Purpose: To evaluate a machine learning algorithm that automatically grades age-related macular degeneration (AMD) severity stages from optical coherence tomography (OCT) scans. Methods: A total of 3265 OCT scans from 1016 patients with either no signs of AMD or with signs of early, intermediate, or advanced AMD were randomly selected from a large European multicenter database. A machine learning system was developed to automatically grade unseen OCT scans into different AMD severity stages without requiring retinal layer segmentation. The ability of the system to identify high-risk AMD stages and to assign the correct severity stage was determined by using receiver operator characteristic (ROC) analysis and Cohen's κ statistics (κ), respectively. The results were compared to those of two human observers. Reproducibility was assessed in an independent, publicly available data set of 384 OCT scans. Results: The system achieved an area under the ROC curve of 0.980 with a sensitivity of 98.2% at a specificity of 91.2%. This compares favorably with the performance of human observers who achieved sensitivities of 97.0% and 99.4% at specificities of 89.7% and 87.2%, respectively. A good level of agreement with the reference was obtained (κ = 0.713) and was in concordance with the human observers (κ = 0.775 and κ = 0.755, respectively). Conclusions: A machine learning system capable of automatically grading OCT scans into AMD severity stages was developed and showed similar performance as human observers. The proposed automatic system allows for a quick and reliable grading of large quantities of OCT scans, which could increase the efficiency of large-scale AMD studies and pave the way for AMD screening using OCT.
The discussion on the use of bevacizumab is still ongoing and often doctors are deterred from using bevacizumab due to legal or political issues. Bevacizumab is an effective, safe and inexpensive treatment option for neovascular age-related macular degeneration (AMD), albeit unregistered for the disease. Therefore, in some countries ophthalmologists use the equally effective but expensive drugs ranibizumab and aflibercept. We describe the economic consequences of this dilemma surrounding AMD treatment from a societal perspective. We modelled cost-effectiveness of treatment with ranibizumab (as-needed), aflibercept (bimonthly) and bevacizumab (as-needed). Effectiveness was estimated by systematic review and meta-analysis. The drug with the most favourable cost-effectiveness profile compared to bevacizumab was used for threshold analyses. First, we determined how much we overspend per injection. Second, we calculated the required effectiveness to justify the current price and the reasonable price for a drug leading to optimal vision. Finally, we estimated how much Europe overspends if bevacizumab is not first choice. Bevacizumab treatment costs €27,087 per year, about €4,000 less than aflibercept and €6,000 less than ranibizumab. With similar effectiveness for all drugs as shown by meta-analysis, bevacizumab was the most cost-effective. Aflibercept was chosen for threshold analyses. Aflibercept costs €943 per injection, but we determined that the maximum price to be cost-effective is €533. Alternatively, at its current price, aflibercept should yield about twice the visual gain. Even when optimal vision can be achieved, the maximum price for any treatment is €37,453 per year. Most importantly, Europe overspends €335 million yearly on AMD treatment when choosing aflibercept over bevacizumab. Bevacizumab is the most cost-effective treatment for AMD, yet is not the standard of care across Europe. The registered drugs ranibizumab and aflibercept lead to large overspending without additional health benefits. Health authorities should consider taking steps to implement bevacizumab into clinical practice as first choice.
Age-related macular degeneration (AMD) is the leading cause of blindness in the Western world. AMD is a multifactorial disorder but complement-mediated inflammation at the level of the retina plays a pivotal role. Oral zinc supplementation can reduce the progression of AMD but the precise mechanism of this protective effect is as yet unclear. We investigated whether zinc supplementation directly affects the degree of complement activation in AMD and whether there is a relation between serum complement catabolism during zinc administration and the complement factor H (CFH) gene or the Age-Related Maculopathy susceptibility 2 (ARMS2) genotype. In this open-label clinical study, 72 randomly selected AMD patients in various stages of AMD received a daily supplement of 50 mg zinc sulphate and 1 mg cupric sulphate for three months. Serum complement catabolism–defined as the C3d/C3 ratio–was measured at baseline, throughout the three months of supplementation and after discontinuation of zinc administration. Additionally, downstream inhibition of complement catabolism was evaluated by measurement of anaphylatoxin C5a. Furthermore, we investigated the effect of zinc on complement activation in vitro. AMD patients with high levels of complement catabolism at baseline exhibited a steeper decline in serum complement activation (p<0.001) during the three month zinc supplementation period compared to patients with low complement levels. There was no significant association of change in complement catabolism and CFH and ARMS2 genotype. In vitro zinc sulphate directly inhibits complement catabolism in hemolytic assays and membrane attack complex (MAC) deposition on RPE cells. This study provides evidence that daily administration of 50 mg zinc sulphate can inhibit complement catabolism in AMD patients with increased complement activation. This could explain part of the mechanism by which zinc slows AMD progression. The Netherlands National Trial Register NTR2605