AI misdiagnosis delays treatment for patient with brain infarction
A 52-year-old patient in China used an AI diagnostic platform to self-diagnose dizziness as vertigo, delaying treatment for three days until severe symptoms prompted hospital admission, where he was found to have a large cerebellar infarction. Hospital experts noted the AI system's inability to perform proper clinical assessment and advised against using AI tools as a substitute for professional medical diagnosis.
What happened
On June 18, 2026, 52-year-old Zhang Yunxiang experienced sudden dizziness at home. Without typical stroke symptoms such as limb numbness, weakness, or facial drooping, he consulted an AI platform to search for symptoms and treatment methods, which led him to self-diagnose as suffering from vertigo.
Three days later, his dizziness worsened and he developed unsteady gait and lethargy. His family took him to Changsha Central Hospital, where imaging revealed left cerebellar hemisphere large infarction with cerebral edema and mass effect, indicating high risk of brain herniation.
Reported impact
The patient's condition was critical upon hospital admission but stabilized with active treatment. Hospital expert Dr. Hu Jue stated that acute vestibular syndrome, which can present as simple dizziness, accounts for approximately 25% of posterior circulation stroke cases, and some cerebellar infarction patients may not exhibit typical early stroke symptoms, making misidentification as benign vertigo common.
Dr. Hu emphasized that delayed recognition of cerebellar infarction can cause patients to miss intravenous thrombolysis and early intervention for posterior fossa edema, critical time windows in stroke treatment.
Index analysis
AI diagnostic platforms rely on patient-entered information and cannot perform face-to-face history taking, physical examination, neurological assessment, and imaging interpretation as physicians do. For acute symptoms such as dizziness, chest pain, abdominal pain, or altered consciousness that may indicate life-threatening conditions, text-based descriptions are difficult to assess accurately for severity.
Governance controls
Hospital experts stated that AI can serve as a health knowledge reference tool but cannot replace professional medical diagnosis and treatment. Dr. Hu cautioned that patients should not self-diagnose or self-medicate based on AI suggestions, nor delay seeking medical care as a result.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.