Large Language Models-Assisted Diagnosis of Catecholaminergic Polymorphic Ventricular Tachycardia in a Pediatric Cardiac Arrest Patient

Background : Catecholaminergic polymorphic ventricular tachycardia (CPVT), a rare hereditary ion channel disorder, is triggered by exercise or stress, causing PVT and sudden death. Diagnosis is tough, especially with cardiac arrest as the initial symptom, and guideline - recommended adrenaline may worsen it. Large language models offers new ways to identify it quickly. Case Presentation : A 7 - year - old boy had sudden cardiac arrest during rope - skipping. After resuscitation, defibrillation, and adrenaline, his circulation returned, but PVT persisted. VA - ECMO in our hospital couldn't control the arrhythmia. Multidisciplinary discussion was inconclusive. ChatGPT and DeepSeek suggested CPVT. After stopping catecholamines and using beta - blockers, arrhythmias decreased. Gene testing confirmed an RYR2 gene mutation (c.6737C>T), diagnosing CPVT. However, due to long - term resuscitation, late diagnosis, and delayed ECMO, the child developed severe complications. Despite successful ECMO weaning, the parents gave up treatment, and the child died. Conclusion: CPVT patients with cardiac arrest are critically ill and hard to diagnose. Early detection and targeted treatment are vital for prognosis. Large language models have value in CPVT diagnosis but should be combined with clinical judgment and further tests. For children with unexplained cardiac arrest, Large language models - assisted consultation can be considered, and clinicians should better understand rare diseases like CPVT for more timely and accurate diagnoses. Clinical trial number No applicable.