Research on predicting risk factors for re-bleeding in the acute phase of intracerebral hemorrhage using machine learning algorithms - Takeaways - MDSpire

Research on predicting risk factors for re-bleeding in the acute phase of intracerebral hemorrhage using machine learning algorithms

  • By

  • Xiong Deng

  • JieYao Xia

  • ZhiJun Liang

  • Can Luo

  • June 3, 2026

  • 0 min

Share

  • 1

    A retrospective analysis of 368 patients with intracerebral hemorrhage identified significant risk factors for rebleeding during the acute phase.

  • 2

    Lasso regression identified six key variables, including hematoma morphology and GCS score, for predicting rebleeding risk.

  • 3

    The GBDT model achieved the highest AUC of 0.967 in validation, indicating strong predictive performance for rebleeding risk.

  • 4

    The final 10-variable GBDT model showed an AUC of 0.85 on the test set, demonstrating stable performance through bootstrap validation.

  • 5

    SHAP values provided insights into variable importance, with HII, GCS, and age being the top predictors of rebleeding risk.

Original Source(s)

Research on predicting risk factors for re-bleeding in the acute phase of intracerebral hemorrhage using machine learning algorithms

  • by Xiong Deng, JieYao Xia, ZhiJun Liang, Can Luo

  • June 3, 2026

Related Content

Differential associations of novel metaflammation indices with incident stroke across traditional Chinese medicine constitutions: a retrospective community-based cohort study of older adults

Chemotherapy-related cognitive impairment and non-pharmacological interventions targeting the nervous system: a systematic review

Transcervical carotid artery stenting compared to transfemoral carotid artery stenting and carotid endarterectomy: perioperative and short-term results from a single center