Back to all papers

Deep Learning Models for Radiomics-Based Segmentation of Vestibular Schwannoma on Magnetic Resonance Imaging: A Systematic Review and Meta-analysis.

November 17, 2025pubmed logopapers

Authors

Hajikarimloo B,Mohammadzadeh I,Shirzadi P,Tos SM,Mortezaei A,Habibi MA

Affiliations (6)

  • Department of Neurological Surgery, University of Virginia, Charlottesville, VA, USA. [email protected].
  • Skull Base Research Center, Loghman-Hakim Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran.
  • Department of Neurological Surgery, Shohada Tajrish Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran.
  • Department of Neurological Surgery, University of Virginia, Charlottesville, VA, USA.
  • Student Research Committee, Gonabad University of Medical Sciences, Gonabad, Iran.
  • Department of Neurosurgery, Shariati Hospital, Tehran University of Medical Sciences, Tehran, Iran.

Abstract

Precise segmentation of the vestibular schwannoma (VS) is essential to optimize therapeutic strategies and enhance outcomes. While manual segmentation is associated with considerable time consumption and interobserver variability, deep learning (DL)-based models can provide faster, more precise segmentation of VS lesions. This study aimed to evaluate the performance of the DL-based models in VS segmentation. A systematic literature search was conducted on May 8, 2025. Studies that developed DL-based models to perform VS segmentation based on the magnetic resonance imaging (MRI)-driven radiomics and reported the mean Dice Similarity Coefficient (DSC) were included. Forty-one studies involving 8028 VS cases were included. The mean DSC ranged from 0.75 to 0.99 across the included studies. The meta-analysis revealed a pooled DSC of 0.89 (95% CI: 0.88-0.91) for the best-performing models. The sensitivity analysis demonstrated that the results were robust and consistent. No significant publication bias was observed. DL-based models have demonstrated encouraging performance for VS segmentation using MRI-driven radiomics. The application of these models in daily clinical workflows can optimize therapeutic stages and enhance patient outcomes.

Topics

Journal ArticleReview

Ready to Sharpen Your Edge?

Join hundreds of your peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.

We respect your privacy. Unsubscribe at any time.