Retinal Physician

JAN-FEB 2017

Issue link:

Contents of this Issue


Page 18 of 71

MMP9 (matrix metalloproteinase-9) was recently identified as associated only with neovascular AMD. To date, this is the first neovascular-specific allele that has been identified. 1 COM M ERCIALLY AVAIL ABLE GENETIC TESTS FOR A M D There are several commercially available genetic tests for AMD. In the United States, Macula Risk PGx (ArcticDx, Toronto) can be ordered by a provider. RetnaGene (Sequenom, San Diego) is no longer avail- able. Asper Biotech based in Tartu, Estonia, offers direct-to-consumer genetic testing worldwide. Prior to 2010, direct-to-consumer genetic AMD tests were an emerging market segment for the labora- tory testing industry. Warning letters, however, issued by the FDA that year to Pathway Genomics Corporation (San Diego), deCODE Genetics (Reykjavik, Iceland), 23andMe (Mountain View, CA), and others stated that the offered tests met the definition of a medical device and, therefore, required FDA clearance or approval in order to be marketed to US consumers. The FDA argued that the tests had not been proved safe, effec- tive, or accurate, and that patients could be put at risk by making medical decisions based on data that had not received independent market review. Since that time, the above tests have been formally withdrawn from the direct-to-consumer marketplace, but a genetic AMD profile can still be obtained by the consumer indirectly from Promethease, a retrieval sys- tem that builds a personal DNA report based on con- necting a file of DNA gene reports (from 23andMe and other US-based companies) to the scientific find- ings cited in SNPedia, a wiki for human genetics. A macula risk test, ordered by an ophthalmologist, analyzes 15 variants across 12 loci and uses patient data such as age, body mass index, smoking history, and educational level, to stratify the patient into one of five "macula risk" categories that are associated with vary- ing degrees of risk of progression to advanced AMD. The company offers a second test, " Vita Risk," which uses variants at CFH and ARMS2 in an effort to pre- dict response to nutritional therapy. In 2002, a task force comprised of CDC and NIH researchers created the "ACCE" model to evaluate the utility of various genetic tests. 12 The model considers four variables. Analytic validity measures the accuracy (sensitivity and specificity) with which the genetic information is detected. Clinical validity measures the extent to which the genetic test predicts the clinical phenotype. Clinical utility measures the ability of the test to improve clinical outcomes. The last variable describes the ethical, legal, and social implications of the test. Applying the ACCE model to the Macula Risk test reveals that the test demonstrates analytic validity because it is technically accurate. One measure of the clinical validity of a predictive model uses area under the curve (AUC), in which an AUC of 0.5 or less indi- cates chance (no accuracy), an AUC of 1 is completely accurate, and an AUC of 0.75 or greater suggests a use- ful model. 13 Models based on either clinical or genetic information may meet this definition of usefulness. For example, some clinical models with no genetic infor- mation have achieved AUCs of 0.76 or greater. 14,15 Alternatively, some purely genetic risk models have achieved AUCs of 0.81 or greater. 16,17 It is reasonable to suspect that models combining clinical and genetic data may yield even more accurate information, but this is not always the case. Combined models have reported AUCs of 0.75 or higher. 18-24 The Macula Risk test is based on a statistical model with a reported AUC of 0.883 for five-year progression and 0.895 for 10-year progression, which suggests good clinical validity. 25 However, one model incorporating only clinical data reported a similar AUC of 0.88 15, suggesting that comparable validity may be achieved without any genetic analysis. The clinical utility of the Macula Risk test is more controversial. One might recommend more frequent examinations for higher-risk patients, but there is little or no peer-reviewed evidence to support such a strategy. The utility of genetic testing for predicting response to nutritional supplementation is discussed in the next section. The ethical, legal, and social implications of genetic testing may be considerable, especially if a genetic test is requested by a young, asymptomatic patient. For these reasons, the American Academy of Ophthalmology 17 R E T I N A L P H Y S I C I A N . C O M | J A N U A R Y / F E B R U A R Y 2 0 1 7 Figure 2. Fundus photo of the le macula showing dry AMD with geographic atrophy, pigmentary changes, and reticular pseudodrusen. P E E R R E V I E W E D

Articles in this issue

Archives of this issue

view archives of Retinal Physician - JAN-FEB 2017