Scholarlink

Purpose <p>Idiopathic intracranial hypertension (IIH) is a condition characterized by increased intracranial pressure without an identifiable cause. This study aimed to assess the efficacy of artificial intelligence (AI)-based algorithms in diagnosing IIH using magnetic resonance imaging (MRI) and magnetic resonance venography (MRV) data.</p> Materials and Methods <p>A total of 194 individuals (74 IHH patients and 120 controls) were examined, and brain MRV and T2-weighted MRI images were analyzed. Two different models were trained using AI algorithms on MRV and T2 images, and these models were evaluated together.</p> Results <p>The accuracy of the models was calculated to be 80.60% ± 9.17 using only MRV data, 79.86% ± 8.50 using only T2 data, and 82.6% ± 8.88 when both datasets were combined, and the area under the curve values were 84.97% ± 9.24, 85.46% ± 8.00, and 90.55% ± 8.53, respectively.</p> Conclusion <p>The results indicate that AI algorithms offer high accuracy and reliability in diagnosing IIH. AI-assisted diagnostic methods can accelerate clinical processes and eliminate observer-related variations.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Automated Detection of Idiopathic Intracranial Hypertension Using Artificial Intelligence: An Approach Based on Magnetic Resonance Imaging and Magnetic Resonance Venography Data

  • 期刊论文

  • 发表日期：

    2026年2月10日

查看全文
  • Habip Eser Akkaya,
  • Onder Polat,
  • Ugur Kesimal,
  • Dorukan Inanc Akpinar,
  • Zeynep Ziroglu,
  • Hafize Nalan Gunes

摘要

Purpose

Idiopathic intracranial hypertension (IIH) is a condition characterized by increased intracranial pressure without an identifiable cause. This study aimed to assess the efficacy of artificial intelligence (AI)-based algorithms in diagnosing IIH using magnetic resonance imaging (MRI) and magnetic resonance venography (MRV) data.

Materials and Methods

A total of 194 individuals (74 IHH patients and 120 controls) were examined, and brain MRV and T2-weighted MRI images were analyzed. Two different models were trained using AI algorithms on MRV and T2 images, and these models were evaluated together.

Results

The accuracy of the models was calculated to be 80.60% ± 9.17 using only MRV data, 79.86% ± 8.50 using only T2 data, and 82.6% ± 8.88 when both datasets were combined, and the area under the curve values were 84.97% ± 9.24, 85.46% ± 8.00, and 90.55% ± 8.53, respectively.

Conclusion

The results indicate that AI algorithms offer high accuracy and reliability in diagnosing IIH. AI-assisted diagnostic methods can accelerate clinical processes and eliminate observer-related variations.