Applications of Artificial Intelligence in the Diagnosis of Idiopathic Intracranial Hypertension: A Review of Neuroimaging and Machine Learning Approaches
Code: G-1937
Authors: Negin Safari Dehnavia , Mohammad Fayaz * ℗
Schedule: Not Scheduled!
Tag: Biomedical Signal Processing
Download: Download Poster
Abstract:
Abstract
Background and Aims: Idiopathic Intracranial Hypertension (IIH), also known as pseudotumor cerebri, is a complex neurological disorder characterized by elevated intracranial pressure without an identifiable cause. The accurate diagnosis of IIH is challenging due to the non-specific nature of its symptoms and reliance on invasive procedures like lumbar puncture. Advances in medical imaging, particularly MRI, CT, and optical coherence tomography (OCT), along with the integration of artificial intelligence (AI), offer new opportunities for non-invasive, accurate, and rapid diagnosis. Method: This review explores recent applications of machine learning (ML) and deep learning (DL) algorithms in the diagnosis and monitoring of IIH, focusing on neuroimaging modalities. A systematic analysis of existing models—including logistic regression, support vector machines (SVM), decision trees, random forests, k-nearest neighbors (KNN), discriminant analysis, and neural networks (CNNs and RNNs)—was conducted, with emphasis on their clinical relevance and diagnostic accuracy. Results: AI models demonstrated promising performance in classifying IIH-related abnormalities across various imaging modalities. In MRI and CT, CNNs and SVMs showed robust feature extraction and classification of papilledema and optic nerve sheath dilation. Logistic regression models were effective in associating imaging findings with clinical outcomes. OCT-based analysis revealed that CNNs, SVMs, and KNNs could accurately grade papilledema and predict Frisén scale values using retinal thickness parameters. Logistic and linear regression approaches also supported surgical decision-making and early detection through predictive modeling. Conclusion: AI, particularly when combined with high-resolution imaging, offers a powerful toolkit for the non-invasive diagnosis of IIH. While several models show clinical potential, further validation using larger, multi-center datasets and explainable AI approaches is needed to ensure real-world applicability and physician trust.
Keywords
Idiopathic Intracranial Hypertension, AI, ML
