Toba S, Mitani S, Yodoya N, et al. - Experts intended to generate and authenticate a quantitative method to prognosticate the pulmonary to systemic flow ratio from chest radiographs using deep learning. Researchers designed a retrospective observational study including 1,031 cardiac catheterizations for 657 individuals between January 1, 2005, and April 30, 2019, at a tertiary center. They randomized 78 individuals (100 catheterizations) for evaluation. They established a deep learning model that predicts the pulmonary to systemic flow ratio from chest radiographs applying the method of transfer learning. They compared the diagnostic concordance rates with 3 certified pediatric cardiologists. They analyzed diagnostic performance for a high pulmonary to systemic flow ratio of 2.0 or more applying cross-tabulation and a receiver operating characteristic curve. The present study showed that deep learning-based examination of chest radiographs prognosticated the pulmonary to systemic flow ratio in individuals with congenital heart disease. These data demonstrate that the deep learning-based method may confer an objective and quantitative evaluation of chest radiographs in the congenital heart disease clinic.