Using machine learning to detect hypoxic-ischemic brain injury after cardiac arrest

Takeaway

  • In adult comatose survivors of cardiac arrest, deep transfer learning accurately identified hypoxic-ischemic brain injury (HIBI) from initial head computed tomography (HCT) scans that were deemed normal.

Why this matters

  • Although rates of resuscitation after cardiac arrest have improved, many survivors have severe neurological disability as a result of HIBI.

  • Earlier identification of patients who develop HIBI will aid in therapeutic decision making.