DEVELOPING CLASSIFICATION CRITERIA FOR THE UVEITIDES

 描述（由申请人提供）：葡萄膜炎是以眼内感染为特征的约30种不同疾病的集合。每种疾病都有其自身的特点、病程、治疗和预后。传统上，葡萄膜炎根据炎症的原发解剖部位分为前葡萄膜炎、中间葡萄膜炎、后葡萄膜炎和全葡萄膜炎。然而，在这方面， 对疾病的这种“归类”有很大的局限性。例如，在后葡萄膜炎中，一些（例如弓形体视网膜炎和巨细胞病毒视网膜炎）是感染性的并且需要用抗微生物剂/抗病毒剂治疗，一些是需要免疫抑制的慢性、假定的免疫介导的疾病（例如鸟射性脉络膜视网膜炎、多灶性脉络膜炎、蜿蜒性脉络膜炎），并且少数是自限性、自发缓解的疾病 预后良好（如急性后部多灶性片状色素上皮病和多发性一过性白色斑点综合征）。因此，精确的诊断对于研究至关重要，包括流行病学，转化发病机制研究，结果研究和疾病特异性临床试验。分类标准是一种用于研究目的的“诊断”标准。尽管分类标准寻求优化灵敏度和特异性，但当需要权衡时，它们强调特异性，以确保正在研究的患者组是同质的。对于复杂疾病的基因组风险因素研究和转化发病机制研究，特别需要精确的表型，因为包含具有不同风险因素和疾病机制的其他疾病会混淆结果。目前还没有广泛接受和验证的分类标准的任何葡萄膜炎。初步数据表明，葡萄膜炎专家对任何一个病例的独立诊断最多只能达到“一般至中等”的一致性（κ值为0.27-0.40），但委员会对>98%的病例的诊断达成一致的能力。“为葡萄膜炎命名标准化（SUN）工作组制定葡萄膜炎分类标准”项目的目标是使用正式、严格的方法为25种主要葡萄膜炎制定分类标准。该项目有4个阶段：1）信息学，开发标准化术语; 2）病例收集，开发每种疾病约250例的初步数据库; 3）病例选择，选择通常被认为是该疾病的每种疾病至少150-200例（使用正式共识技术）从初步数据库到最终数据库;以及4）使用机器学习方法对最终数据库进行数据分析，以针对每种疾病开发一组最小化错误分类的简约标准。该项目的信息学和案例收集阶段已经完成。病例选择阶段正在顺利进行，并使用在线投票和协商一致电话会议，以实现对最终数据库中所有病例的绝大多数接受。该应用程序的目标是完成病例选择和数据分析，并为25种主要葡萄膜炎制定分类标准。这些结果对葡萄膜炎领域的未来临床研究至关重要。

 DESCRIPTION (provided by applicant): The uveitides are a collection of ~30 distinct diseases characterized by intraocular infection. Each disease has its own features, course, treatment, and prognosis. Traditionally, the uveitides have been grouped by the primary anatomic site of inflammation as anterior uveitis, intermediate uveitis, posterior uveitis, and panuveitis. However, there are substantial limitations to this "lumping" of diseases. For example, among the posterior uveitides, some (e.g. toxoplasmic retinitis and cytomegalovirus retinitis) are infectious and require treatment with antimicrobial/antiviral agents, some are chronic, presumed immune-mediated diseases that require immunosuppression (e.g. birdshot chorioretinitis, multifocal choroiditis, serpiginous choroiditis), and a few are self-limited, spontaneously-remitting diseases with a good prognosis (e.g. acute posterior multifocal placoid pigment epitheliopathy and multiple evanescent white dot syndrome). As such precise diagnosis is critical for research, including epidemiology, translational pathogenesis research, outcomes research, and disease specific clinical trials. Classification criteria are a type of "diagnostic" criteria used for reserch purposes. Although classification criteria seek to optimize sensitivity and specificity, when a trade-off is required, they emphasize specificity in order to ensure that a homogeneous group of patients is being studied. A precise phenotype is required particularly for genomic risk factor studies of complex disorders and translational pathogenesis research, as inclusion of other diseases with different risk factors and disease mechanisms would confound the results. Currently there are no widely-accepted and validated classification criteria for any of the uveitides. Preliminary data indicate "fair to moderate" agreement at best on the independent diagnosis of any one case by uveitis experts (κ's 0.27-0.40), but the ability of committees to reach agreement on the diagnosis of >98% of cases. The goal of the "Developing Classification Criteria for the Uveitides" project is for the Standardization of Uveitis Nomenclature (SUN) Working Group to develop classification criteria for the 25 leading uveitides using a formal, rigorous approach. There are 4 phases to the project: 1) informatics, to develop a standardized terminology; 2) case collection, to develop a preliminary database of ~250 cases of each disease; 3) case selection, to select at least 150-200 cases of each disease that are generally accepted to be the disease (using formal consensus techniques) from the preliminary database into a final database; and 4) data analysis, using machine learning approaches, of the final database to develop a parsimonious set of criteria for each disease that minimizes misclassification. The informatics and case collection phases of the Project are complete. The case selection phase is well underway and uses online voting and consensus conference calls to achieve supermajority acceptance on all cases included in the final database. The goals of this application are to complete case selection and data analysis and develop classification criteria for the 25 of the major uveitides. These results are crucial to future clinical research i the field of uveitis.