Advances in imaging technologies improve screening, detection of diabetic eye disease

For years, the main reason for ophthalmic screening in patients with diabetes has been to detect the presence of diabetic retinopathy or diabetic macular edema — both of which may lead to blindness if untreated — relying on either a dilated fundus examination or color photography.

One expert estimated that a dilated fundus examination performed by an ophthalmologist detects diabetic retinopathy in more than 90% of cases. Even so, subtle diabetic macular edema may not be picked up without an optical coherence tomography examination, according to David M. Brown, MD, FACS, of Retina Consultants of Houston.

“In 30% of cases of diabetic macular edema, you will miss it without either a good exam or an OCT,” Brown said.

 

Even as a growing number of clinicians are enthusiastically adopting photographic screening programs for diabetic retinopathy, these programs “are not considered a replacement for a comprehensive eye evaluation by an ophthalmologist experienced in managing diabetic retinopathy,” according to the American Academy of Ophthalmology Preferred Practice Pattern.

On the other hand, in a review of diabetic retinopathy management guidelines that appeared in Expert Review of Ophthalmology, the authors noted, “While the slit lamp has advantages of availability and affordability compared with photography, the disadvantages of its routine use in a low-resourced setting included availability of trained ophthalmic staff and need for pupil dilation.”

Even with guidelines and published preferred practices, U.S. ophthalmologists continue to customize their screening practices, depending on personal and patient circumstances.

Type 1, type 2 diabetes

“[In type 1 diabetes], you typically know the onset date,” Brown said, adding that it would be unusual for a patient with type 1 diabetes to acquire diabetic retinopathy within the first 5 years of onset. “Thus, these patients need screening not initially but fairly soon afterward.”

If there is absolutely no retinopathy upon the initial screening, Brown normally proceeds with screenings once every 2 years in patients with type 1 diabetes. “But if you have any retinopathy, you need to be seen yearly,” he said.

In contrast, type 2 diabetes has an insidious onset, and patients typically do not know how long they have had the disease.

“They might have had it 10 years before diagnosis, so these patients need screening immediately and yearly,” Brown said.

Brown said the advantage of early screening in both types of diabetes is that patients become better educated about managing their blood glucose level, and among diabetic children, education is even more important. For women with diabetes who become pregnant, Brown screens every trimester because they “are at greatly increased risk of trouble.”

Scheduled screenings are critical, so that treatment can commence if needed, Brown said.

“We now have very good treatments for diabetic macular edema,” he said. “[Anti-VEGF drugs] improve vision by two lines or more consistently, but left untreated, you really lose the magnitude of those gains.”

In clinical trials, most patients who delayed treatment by 2 years gained less than half the vision of patients who started treatment immediately.

“Unfortunately, though, there is not much, if any, screening for diabetic macular edema because the best tool is an OCT; however, insurance companies and Medicare only pay if you have a diagnosis,” Brown said. “So if you screen and results are normal, you do not get paid.”

Brown said the most widely used and easiest imaging technology for detecting diabetic retinopathy is non-mydriatic fundus photographs.

“These pictures are better than nothing,” he said, but there is a greater chance of missing pathology. “In other words, you really cannot see if someone has macular edema unless they have hard exudates, and often you will miss inferior vitreous hemorrhage.”

A better option is widefield imaging, which can show inferior vitreous hemorrhage, Brown said, but it can also miss some macular edema.

Widefield imaging

“In the past, we did not have imaging modalities to image the far periphery of the retina,” OSN Retina/Vitreous Board Member Seenu M. Hariprasad, MD, said. “The older non-mydriatic imaging technologies limited our view to the posterior pole of the retina.”

Over the past 3 years, new-generation imaging technologies have emerged, including ultra-widefield imaging and angiography systems, which enable visualization of the far periphery of the retina, where a lot of pathology in diabetic retinopathy occurs, according to Hariprasad.

“A lot of the pathology in diabetic retinopathy occurs in the far edges of the retina. We want to make sure we are not missing anything,” he said.

Kernt and colleagues demonstrated that non-mydriatic ultra-widefield retinal imaging is comparable to the gold standard of dilated ETDRS seven-field 35-mm stereoscopic color fundus photography, according to Hariprasad.

“This gold standard of imaging the retina consists of taking seven pictures of the retina while the patient is dilated; then, a computer program fuses the images together so you can see all the edges of the retina. Clinicians have used the seven standard field very reluctantly, in part because it is time-consuming and it takes someone with knowledge to administer drops to the eye for dilation,” Hariprasad said.

A decade ago, the AAO did not advocate screening programs for diabetic retinopathy, according to Hariprasad, but the latest Preferred Practice Pattern states: “Future research also should include establishing standardized protocols and satisfactory performance standards for diabetic retinopathy screening programs.”

Hariprasad called this stance “a very important paradigm shift” for the AAO.

“With new-generation ultra-widefield imaging, we need to revisit the applicability of diabetic screening programs,” he said.

Furthermore, because imaging can be done without dilation, diabetic screenings can now be performed on a community-wide basis, Hariprasad said. However, some people feel community screening is a disservice unless screening includes OCT and IOP measurements.

“There are shortcomings of screening programs, despite our best ability,” Hariprasad said. “OCT is costly to implement and measuring IOP is difficult because most of the technologies require contact with the eye and necessitate a higher level of training. If cost was not a barrier, I would certainly be an advocate of measuring IOP and macular thickness using OCT. I am not an advocate of fluorescein angiography for screening as it is too invasive.”

Hariprasad’s ideal screening program would consist of three tests: ultra-widefield imaging, macular OCT and an IOP check.

“Incorporating these tests would provide us with a world-class screening program,” he said.

Telemedicine

With improvements in imaging systems and connectivity, telemedicine screening for diabetic retinopathy is an option to consider to improve access and reduce barriers for care, according to OSN Retina/Vitreous Board Member Judy E. Kim, MD. With the support from the Vitality Award and a grant from the Healthier Wisconsin Partnership Program, she has been evaluating successful implementation of telemedicine for eye screening in medical and community settings.

“In primary care settings, as a part of a patient’s diabetes care, color fundus photos are taken by a medical assistant, technician or nurse,” Kim said. The images are transmitted to her for review, and the results are sent back to the practice.

“Patients are either referred to an ophthalmologist or they schedule a 1-year follow-up screening exam, depending on the findings of the photographs,” Kim said.

A pilot study headed by Kim assessed the sensitivity and specificity of detecting diabetic retinopathy via telemedicine using photographs compared with detection via slit lamp examination by eye care specialists at an eye clinic.

“When patients were dilated, telemedicine was very accurate,” Kim said. “I was able to grade all the images and distinguish presence or absence of retinopathy with greater than 98% correlation with the eye care specialists’ findings. In fact, sometimes we actually saw more on the picture than the in-person examination. Some patients are difficult to examine at the slit lamp due to poor cooperation or excessive blinking because of the bright lights. On the other hand, the camera can capture the fundus image in a split second and document the fundus findings for more detailed examination. In certain instances, we were able to detect microaneurysms and hemorrhages on photographs that were missed at the slit lamp.”

According to Kim, fluorescein angiography is helpful for better classification and management of diabetic retinopathy but is not practical for screening purposes.

“It is an invasive procedure that can result in potentially significant side effects. It requires a skilled photographer, and the camera is more expensive. It is not cost-effective as a screening modality at this time,” she said.

Kim believes telemedicine screening is the wave of the future, as more inexpensive, automated and portable cameras develop.

She said that telemedicine does not replace a complete ophthalmic examination by an ophthalmologist.

“In addition to detecting the presence of other eye diseases besides diabetic retinopathy, an annual examination in the office provides an opportunity for us to further educate our patients on the importance of controlling blood sugar and blood pressure,” Kim said. “However, diabetes is becoming an epidemic, with a significant increase in the numbers of persons with diabetes being expected here in the United States and worldwide. We know that early detection and treatment of sight-threatening diabetic retinopathy is crucial. Given limited resources and access to health care providers, as well as a less-than-acceptable rate of diabetic retinopathy screening in the current delivery of the health care model, we need to consider screening modalities that are more cost-effective and accessible. Telemedicine may be the answer.”

Despite the potential benefits of telemedicine in screening for diabetic retinopathy, Kim believes the absence of or poor reimbursement for screening in many states is a current deterrent for wide acceptance of this technology.

“I believe there needs to be a policy change that values screening and preventive care,” she said.

American Diabetes Association

The 2014 American Diabetes Association (ADA) recommendations for screening diabetic retinopathy are similar to those of the AAO and have not changed for many years, according to Endocrine Today Board Member Zachary T. Bloomgarden, MD, MACE, an endocrinologist and clinical professor of medicine at the Icahn School of Medicine at Mount Sinai in New York.

“The ADA encourages an initial dilated and comprehensive eye examination by an ophthalmologist or optometrist within 5 years after diabetes onset in adults with type 1 diabetes and shortly after diagnosis of diabetes in patients with type 2 diabetes,” Bloomgarden said.

If there is no evidence of retinopathy at the initial examination, then an examination every 2 years can be considered. But if diabetic retinopathy is present, subsequent examinations for patients with type 1 and type 2 diabetes should be repeated annually by either an ophthalmologist or optometrist, he said. More frequent examinations are required if the retinopathy is progressing.

The ADA suggests that high-quality fundus photography can detect most clinically significant diabetic retinopathy.

“Some ophthalmologists, though, advocate going directly to OCT,” Bloomgarden said. “However, the minor degrees of macular edema detected by OCT are of uncertain significance, so OCT has not been adopted as a screening tool at the present time. Fundus photography or binocular fundus examination is the way of looking for macular edema.”

“I feel ophthalmologists absolutely do a great job at screening and detecting for diabetic retinopathy,” Brown said. “The problem is that very few diabetics are seen by an ophthalmologist.”

Most patients with diabetes are primarily cared for by general medical doctors or endocrinologists.

“It is crucial for these physicians to have an awareness of the importance of diabetic eye disease,” Bloomgarden said. “Referring people with diabetes for eye care absolutely makes sense because between 25% and 50% of people with type 1 diabetes also have retinopathy, and probably around 5% have either proliferative retinopathy or macular edema.”

The targeted population for referrals to ophthalmologists are diabetic patients with elevated glucose levels.

“We should focus on patients with high blood glucose because they are the most likely to have diabetic eye disease,” Bloomgarden said.

Widefield angiography

Szilárd Kiss, MD, director of clinical research and an associate professor of ophthalmology at Weill Cornell Medical College in New York, has devoted a considerable amount of his research since 2008 to ultra-widefield imaging, including ultra-widefield fluorescein angiography.

“It tells you a lot more than 7SF stereoscopic imaging,” he said. “It also tells you more than a dilated exam.”

Ultra-widefield angiography entails injecting a dye in the hand and then taking pictures as the dye travels through the eye.

“You are looking at the blood vessels, whether they leak or not, whether the blood-retinal barrier is being maintained,” Kiss said. “You are also looking at areas where there are no blood vessels and areas of abnormal blood vessel growth.”

Kiss said diabetes causes blood vessels to leak and subsequently leads to macular edema, as well as blood vessels to block off, both resulting in retinal ischemia.

“These two events lead to abnormal blood vessel formation and neovascularization,” he said.

Ultra-widefield angiography will have an impact on screening for diabetic retinopathy, which affects the entire fundus, Kiss said.

“Traditional imaging misses a lot of pathology,” he said. “In fact, the classification of diabetic retinopathy is wrong in about 10% to 15% of cases unless you do ultra-widefield imaging.”

Traditional imaging provides a maximum field of view of about 50°, and “even with the seven standard field, you only get out to about 100°,” Kiss said. “With ultra-widefield imaging, you get out to 200° within one shot, so you are able to view over twice as much of the retina or 80% of the entire retina in one shot. You can imagine the amount of pathology that one sees by looking at more of the retina. This considerably greater view also has some impact in follow-up and treatment.”

Furthermore, compared with a montage of photographs via normal angle fundus cameras, ultra-widefield fundus imaging is often easier for the patient, he said.

Kiss is the lead author of an article on ultra-widefield fundus imaging for diabetic retinopathy in Current Diabetes Reports, which concludes, “As the population of diabetic individuals continues to grow in the United States and the rest of the world, the ability to capture a widefield image in a single photograph may improve our ability to diagnose, grade, and treat diabetic eye disease both here and abroad.”

A combination of an OCT and a widefield picture likely provides a good screening, Brown said.

“But is it as good as a dilated fundus examination by someone highly competent? Probably not,” he said.

Checking for visual acuity is another reasonable part of screening, according to Brown.

“I think a widefield image and a normal vision is a pretty good screen for the masses,” he said. “However, currently only about 50% of diabetics are screened by any method. The easier we can make the screening for that other 50% of patients, the more disease we will pick up and hopefully save vision.”

“By detecting disease early using various imaging modalities, we can prevent vision loss and improve quality of life for our patients while saving a significant amount of money for the insurance companies, the society and the patients. I definitely think we need to emphasize and move forward with early screening. This will require a team approach, better education for our patients and the medical colleagues on the importance of eye examinations, and thinking outside the box regarding the delivery of screening modalities, such as telemedicine,” Kim said. – by Bob Kronemyer

References:

For more information:

What is the best way to ensure that more patients with diabetes are screened for retinopathy?

Telemedicine is best approach

The best approach to screening persons with diabetes is through telemedicine, whether in a kiosk at the local mall or in a primary care physician’s office. However, primary care physicians need advanced training to accurately interpret what they see.

One of the problems with telemedicine historically is this: Because you can screen all people, all people get screened. That creates a very large burden on the screening populations for a low yield on actionable items, unlike screening for retinopathy of prematurity, which is a high-yield event for preventing blindness.

The challenge in screening for diabetic retinopathy is deciding whether to screen all patients or only those with diabetes. A second problem is that a distinction between proliferative and nonproliferative disease needs to be made once retinopathy is identified. A third component is the need to identify macular edema. To screen patients with diabetes only, though, you are going to detect a lot of patients with retinopathy who do not necessarily need to be seen or treated because the retinopathy is nonproliferative. For other patients, to be absolutely certain, you need some kind of confirmatory exam, either in person or with fluorescein angiography.

Additionally, the camera is inferior by itself at detecting the leading cause of blindness from diabetes, which is macular edema. For that, you need an optical coherence tomography device, which is severely limited in where it can be placed. The purpose of screening should be not to identify retinopathy, but to identify intervention points; otherwise, you will be swamped with disease that does not need to be seen.

Another option is for trained people to go out in the community and actually evaluate the patients themselves. But if that was easy and affordable and economical, we would already be doing it. Ultimately, we are headed toward home monitoring, perhaps using a camera that looks in the eye once a month. I believe home monitoring will be the simplest screening method. If there is an ocular change, the primary care physician may not know what the change represents, but he can get the patient plugged in to the health care system.

Darius M. Moshfeghi, MD, is the director of telemedicine at the Department of Ophthalmology, Byers Eye Institute, Stanford University School of Medicine. Disclosure: Moshfeghi has no relevant financial disclosures.

Make screening a health care priority

There are excellent national programs throughout the world that guarantee widespread screening of people with diabetes for retinopathy. The program in the United Kingdom is considered a case model for annual screening and has been in existence since 2008. Such a program relies on government to make sure that the resources are allocated in a way that everyone with diabetes receives screening on a regular basis. Because the U.K. has a national health system, the program is relatively efficient and effective.

But in the United States, screening for retinopathy and preventing blindness among people with diabetes is more of a challenge: People have varying types of insurance, and therefore we do not have a clear path for screening on a large scale like in the U.K. One major obstacle is a lack of a clear reimbursement paradigm for people who want to do the screening effectively. At present, the most efficient method to screen people with diabetes for retinopathy is to take pictures of the retina and have them graded at a central location. However, reimbursement for this activity is a complicated process, and often such service is not covered. If the United States wants more people with diabetes to be screened for eye disease, then it must become a health care priority. Once this decision is made, systems can be put into place that are run either at the state or federal level.

Screening for diabetic retinopathy is not complex. The cameras used to take images of the fundus are exceedingly good and many of them can be used without dilating the eye. In many countries with routine screening, the need to perform vitrectomy for diabetic retinopathy is incredibly rare.

There is no question that if everyone with diabetes knew they needed an eye exam, more people would avail themselves of the service. But, public health screenings that depend on the patient to report for care are more likely to fail. In some parts of the United States, there are no convenient locations for a screening exam. Easy access to care is important if screening is a priority. Removing the friction from the process will allow greater access to services. An especially important problem for some with no or limited insurance is what to do if they have diabetic retinopathy. Identifying an affordable provider who will accept insurance or payment plan places additional barriers in the way of blindness prevention.

Screening for diabetic eye disease is known to be highly cost-effective, so earmarking the resources for identifying and treating disease before vision loss occurs should save money to society. Clearly we can do better than we are doing now.

Derived from an interview with David S. Friedman, MD, MPH, PhD, the director at the Dana Center for Preventive Ophthalmology, Wilmer Eye Institute and Bloomberg School of Public Health. Disclosure: Friedman has no relevant financial disclosures.