Early Detection of AMD Is Crucial to Preserve Vision
Even with better treatment options, early detection is key to preserving vision and maintaining quality of life.
Choroidal neovascularization (CNV) is the main cause of vison loss in individuals with age-related macular degeneration (AMD), occurring in 18% of patients over 5 years.1,2 We have available treatments for CNV, however, they are not capable of reversing CNV-induced injury, and in the majority of cases, the treatments do not improve vision. Early detection among AMD patients is the key to preserving functional vision. The benefit of treating AMD lesions while they are still relatively small has been well established,3-5 therefore, identifying CNV before the onset of significant visual loss is crucial.
The natural history of CNV is such that it increases in size by an average of 10 to 18 µm daily;6,7 it grows closer to the foveal center in more than half of cases,7 and it causes deterioration of visual acuity.8
AMD TREATMENT TODAY
Antiangiogenic therapy has essentially replaced photodynamic therapy (PDT) as the AMD treatment of choice. Indirect evidence shows that earlier detection enhances the benefit of treatment. Data from the Treatment of AMD with Photodynamic Therapy study (TAP) and Verteporfin (Visudyne; Novartis, East Hanover, NJ) in Photodynamic Therapy (VIP)2 showed that the larger the lesion at baseline, the smaller the treatment benefit in terms of the absolute visual acuity 2 years posttreatment initiation. The VEGF Inhibition Study in Ocular Neovascularization (VISION) trial with pegaptanib (Macugen; OSI/Eyetech, New York, NY)3 revealed that therapeutic benefit in the patient subgroup with early, smaller lesions was superior to that for the overall population. In an article on early retreatment with verteporfin, Michels et al suggested that there was more vision loss with increasing baseline lesion size in the both the early retreatment and the standard retreatment group.
Ranibizumab (Lucentis; Genentech, South San Francisco) is a humanized antivascular endothelial growth factor (anti-VEGF) antibody fragment that inhibits VEGF by competitively binding with VEGF. Ranibizumab, derived from bevacizumab (Avastin; Genentech), is the first US Food and Drug Administration-approved treatment for neovascular AMD that maintains or improves vision in ≥90% of patients and improves visual acuity.9-11 In two pivotal phase 3 trials,9-11 ranibizumab was associated with a <1.7% rate of major serious ocular adverse events. Ranibizumab treatment, however, resulted in improvement in only 40% of patients with predominantly classic AMD11 and in only 34% of patients with minimally classic and occult AMD.
A recent subgroup analysis of the Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular AMD (MARINA) study showed that treatment benefit was greater among patients with higher baseline visual acuity and smaller lesions.12 Therefore, to preserve our patients' visual acuity, we need to detect lesions when visual acuity is still good.
WHY EARLY DETECTION?
For the first time, CNV treatment has resulted in an average improvement in visual acuity. With close to a nine-letter improvement on the Early Treatment Diabetic Retinopathy Study (ETDRS) chart, early detection's impact on visual acuity can be considerable. For an average patient, earlier detection means a significant gain of vision.
The economic impact of early detection can also be significant. Javitt et al showed that the yearly Medicare-only related costs for a patient with severe visual impairment is >$4,000.13 If approximately 50,000 patients are affected in their second eye every year and those patients have a 10-year life expectancy postdiagnosis, the annual savings for Medicare from early detection could be $2.25 billion. We should also not forget that CNV with its resultant visual loss causes significant deterioration in quality of life.14,15
How can we detect a developing lesion while it is small and visual acuity is still preserved? We must have a sensitive and specific monitoring tool, we need to increase the frequency at which we see at-risk patients, and we need to increase patient awareness.16 Haddad et al17 reported that only one third of AMD patients seek ophthalmologic examination within the first month of symptoms. This delay in presentation clearly causes delay in diagnosis and treatment, resulting in avoidable visual loss.18 It is also essential to increase awareness among health care providers of the importance of early detection and the need for prompt referral.
AMD DETECTION MODALITIES
One aspect of clinical evaluation involves instructing patients to report symptoms. Many investigators have reported that patients who notice symptoms are detected relatively late in the stage of the disease.19-21 The reasons for this are that the recognition of an existing scotoma is impeded by the cortical phenomenon of completion22 and crowding.23 Patients often are unable to maintain proper fixation during peripheral visual field testing, so they scan the world with their fovea and ignore retinal defects until the defects are close to the fovea or are large. The noninteractive nature of the Amsler grid renders it unsuitable for patient monitoring because certain factors such as quality of examination performance and reliability measures cannot be assessed.
An alternative would be to examine aging patients regularly to look for CNV. Although yearly or even biyearly follow-up exams would be helpful, it is unlikely that the timing of a visit would coincide with transformation from AMD to CNV.2 Even if the conversion had recently occurred, it can be difficult to diagnose with biomicroscopy alone.
Several tools can be used to monitor a patient at high risk for developing CNV. We could perform fluorescein angiography at each visit, but it is expensive, has side effects, and is time consuming. The clinical utility of biomicroscopy is variable and is highly examiner dependent. A nonmydriatic digital camera provides limited stereopsis, and the photographs are frequently nongradable, especially in the presence of small nonpharmacologically dilated pupils and/or media opacities.24,25 Optical coherence tomography (OCT) has emerged as an important means of detecting and monitoring patients with AMD by showing specific patterns in high-risk characteristics of nonneovascular AMD26 that are capable of assisting in the decision making of whether to retreat CNV.27,28 The reproducibility of OCT findings and their application to clinical practice, however, have yet to be demonstrated.
PREFERENTIAL HYPERACUITY PERIMETER
Our group recently demonstrated that the preferential hyperacuity perimeter (PHP) (Foresee; NotalVision, Tel Aviv, Israel) has potential for assisting in early detection of CNV in AMD.29-31 The technology is based on the phenomenon of hyperacuity, the ability we have to perceive minute differences in the relative spatial localization of two objects in space.29-31 The patient is presented with a pattern of dotted lines projected for 160 milliseconds to the central 14° of the visual field. Each line contains an artificial distortion at a different magnitude, and the distortion serves as a competitive stimulus to any pathologic distortion that might appear on the pattern. When there is a CNV lesion, attention competition between the artificial distortion and the pathologic distortion takes place in the patient's brain. In general, the brain ignores smaller stimuli when there is a larger one. This is true for everyday stimuli, and this phenomenon is exploited in PHP testing to assess the magnitude of the pathologic distortion: varying sizes of artificial distortion are presented, allowing quantification of the pathologic distortion by analyzing the patient's responses. Based on these responses, a visual field map is constructed, analyzed, and compared with normative data, thereby determining the likelihood of this defect being CNV.
The first PHP study29 was aimed at evaluating whether PHP is better than the Amsler grid under supervised office conditions. The results showed that PHP was more sensitive than the Amsler grid in each category of AMD, and that it was twice as sensitive for detecting CNV.29 To validate this study, which was performed in our facilities, we performed an international multicenter study. Although this study confirmed the results of the previous study, it also included a significant rate of false-positive responses.30 The algorithm was then amended. The PHP was further developed to be able to differentiate between patients at risk for CNV and patients with recent-onset CNV, to evaluate whether the PHP can be used as a tool for monitoring at-risk patients.
This investigation31 was an international multicenter study performed in the United States, Europe, and Asia, and it showed that PHP can detect recent-onset CNV with high sensitivity (82%) and differentiate these patients from those with intermediate AMD with high specificity (88%). We also investigated whether reviewing color stereo photographs (simulating biomicroscopy) was effective for monitoring AMD patients. All color stereo photographs of the study patients were graded in a masked manner. Analysis showed that grading was less sensitive than the PHP (sensitivity of 70% [95% CI, 58%-81%] compared with 82% for the PHP, P=.09, McNemar) and that the specificity was higher (specificity of 95% [95% CI, 85%-99%] compared with 88% for the PHP (P=.05, McNemar). CNV could easily be missed in real-life situations in which biomicroscopy is frequently performed by less skilled practitioners.
PHP does have limitations: Like any other diagnostic device, there is a learning curve in its use and in interpreting the visual field results. Exam scheduling in clinics will need to be altered to integrate PHP into the diagnostic protocol. PHP is, however, a promising tool for the early detection of CNV in the era of new anti-VEGF treatments.
Newly emerging treatment options have made it incumbent upon eye care professionals to upgrade monitoring techniques for earliest possible detection of CNV. Our patients must also be encouraged to be more vigilant to any changes in vision, however subtle, to halt the progression of AMD and preserve vision in our graying society. PHP use in the clinic will significantly enhance the rate of early detection of CNV development, leading to earlier and more effective treatment and better final visual acuity. Finally, similar to other medical disciplines where home monitoring has improved disease management,32 additional development of PHP technology for use at home can improve even further the rate of early detection and consequently enhance visual outcome following treatment.
This article was adapted from the Richard & Hinda Rosenthal Foundation Lecure given by Dr. Loewenstein at the Macula Society 29th Annual Meeting and published in Retina. 2007;27:873-878.
Anat Loewenstein, MD, is from the Department of Ophthalmology, Tel Aviv Sourasky Medical Center, Faculty Medicine, Tel Aviv University. Dr. Loewenstein may be reached at firstname.lastname@example.org.