Predicting Adverse Outcomes in Neurosurgical Patients Using Machine Learning: Exploring the Impact of Liver Function Indicators - Takeaways - MDSpire

Predicting Adverse Outcomes in Neurosurgical Patients Using Machine Learning: Exploring the Impact of Liver Function Indicators

  • By

  • Yibo Fan

  • Jingyue Zhang

  • Lin Wu

  • Shuo An

  • Yingsheng Wei

  • Jian Sun

  • Ye Tian

  • Hanxu Zhang

  • April 29, 2026

  • 0 min

Share

  • 1

    A machine learning model was developed to predict unfavorable outcomes in neurosurgical patients using liver function indicators.

  • 2

    The CatBoost model achieved an AUC of 0.932, demonstrating strong predictive performance for adverse outcomes.

  • 3

    Key predictive features included Glasgow Coma Scale score, age, and liver function markers such as AST and albumin.

  • 4

    Lower GCS scores and older age were associated with worse outcomes, while higher AST and ALKP levels indicated increased risk.

  • 5

    Future research should validate these findings across multiple centers and investigate the liver-brain axis in neurosurgical prognosis.

Original Source(s)

Predicting Adverse Outcomes in Neurosurgical Patients Using Machine Learning: Exploring the Impact of Liver Function Indicators

  • by Yibo Fan, Jingyue Zhang, Lin Wu, Shuo An, Yingsheng Wei, Jian Sun, Ye Tian, Hanxu Zhang

  • April 29, 2026

Related Content

Defining the target for hemifacial spasm –regarding microvascular decompression

Impact of onset timing and operative duration on outcomes in chronic subdural hematoma: an exploratory retrospective study

CT Perfusion Detects Tissue at Risk in Mild Stroke

Perfusion imaging adds prognostic information in patients with low National Institutes of Health Stroke Scale scores.