Related Experiment Video
Updated: Jun 13, 2026
14:08
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
Clinical Subgroups and Treatment Outcomes in Idiopathic Normal Pressure Hydrocephalus: Application of Machine
Emalee J Burrows1,2, Linda D'Antona1,2, James Booker1,3
1UCL Queen Square Institute of Neurology, London, UK.
Neurosurgery Practice
|June 12, 2026
Summary
Artificial intelligence identified four distinct patient subgroups for idiopathic normal pressure hydrocephalus (iNPH), revealing varied outcomes after shunt surgery. This challenges the single-disease view and aids personalized iNPH management.
Area of Science:
- Neurology
- Radiology
- Artificial Intelligence
Background:
- Idiopathic normal pressure hydrocephalus (iNPH) presents diagnostic and management challenges due to variable clinical phenotypes, radiological features, and treatment outcomes.
- Unsupervised machine learning is crucial for dissecting complex neurological conditions like iNPH.
Purpose of the Study:
- To employ unsupervised machine learning clustering to identify distinct clinical subgroups within iNPH patients.
- To analyze outcome trajectories following ventriculoperitoneal shunt insertion in these identified subgroups.
Main Methods:
- A retrospective single-center case series analyzed 187 patients with shunt-responsive iNPH.
- Clustering methods were applied to demographic, clinical (preoperative symptoms, postoperative shunt response, long-term outcomes), and radiological variables.
- Radiological data were assessed by senior neuro-radiologists blinded to clinical information.
Main Results:
- Four distinct patient subgroups with differing clinical outcome trajectories were identified using Ward's hierarchical agglomerative clustering.
- These subgroups included: disproportionately enlarged subarachnoid space hydrocephalus-positive, dilated sylvian fissure hydrocephalus, small-vessel disease with ventriculomegaly, and marked ventriculomegaly.
- The mean patient age was 75.6 years, with a median follow-up of 28 months.
Conclusions:
- Findings suggest iNPH is not a single entity, offering new insights into its pathophysiology.
- While disproportionately enlarged subarachnoid space hydrocephalus indicates shunt responsiveness, other phenotypes are critical for comprehensive assessment.
- Cluster-based phenotyping can enhance prognostic prediction, patient counseling, and targeted management strategies for iNPH.
