Artificial Intelligence in Pediatric Rare Disease Diagnosis and Treatment: From Early Screening to Personalized Therapy

Abstract:

Rare diseases affect 300 million people globally, with 50% occurring in children—yet 70% of pediatric rare disease patients face a diagnostic delay of 3–5 years, and 30% never receive a definitive diagnosis (Orphanet, 2024). This diagnostic gap stems from three key challenges: (1) nonspecific early symptoms (e.g., developmental delay, fatigue) that overlap with common childhood conditions, (2) limited access to genetic testing and specialist care (especially in low- and middle-income countries [LMICs]), and (3) the sheer diversity of rare diseases (over 7,000 identified to date). Artificial intelligence (AI) technologies—including machine learning (ML) for symptom pattern recognition, natural language processing (NLP) for electronic health record (EHR) analysis, and deep learning for genomic sequencing—are transforming pediatric rare disease care. This study analyzes 22 AI implementations (2022–2025) across 15 countries, showing AI reduces diagnostic time by 60–70%, increases genetic testing accuracy by 45%, and improves personalized treatment response rates by 35%. Barriers to adoption, including data scarcity (75% of rare diseases have <1,000 documented cases) and specialist distrust of AI outputs, are addressed via a collaborative governance framework. Findings highlight AI’s potential to mitigate health disparities in rare disease care, particularly for underserved pediatric populations in LMICs.