Sub-Phenotyping of Pediatric Celiac Disease with Topological Data Analysis

Celiac Disease (CD) is an autoimmune disorder triggered by gluten ingestion in genetically susceptible individuals. Its heterogeneous presentation, especially in pediatric patients, limits the effectiveness of current symptom-based classifications such as the Oslo definitions. This study proposes a framework based on Topological Data Analysis (TDA) to identify new, clinically meaningful sub-phenotypes of pediatric CD. We used a multicentric dataset of over 3,000 children. Compared with standard clustering (DBSCAN, Agglomerative, K-Medoids), our TDA Mapper revealed stable, interpretable communities reflecting serological and clinical patterns. Results indicate that serology, histology, and comorbidities jointly define distinct pediatric phenotypes, supporting data-driven approaches for precision medicine in CD.