The application of the facial analysis program Face2Gene in a single genetic counseling center: a retrospective study

Abstract Background Face2Gene (F2G) is a software program, widely used in clinical genetics and dysmorphology for recognizing children with genetic disorders. In order to assess its accuracy in real clinical context, we applied F2G to patients from our genetic counseling center. Results We randomly selected 151 children, who were referred to genetic counseling because of dysmorphic features and later diagnosed with a particular genetic condition. A frontal photograph was uploaded to the program, and at least three phenotypic traits were selected for each case. Date of birth, sex, weight, height, and head circumference were also entered in the software. The efficacy of the program to correctly diagnose the syndrome based on the facial analysis and/or phenotypic traits was assessed. In 56% (84 cases) of the patients, the F2G program accurately identified the correct diagnosis in its top three suggestions. Forty-seven percent (71 cases) of the patients had the correct diagnosis after applying only facial analysis. There was a statistically significant difference between the two types of analysis—p = 0.001. In 19 of the cases where F2G was unable to identify the correct diagnosis among the top three options based on phenotypic and facial analysis, the diagnosis was included among the 30 suggested syndromes, yielding a total success rate of 68%. The diagnosis was found in the ultra-rare syndromes’ suggestions area in six more cases. Conclusions Our results show that F2G has a good overall performance, but adding phenotypic features to the case under study may increase even further its diagnostic capacity.