Katsuyuki KUNIDA

Background: Fulminant myocarditis (FM) is a rare but life-threatening pediatric condition that rapidly progresses to cardiogenic shock and fatal arrhythmias. Early identification of prognostic biomarkers is vital for timely intervention and better outcomes. Although inflammatory cytokines contribute to FM pathogenesis, their prognostic value remains unclear. This study aimed to identify mortality-associated markers by integrating cytokine profiles and clinical variables through a machine learning approach.Methods: We retrospectively analyzed 21 pediatric FM cases from two tertiary centers (2012-2022). At admission, 37 cytokines and 14 clinical parameters were assessed. Partial least squares discriminant analysis was employed to identify prognostic features, with variable importance in projection scores quantifying their contribution. Model performance was evaluated using leave-one-out cross-validation. Statistical significance was determined via the Benjamini-Hochberg method at a false discovery rate of 0.05.Results: Of the 51 features analyzed, 23 emerged as key predictors with variable importance in projection scores above 1.0, including 20 cytokines and three clinical parameters. Six cytokines (TNFｰα, M-CSF, MIP-1α, IL-8, IL-6, and IL-15) were both statistically significant and highly important. Elevated CK-MB and lactate levels and lower pH were also linked to poor outcomes. The model performed robustly, with an AUC of 0.92, 85.7% accuracy, 92.9% sensitivity, and 71.4% specificity.Conclusions: TNF-α emerged as a key cytokine linked to mortality in pediatric FM, supporting its role as a prognostic biomarker and potential therapeutic target.

Abstract Recent advancements in machine learning-based data processing techniques have facilitated the inference of gene regulatory interactions and the identification of key genes from multidimensional gene expression data. In this study, we applied a stepwise Bayesian framework to uncover a novel regulatory component involved in differentiation of specific neural and neuronal cells. We treated naive neural precursor cells with Sonic Hedgehog (Shh) at various concentrations and time points, generating comprehensive whole-genome sequencing data that captured dynamic gene expression profiles during differentiation. The genes were categorized into 224 subsets based on their expression profiles, and the relationships between these subsets were extrapolated. To accurately predict gene regulation among subsets, known networks were used as a core model and subsets to be added were tested stepwise. This approach led to the identification of a novel component involved in neural tube patterning within gene regulatory networks (GRNs), which was experimentally validated. Our study highlights the effectiveness of in silico modeling for extrapolating GRNs during neural development.