopathy screening is one of the first and most beneficial uses of telemedicine in ophthalmology. In the United Kingdom, diabetic retinopathy is no longer the leading cause of vision loss in working age adults in part due to a robust national telemedicine diabetic retinopathy screening (Liew, Michaelides, Bunce. BMJ Open. 2014;4:e004015).

Telemedicine programs allow patients with diabetes to receive appropriate and timely screening for retinopathy so that necessary intervention can be performed when it is most likely to be beneficial.

Telemedicine with photographic documentation for ROP screening offers increased access for neonatal intensive care units where specialty coverage is often limited.
	Telemedicine ROP evaluations may become part of the gold standard for diagnosis and monitoring of the disease given the benefits of retinal imaging over traditional retinal drawings.


	By 2030, nearly 440 million individuals worldwide will have diabetes. It is projected that by 2020 there will be a significant undersupply of ophthalmologists worldwide. The task of detecting and evaluating diabetic retinopathy will create a resource and economic burden to healthcare systems and telemedicine screening programs will be crucial to meet the demand.

The FDA recently cleared the first artificial intelligence system for use in the United States for diabetic retinopathy screening. This is expected to enhance the efficiency of diabetic retinopathy screening as physician review of every image will no longer be necessary.


	Urgent eye problems account for a significant percentage of ED to ED transfers from community hospitals to tertiary care hospitals due to the lack of ophthalmology coverage.

Telemedicine provides an opportunity to use an external high-resolution camera for a video visit and consultation to avoid transferring the patients from one ED to another.

It is anticipated there will be significant opportunity for growth of telemedicine in ophthalmology and those practices flexible enough to embrace it and integrate new technologies will come out ahead. â 


	This article was adapted from Dr. Zimmer-Galler’s presentation at the 2018 meeting of the Johns Hopkins’ Wilmer Eye Institute Current Concepts in Ophthalmology in Baltimore. Dr. Zimmer-Galler has no financial interests to disclose.

Telemedicine, teleophthalmology programs in action at Johns Hopkins

Artificial intelligence (AI) is a technology in which machines and equipment can “learn” from experience and adjust accordingly.

The technology has the potential to have a significant impact on ophthalmology in the coming years, according to Dimitri Azar, MD, MBA.

Three categories of algorithms exist for AI and machine learning:

Unsupervised learning, which groups data that has not been labeled and includes methods such as clustering,

Supervised learning, which infers a function from labeled training data and maps an input to an output based on example input-output pairs, includes linear regression analysis, support vector machine analysis, decision trees and random forests, and convoluted neural networks and deep learning, relying heavily on labeled data,

Semi-supervised learning, which uses mostly unlabeled data with a small amount of labeled data.

In ophthalmology, the neural network approach, deep learning, has surpassed other methods in recent years, said Dr. Azar, senior director of ophthalmic innovations, and clinical lead of ophthalmology, A

, and distinguished university professor and B.A. Field Chair of Ophthalmologic Research, and former dean, 

College of Medicine, University of Illinois, Chicago.

It requires a level of computing that was not available to most researchers in the past, so other approaches were more likely to be used. With today’s increased access to big data and analytics, there has been a plateau in the traditional methods, and deep learning has surpassed it simply because it requires more analytical abilities, Dr. Azar said.

“One thing about these learning networks that I was impressed with is that as the training commences, neural networks start off without any fine tuning, and return random results,” he said. “The neural network progressively learns the combinations and permutations of important features.”

A difference between neural networks and human beings is humans have trouble letting go of inaccurate information that was thought to be useful in the past. The convoluted neural networks, as they progressively learn combinations and permutations of the important features, learn to ignore the unimportant features in order to make better algorithms, he noted.


	The process for networks features a weight given to particular features-the features the network finds most important are given the most weight.

As the network learns more, these weightings shift, so a feature that early on was thought to be important will be given less weight, as other features are given more weight. This allows the networks to make more accurate predictions.

The algorithm-generation process begins with an input layer. Before there is any output, there are several hidden layers. While the applications operate like black boxes, the results are not always given with an explanation. This makes detection of inappropriate outcomes difficult.

As algorithms become more powerful, the methods for troubleshooting them may lag behind. This can be a potential limitation of this approach, and may have implications in obtaining approval.


	Dr. Azar did a literature search on the use of AI in ophthalmology, in the topics of AMD, diabetic retinopathy, retinopathy of prematurity, dry eye, keratoconus and corneal topography, glaucoma and visual fields. The search showed that from 2016 through the first half of 2018, a period of 30 months, there were about twice as many publications as in the 60 months from 2006 through 2010.

AMD and diabetic retinopathy were the most frequent topics found in the search. There are AI applications in corneal topography (some of which stem from his early work with Dr. Paul Lu), and dry eye as well. He proposes future studies using AI that could simplify the diagnosis and classification of different types of dry eye, and their severity, potentially replacing the current methods that are based on the DEWS II report.

In glaucoma there are studies utilizing deep learning. Some look at the nerve fiber layer, and others focus on the optic nerve. Future studies will include using OCTA to examine blood flow in the retina, adjacent to the nerve.


	When it comes to AI, a major problem in glaucoma is that the data comes from multiple sources, which frequently do not connect with each other. There may be billing data, patient information, imaging data, etc., all coming from different sources. While it would seem easy to connect these sources, doing so in a HIPAA-compliant way is not always easy.

A group at the University of Illinois has developed a machine learning method to collect the data that comes from various diagnostic tools and records, under one name or one eye, staying within HIPAA regulations.

Currently, most physicians look at data from multiple sources to make a diagnosis. The time-consuming visual field test is probably the most reliable way of detecting whether or not someone is progressing.

By linking data together, with AI applications it should become easier to find shortcuts that can predict what will happen in glaucoma, like current trends in diabetic retinopathy, he said.


	Fundus photography is a big area of AI applications in ophthalmology. The Google group has published articles on the use of AI in several areas, including refractive error prediction, cardiovascular risk detection, identification of retinal lesions, and diabetic retinopathy.

An article published in IOVS in 2018, using data from the UK biobank and the AREDS study, looked at attention maps that predicted refractive error, and the refractive error prediction was good.

 One of the figures in the article shows attention maps for myopic, emmetropic (neutral), or hyperopic patients.

Even with the computer pointing out the areas it used to make a diagnosis, it can still be difficult for an ophthalmologist to make a diagnosis. The computer is able to do so.

In a landmark article published in Nature Medicine in 2017 the researchers used deep learning architecture to make referral recommendations in a group of OCT scans, done with multiple devices.

By using tissue segmentation they were able to give a probability of a diagnosis, and accordingly make an urgent, semi-urgent, routine, or observation only referral suggestion.
	If “trained” properly, the machine was able to have an error rate in the patient referral decisions of only 5.5%. This was better than 80% of the retina specialists and all of the optometrists, who were given not only the OCT data, but also fundus data and notes.

Dr. Azar also discussed the article he called “transformational.” It was published in JAMA 2016 and focused on using deep learning for diagnosing DR.3 The study results showed very high sensitivity and specificity, and indicated that diagnoses were as good or better than those of the retina specialists who were convened there.
	An area receiving the most attention in retinal AI applications is OCT. One study looked at the severity, characterization and estimation of 5-year risk of AMD progression using AI and found that the machine did very well.

Another group studied the prediction of individual disease conversion in early AMD.5 The authors found the most critical quantitative features for progression were retinal thickness, hyper-reflective foci, and drusen areas. Dr. Azar said, “The interesting part here is this is not only predicting a disease, but also includes discovery.”


	There are great applications of AI in ophthalmology, and the uses will continue to expand. But there are several limitations, including:

The quality and diversity of training sets

Because the statistics are very good, people may erroneously conclude that the system is not making errors

The black box effect of convoluted neural networks


	More than 100 years ago, the famous Flexner report established the biomedical model of education, training, and research as an enduring basis of medical education.6

Dr. Azar described this as a cross-disciplinary convergence in ophthalmology, but said there are caveats when it comes to the education of ophthalmology fellows, residents, and students. There needs to be an understanding that the knowledge base these machines can provide should only be a foundation to facilitate interpretation of data.
	He emphasized the importance of the humanistic elements of medicine-professionalism, communication, empathy, compassion, and respect.

Dr. Azar said these important topics should now be included in the curriculum of all medical students, because much of the information doctors previously needed to learn will be easily obtained through AI.


	This article was adapted from Dr. Azar’s presentation at the 2018 Johns Hopkins Wilmer Eye Institute’s Current Concepts in Ophthalmology meeting in Baltimore. At the time of the presentation, Dr. Azar was on the Board of Directors of Novartis, and the Board of Directors of Verb Surgical, and received NIH funding including an RO1 grant. He is currently an employee of Alphabet Verily.

Artificial intelligence has the potential to have a significant impact on ophthalmology in the coming years.

Artificial intelligence gains more acceptance in ophthalmology

Researchers explored messenger RNA in pediatric cataracts linked to multiple genes that have been previously identified as being associated with pediatric cataracts. 

Pediatric cataracts are the cause of up to 15.3% of childhood blindness cases, and while different types of pediatric cataracts have distinct etiologies and morphologies, almost nothing is known about the genetic pathways that underlie these cataracts.

While pediatric cataracts can be treated, surgery is almost always an emergency.

“We have studies of gene mutations in human families,” said Joyti Matalia, MBBS, 

The mutations have been studied extensively in animal models, according to Dr. Matalia.

“In humans, these studies have been done only in blood and never in lens matte,” she said. “The causes of pediatric cataract have been investigated. There is no report on the molecular mechanisms of congenital cataracts or on gene expression in these cataracts.”

Dr. Matalia described an investigation of gene expression and molecular pathways that lead to different types of pediatric cataracts. Researchers explored messenger RNA in pediatric cataracts linked to multiple genes associated with pediatric cataracts.


	The study classified pediatric cataracts into five different categories.

Congenital cataracts were classified as either infectious, associated with rubella, cytomegalovirus, or a combination of the two infections, or noninfectious, whether linked with an inherited condition or lens opacification due to posterior capsular anomalies.

Acquired cataracts were classified as postnatal, cataracts which developed anytime during childhood after the first birthday, secondary cataracts that developed following other ocular diseases such as uveitis, or traumatic cataracts.

Three types of genes are known to be associated with pediatric cataracts. Structural genes maintain transparency of the lens.

Transcription factor genes play roles in lens development. Fibrotic genes maintain the physical integrity of the lens.
	Errors in the expression of any of these genes has the potential to result in early cataracts.


	The group explored three lens structural genes, including:
	> AQP-0 is responsible for aquaporin-0, a protein which maintains lens transparency. Defects in this protein can result in inherited cataracts and congenital glaucoma.
	> HSPA-4 is responsible for heaths shock protein A-4, found in lens epithelium and fibers. Mutations are reported to result in inherited cataracts.
	> CRYG-C is responsible for crystallin gamma-C, a protein which maintains lens transparency. Mutations are associated with autosomal dominant nuclear cataracts.

Four lens transcription factor genes were also included. They are:
	> MAF codes for musculoaponeurotic fibrosarcoma oncogene, which is involved in the regulation of crystallins. Defects cause cataracts and anterior segment dysgenesis.
	> TDRD-7 codes for Tudor domain containing protein 7, which regulates genes critical for lens development. Defects can give rise to cataracts, microcornea and coloboma.
	> FOXE-3 codes for forkhead box 3, which may play a critical role in lens development. Defects have been associated with cataracts as well as anterior segment dysgenesis.
	> PITX-3 codes for pituitary homeobox 3, also suspected of playing a critical role in lens development. Defects have been associated with cataracts and anterior segment abnormalities.

Four genes have been associated with profibrotic factors. These factors are typically seen in a fibrotic cascade in skin and other tissues, Dr. Matalia said, and might be active in cataract formation. These genes include:
	> TGF-Î² codes for transforming growth factor beta.
	> Î±-SMA codes for alpha smooth muscle actin.
	> VIM codes for vimentin
	> BMP-7 codes for bone morphogenic protein-7.

Researchers conducted a prospective cross-sectional study using lens material extracted during routine cataract surgery on 90 patients younger than 16 years of age. There were eight study groups, postnatal cataract (nine patients), secondary cataracts (13 patients), traumatic cataracts (13 patients), rubella cataracts (nine patients), rubella + CMV (eight patients), CMV cataracts (nine patients), prenatal cataracts (nine patients), and posterior capsular anomalies (10 patients).

A control group of 10 patients had clear subluxated lenses.

Among the structural genes, only AQP-0 and CRYG-C are highly expressed in infectious CMV and prenatal cataracts, Dr. Matalia noted, concluding that prenatal cataracts likely derive from transcriptional hyperactivity of AQP-0 and CRYG-C. HSPA-4 appears to play no role in the development of prenatal cataracts.

Among the transcription factor genes, TDRD-7 is expressed in the CMV group, supporting the involvement of both AQP-0 and CRYG-C. This same pathway is inhibited in rubella cataracts.

FOXE-3 also shows higher expression in the prenatal group, but PITX-3 and MAF expression is similar to the control group.
	Fibrotic gene expression shows a significant decrease in both the rubella and rubella + CMV groups, suggesting that the fibrotic pathways seen in other tissues do not contribute to the formation of cataracts in younger patients.

“This is the first preliminary study to talk about gene expression in lens matter from pediatric human cataracts,” Dr. Matalia said. “There is definitely a correlation between gene expression and cataract morphology. If we are able to find the precise pathway and a drug to block it, we can modify gene expression using target-specific therapy.” 


	This article was adapted from Matalia’s presentation at the 2018 meeting of the American Academy of Ophthalmology. Matalia has no financial interests to disclose. 

Differing signaling pathways could lead to different types of cataracts in children.

Exploring correlation between gene expression, cataract morphology

Deep phenotyping is key to correct diagnosis with corneal opacities and can ensure physicians do not overlook the root causes of some eye issues.

Deep Phenotyping is shedding new light on congenital corneal opacities, said Kanwal K. Nischal, MD, FAAP, FRCOphth.

Deep Phenotyping is shedding new light on congenital corneal opacities, and with the initial diagnoses, physicians should also delve deeper to avoid overlooking the causes of some ocular anomalies.

Ophthalmology Times, April 15 2019

Managing unique challenges of pediatric congenital cataract

Pharmacologic pipeline makes waves in glaucoma

Telemedicine, teleophthalmology programs in action at Johns Hopkins

Artificial intelligence gains more acceptance in ophthalmology

Exploring correlation between gene expression, cataract morphology

How deep phenotyping of pediatric corneal opacities targets diagnosis

‘Spark joy’ in your ophthalmic practice with these frame board tips