Classification Criteria for Fuchs Uveitis Syndrome

center dot PURPOSE: To determine classification criteria for Fuchs' uveitis syndrome. center dot DESIGN: Machine learning of cases with Fuchs' uveitis syndrome and 8 other anterior uveitides. center dot METHODS: Cases of anterior uveitides were collected in an informatics-designed preliminary database, and a final database was constructed of cases achieving supermajor-ity agreement on the diagnosis, using formal consensus techniques. Cases were split into a training set and a val-idation set. Machine learning using multinomial logistic regression was used on the training set to determine a parsimonious set of criteria that minimized the misclassi-fication rate among the anterior uveitides. The resulting criteria were evaluated on the validation set. center dot RESULTS: One thousand eighty-three cases of ante-rior uveitides, including 146 cases of Fuchs' uveitis syn-drome, were evaluated by machine learning. The overall accuracy for anterior uveitides was 97.5% in the train-ing set and 96.7% in the validation set (95% confidence interval 92.4, 98.6). Key criteria for Fuchs' uveitis syn-drome included unilateral anterior uveitis with or with-out vitritis and either: 1) heterochromia or 2) unilateral diffuse iris atrophy and stellate keratic precipitates. The misclassification rates for Fuchs' uveitis syndrome were 4.7% in the training set and 5.5% in the validation set, respectively. center dot CONCLUSIONS: The criteria for Fuchs' uveitis syn-drome had a low misclassification rate and appeared to perform well enough for use in clinical and translational research. (Am J Ophthalmol 2021;228: 262-267. (c) 2021 Elsevier Inc. All rights reserved.)