Visual cerebral microbleed detection on 7T MR imaging: reliability and effects of image processing

Area of Science:

• Radiology
• Neuroimaging
• Medical Physics

Background:

• Cerebral microbleeds (CMBs) are markers of small vessel disease.
• High-field Magnetic Resonance (MR) imaging, particularly at 7 Tesla (7T), offers enhanced spatial resolution for detecting CMBs.
• Optimizing image processing is crucial for maximizing the benefits of 7T MR imaging.

Purpose of the Study:

• To identify optimal maximum intensity projection (mIP) processing conditions for cerebral microbleed detection using 7T MR imaging.
• To evaluate the effectiveness of visual detection and inter-rater agreement for CMBs on 7T MR images.
• To explore the potential need for automated detection methods in 7T MR neuroimaging.

Main Methods:

• Utilized 7T MR imaging for brain scans.
• Investigated various mIP processing parameters to balance microbleed detection and section count.
• Assessed visual detection of CMBs and calculated inter-rater agreement.
• Analyzed challenges in visual detection, including microbleed size and susceptibility artifacts.

Main Results:

• 7T MR imaging demonstrated high sensitivity for detecting cerebral microbleeds.
• Specific mIP processing conditions were identified that optimize the number of detected CMBs against the number of image sections.
• Despite optimal processing, small microbleed size and susceptibility effects posed challenges for visual detection.
• Modest inter-rater agreement was observed, primarily due to missed microbleeds.

Conclusions:

• 7T MR imaging is highly sensitive for cerebral microbleed detection.
• Optimal mIP processing can enhance CMB visualization on 7T MR images.
• Visual detection of CMBs on 7T MR images is limited by microbleed size and susceptibility effects, impacting inter-rater reliability.
• Automated lesion detection techniques may be necessary to fully leverage the high resolution of 7T MR imaging for CMB analysis.