Prediction of intracerebral hemorrhage hematoma expansion: value of a novel deep learning system score

Key Points

• Hematoma expansion risk was predicted with high accuracy using deep learning methods.
• The model achieved a sensitivity of 90% and specificity of 85% in predicting expansion events.
• Evaluation utilized a machine learning algorithm on clinical imaging data and patient histories.
• This predictive model could enhance clinical decision-making, but external validation is necessary.