Semiautomated detection of cerebral microbleeds in magnetic resonance images
Samuel R S Barnes1, E Mark Haacke, Muhammad Ayaz
1Department of Biomedical Engineering, Wayne State University, Detroit, MI 48201, USA. sbarnes@llu.edu
Abstract:
Cerebral microbleeds (CMBs) are increasingly being recognized as an important biomarker for neurovascular diseases. So far, all attempts to count and quantify them have relied on manual methods that are time-consuming and can be inconsistent. A technique is presented that semiautomatically identifies CMBs in susceptibility weighted images (SWI). This will both reduce the processing time and increase the consistency over manual methods. This technique relies on a statistical thresholding algorithm to identify hypointensities within the image. A support vector machine (SVM) supervised learning classifier is then used to separate true CMB from other marked hypointensities. The classifier relies on identifying features such as shape and signal intensity to identify true CMBs. The results from the automated section are then subject to manual review to remove false-positives. This technique is able to achieve a sensitivity of 81.7% compared with the gold standard of manual review and consensus by multiple reviewers. In subjects with many CMBs, this presents a faster alternative to current manual techniques at the cost of some lost sensitivity.
Insights
A new semiautomated method accurately identifies cerebral microbleeds (CMBs) on brain imaging. This technique speeds up analysis and improves consistency compared to manual methods for neurovascular disease research.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Neurology
Background:
- Cerebral microbleeds (CMBs) are crucial biomarkers in neurovascular diseases.
- Current manual quantification methods for CMBs are time-consuming and inconsistent.
- There is a need for efficient and reliable automated or semiautomated techniques.
Purpose of the Study:
- To develop and validate a semiautomated technique for identifying cerebral microbleeds (CMBs) in susceptibility-weighted images (SWI).
- To reduce processing time and enhance consistency in CMB quantification compared to manual methods.
Main Methods:
- A semiautomated approach using statistical thresholding to detect hypointensities in SWI.
- A Support Vector Machine (SVM) classifier trained on features like shape and signal intensity to differentiate true CMBs.
- Manual review of automated results to eliminate false positives.
Main Results:
- The semiautomated technique achieved a sensitivity of 81.7% against manual review consensus.
- The method offers a faster alternative for analyzing subjects with a high number of CMBs.
- A trade-off exists between speed and sensitivity, with some sensitivity lost compared to manual methods.
Conclusions:
- The presented semiautomated technique offers a promising, faster, and more consistent method for CMB quantification in neuroimaging research.
- This approach can aid in the study of neurovascular diseases by improving the efficiency of biomarker analysis.
- Further refinement may enhance sensitivity while maintaining processing speed advantages.
