Related Experiment Video
Updated: Jun 26, 2026
High-resolution Functional Magnetic Resonance Imaging Methods for Human Midbrain
Published on: May 10, 2012
Low-field strength MRI (0.55T) for stereotactic and functional neurosurgery using deep learning-based reconstruction
Thomas Kinfe1, Miriam Ratliff2, Andreas Stadlbauer2
1From the Mannheim Center for Neuromodulation and Neuroprosthetics (MCNN), Department of Neurosurgery (T.K., M.R., S.B.), Mannheim Center for Translational Neuroscience (MCTN)(T.K.), Mannheim Comprehensive Medical Systems Technology Campus (MCSC)(T.K., S.S.), Department of Radiology (S.S.), Medical Faculty Mannheim, Heidelberg University, Germany; Institute of Medical Radiology (A.S.), University Clinic St. Pölten, Karl Landsteiner University of Health Sciences, St. Pölten, Austria; Department of Neurosurgery (J.M.), SUNY Upstate Medical University, Syracuse, New York; Department of Neurosurgery (S.R.), Clinical Neurosciences Center, University of Utah, Salt Lake City, Utah, USA; Siemens Healthineers (H.-P.F., T.V.), Erlangen, Germany and Department of Radiology (M.U.), University Hospital Erlangen, Erlangen, Germany. thomas.kinfe@medma.uni-heidelberg.de.
Low-field 0.55 T MRI, enhanced by deep learning, shows feasibility for visualizing critical anatomical landmarks in stereotactic neurosurgery. This approach offers potential advantages over high-field MRI, warranting further clinical investigation.
Area of Science:
- Neurosurgery
- Radiology
- Medical Imaging
Background:
- High-field MRI (1.5 T+) is trending for stereotactic neurosurgery.
- Low-field (0.55 T) MRI offers potential benefits in availability, cost, and reduced artifact.
- Clinical utility of 0.55 T MRI for stereotactic neurosurgery remains largely uncharacterized.
Purpose of the Study:
- To evaluate the feasibility and effectiveness of low-field 0.55 T MRI with deep learning reconstruction for stereotactic neurosurgery.
- To compare image quality and visualization of key anatomical structures across different MRI field strengths (0.55 T, 1.5 T, 3.0 T).
Main Methods:
- Optimized MRI parameters using a deep learning algorithm (Deep Resolve Boost) across four protocols.
- Acquired 0.55 T, 1.5 T, and 3.0 T MRI scans in healthy adults.
- Assessed image resolution, contrast, and visualization of stereotactic neurosurgery landmarks by blinded investigators.
Main Results:
- Higher field strengths correlated with higher resolution and shorter scan times.
- Image quality was rated similarly across all field strengths (ICC=0.947), indicating excellent rater agreement.
- The 0.55 T MRI protocol with deep learning enabled sufficient visualization of all relevant anatomical structures in all planes.
Conclusions:
- Preliminary findings indicate 0.55 T MRI is feasible for visualizing key stereotactic anatomical landmarks in healthy subjects.
- Deep learning-enhanced 0.55 T MRI demonstrates potential for stereotactic neurosurgery applications.
- Further evaluation in real-world clinical settings is recommended.
