Clinicians' confidence in synthetic neuroimaging: a survey from the EANS emerging Technologies and Innovations in Neurosurgery (ETIN) taskforce
Related Experiment Video
Updated: Oct 31, 2026
A Pipeline for 3D Multimodality Image Integration and Computer-assisted Planning in Epilepsy Surgery
Published on: May 20, 2016
Michelle D Poelman1,2, Massimo Bottini2, Olivier Zanier2
1Department of Neurosurgery, Park Medical Center, Rotterdam, the Netherlands.
Introduction:
Neurosurgery relies heavily on imaging for diagnostic, planning, intra-operative decision-making, and image-guided interventions. Machine learning (ML) enables synthetic imaging, such as synthetic CT (sCT) generated from MRI or biplanar X-rays, potentially reducing radiation exposure and improving imaging access. However, clinicians' confidence in synthetic imaging remains poorly understood.
Research Question:
This study assessed respondents attitudes towards synthetic neuroimaging.
Material And Methods:
An international cross-sectional survey was distributed via the European Association of Neurosurgical Societies (EANS) and a global neurosurgeon database. Nineteen questions covered demographics, general attitudes, and two algorithms: MRI-based sCT, and biplanar X-ray-based sCT. Confidence was reported using a 5-point Likert scale (1 = not confident, 5 = completely confident). Responses were summarized using medians, interquartile ranges (IQR), counts, and percentages.
Results:
Seventy-two respondents completed the survey; 61% were board-certified neurosurgeons, and 76% worked in academic hospitals. Confidence varied across the two clinical scenarios. Confidence for MRI-based spinal decision-making was moderate (median 3, IQR 3-4), whereas confidence in X-ray-based sCT for cranial applications was lower (median 2, IQR 2-3). Clinical and external validation were frequently selected as factors supporting confidence, whereas hallucinations and lack of validation limited confidence.
Discussion And Conclusion:
Confidence in synthetic neuroimaging varied across clinical scenarios. Differences may reflect clinical context, perceived consequences of error, input modality, and information provided in each scenario. While synthetic imaging may improve access, reduce radiation exposure, and support decision-making, adoption requires robust, application-specific validation. These findings guide future implementation of synthetic imaging in neurosurgery.
