Smartphone Screening for Eye Diseases

Millions of Americans are at risk for sight-threatening diseases and are either unable to access eye care services or unaware that eye care screening could reduce their risk of vision loss. In addition, there is a national shortage of eye care providers. Primary care providers lack resources to effectively screen for sight-threatening conditions. Of the three leading causes of permanent vision loss and blindness in adults in the United States (namely, diabetic retinopathy, glaucoma, and age-related macular degeneration), all three diseases largely affect structures in what is referred to as the "posterior part of the eye." This encompasses the retina and the optic nerve, as well as accompanying vasculature. The posterior part of the eye requires advanced instrumentation, training, and technological advances to visualize pathology and optimize examination in the posterior part of the eye, and accordingly screen for sight-threatening eye diseases in a manner that was not feasible previously. However this is still largely restricted to eye care-specific settings that have the training and resources available to invest in these technologically-advanced instruments. Thus, there remains a critical need to develop safe, effective, readily available, and efficient methods and tools to optimize examination and screening for sight threatening diseases. In addition, these tools should be applicable in a variety of eye care settings, including inpatient and outpatient clinical contexts in order to optimize eye care. In the absence of such tools, patients will continue to have limited access to eye care and remain at risk for sight-threatening eye conditions that otherwise could be prevented and managed appropriately.

Diabetes mellitus is a leading cause of death and disability worldwide. The World Health Organization estimates that more than 346 million people worldwide have DM and it is estimated that 552 million people may be affected by 2030. Diabetic retinopathy is the most common microvascular complication of diabetes, affecting nearly all patients with a history of diabetes for 15 years or more. Among US adults aged 40 or older, an estimated 28.5%-40.3% of diabetics have diabetic retinopathy and 4.4%-8.2% have vision-threatening retinopathy, with higher estimates among non-Hispanic black individuals. Epidemiological studies from the National Eye Institute reveal that diabetes is the leading cause of vision loss in working-age adults in the U.S., and guidelines recommend that patients with diabetes undergo, at minimum, a yearly dilated examination of the posterior part of the eye. This is particularly important to detect and treat diabetes-associated eye conditions, as patients are often asymptomatic until an advanced stage, at which point vision recovery may be limited. Despite the absence of visual symptoms in those patients with no or with early diabetic eye complications, conditions such as diabetic retinopathy have a recognizable latent stage, which is readily diagnosed and graded by retinal imaging (fundus photography). Given its prevalence and clinically-silent (but examination-apparent) progression, diabetic retinopathy is an ideal screening disease. Furthermore, visual outcome favors early detection and treatment.. There are effective, universally agreed-upon treatments for diabetic retinopathy which include laser photocoagulation for proliferative retinopathy, intravitreal injections and focal photocoagulation for diabetic macular edema, and improved glucose and blood pressure control. Despite guidelines for annual screening and benefits of early detection as it relates to vision outcomes and overall diabetic disease control, there is a substantial gap in care in regards to annual retinal screening for diabetics. It is estimated that nearly half of adults with DM in the US do not receive recommended screenings. Screening may be even more sporadic in areas of the world with a shortage of ophthalmologists.

This study seeks to determine whether a smartphone camera and an attachable device with optics and illumination specific for imaging the posterior part of the eye is an effective screening tool that could be used to optimize the posterior eye exam, and to improve access, ease of diagnosis, and early intervention in eye diseases.

Glaucoma is another disease affecting the posterior part of the eye, specifically the optic nerve, and is the second leading cause of blindness worldwide. In particular, open-angle glaucoma is a chronic, insidious disease with progressive loss of peripheral to central vision. When diagnosed and treated early, however, permanent vision loss, which manifests in later stages of the disease, may be prevented. The diagnosis of glaucoma typically relies on both structural changes to the optic nerve and functional visual field deficits. Diagnosing and screening glaucoma in resource-poor settings can be challenging. For example, a study of the Frequency Doubling Technique (FDT) perimetry in rural India found very low sensitivity for this test (7%), whereas studies in industrialized settings have found much higher sensitivity (59%). Furthermore, many structural testing modalities such as OCT of the retinal nerve fiber layer and stereoscopic optic nerve photography require both specialized equipment and expert analysis in order to be effectively used in diagnosis.

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness in the developed world. Worldwide, 14 million people are blind due to wet AMD. By 2020, an estimated 196 million people will suffer from AMD worldwide. The US population with AMD will double by 2050. Dry AMD accounts for 90% of AMD patients, but wet AMD causes the majority of visual impairment. Dry AMD is characterized by drusen of various size and characteristics and pigmentary changes with a classification system. Fundus photography is an established method of performing this characterization. Patients with dry AMD are at risk for progression to wet AMD, which occurs at a rate of 13% within 5 years. When dry AMD progresses to wet AMD, changes occur in the fundus appearance including the presence of retinal fluid and bleeding. Wet AMD requires prompt treatment to prevent vision loss. The visual acuity at initiation of anti-VEGF therapy is the best predictor for final visual acuity 2 years after therapy. Treatment of wet AMD with anti-VEGF therapy results in improvement or stabilization of vision in over 90% of wet AMD patients. Currently patients with dry AMD are evaluated annually for possible progression and then are instructed to monitor their vision daily at home and approximately half of cases of wet AMD come through patient symptoms and half through asymptomatic physician follow-up. Several home monitoring systems have been developed, including Amsler grid and Foresee Home monitoring using Vernier hyperacuity. Fundus photography is another method for monitoring due to the visually apparent changes that can occur, and utilization of a low-cost, smartphone-based fundus photography system would greatly increase access.

Additionally, within pediatric populations, there is a similarly critical need to develop effective, efficient, and child-friendly methods to evaluate for pediatric eye (particularly intraocular/posterior eye) diseases. Pediatric patients can pose a challenging exam, where general anesthesia is sometimes needed if a child is unable to tolerate a complete eye exam. Refined instrumentation and technological advances that are more comfortable for the pediatric patient may make it possible to explore novel, noninvasive photographic tools to optimize eye care in this special patient population within a wider range of established care settings and minimize the need for exam under anesthesia or distress to the patient. For example, leukocoria (or a white red reflex) can be an important referral to ensure that a patient does not have an eye cancer called retinoblastoma. Given the difficulty in examining some patients, providers at times have to take these patients to the operating room for exams under anesthesia, which has risks of anesthesia along with high cost.

Importantly, Retinopathy of prematurity (ROP) occurs when abnormal blood vessels grow over the retina and cause blindness in premature babies. ROP happens more often because more babies are surviving at younger ages due to advances in neonatal care. Screening for ROP is challenging because many hospitals do not have children's eye specialists on staff or nearby. This is a much larger public health problem in developing countries where the neonatal care has improved but ophthalmic expertise in pediatric eye diseases is lacking. ROP telemedicine program allows physicians to send images to be interpreted at larger medical centers, like the U-M Kellogg Eye Center. Babies born before 31 weeks and weighing less than 1,500 grams undergo repeated screenings to monitor for ROP. Photographs of the babies' eyes are taken using a pediatric retinal camera. Kellogg team routinely evaluates images together or remotely to make follow up and treatment decisions.

As the above examples indicate, there is a need for cheap, reliable, and user-friendly diagnostic modality for screening for posterior eye disease, particularly in resource-poor communities and internationally, in order to prevent blindness. Increasing the accessibility and portability of the eye exam holds great promise in improving the reach of ophthalmic care, in both the remote community and inpatient settings. Ophthalmology is uniquely suited for remote screening and evaluation given the optical transparency of the eye and the importance of the ocular examination in diagnosing eye disease. This study evaluates a smartphone-based camera capable of capturing high quality, wide field images of the retina.

This is a multi-site, cross-sectional diagnostic study being conducted at multiple sites . At each site, patients will be approached in the clinic, inpatient, and operating room settings. Photographs will be taken with our smartphone-based camera along with the standard of care, including traditional desktop fundus photography. The feasibility and accuracy of a smartphone-based camera for diagnosing eye diseases will be tested.

Several ophthalmology specialists will grade the smartphone fundus photographs, the traditional retinal photographs, and the documented eye examination for eye disease in question (such as diabetic retinopathy, age related macular degeneration, glaucoma, retinopathy of prematurity, etc.). The agreement between the graders for the diagnosis of eye disease will be assessed. The sensitivity and specificity of diabetic retinopathy or glaucoma diagnoses with the smartphone will also be assessed, using traditional retinal imaging as the reference standard and in a separate analysis using the ophthalmologist's examination as the reference standard. The following are the aims of the project:

Aim 1: to assess the sensitivity and specificity of diagnosis of posterior eye disease (e.g., diabetic retinopathy, glaucoma, macular degeneration) with a smartphone camera relative to (A) ophthalmologist examination in outpatient and inpatient care settings and (B) traditional fundus photography when feasible.

Aim 2: to assess the feasibility of our smartphone-based camera as a pediatric-friendly approach for imaging of posterior eye disease (retina and optic nerve head) in the outpatient, inpatient, and operative care settings, and to assess sensitivity and specificity of diagnosis in screening for important pediatric conditions (e.g. leukocoria and retinopathy of prematurity).

Aim 3: to assess the agreement between masked graders in diagnosis of different diseases involving the posterior eye as imaged by our smartphone-based camera.