Related Experiment Video
Updated: Mar 19, 2026
09:33
Diffusion Tensor Magnetic Resonance Imaging in the Analysis of Neurodegenerative Diseases
Published on: July 28, 2013
29.4K
Improved Automatic Detection of New T2 Lesions in Multiple Sclerosis Using Deformation Fields.
M Cabezas1,2, J F Corral3, A Oliver2
1From the Section of Neuroradiology, Department of Radiology (M.C., J.F.C., C.A., D.P., À.R.) mariano.cabezas@vhir.org mcabezas@eia.udg.edu.
AJNR. American Journal of Neuroradiology
|June 11, 2016
Summary
This study introduces an improved automated method for detecting new T2 lesions in multiple sclerosis (MS) brain MR imaging. The new approach enhances accuracy and reduces variability, aiding in disease activity monitoring.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Multiple Sclerosis Research
Background:
- Detection of new or enlarging T2 lesions on brain MR imaging is crucial for monitoring multiple sclerosis (MS) disease activity.
- Current detection methods can be limited by factors that image subtraction techniques aim to overcome.
- Automated detection aims to reduce inter- and intraobserver variability.
Purpose of the Study:
- To enhance automated detection of new T2 lesions in MS.
- To minimize user interaction in the lesion detection process.
- To reduce inter- and intraobserver variability in MS lesion detection.
Main Methods:
- Multiparametric brain MR imaging from 36 MS patients with new T2 lesions was analyzed.
- Images were registered using affine transformation and the Demons algorithm.
- A novel pipeline involving image subtraction, thresholding, and refinement using deformation fields was developed and compared to existing methods.
Main Results:
- The proposed deformation field-based pipeline achieved a Dice similarity coefficient of approximately 0.70.
- The method demonstrated a true-positive detection rate of 70.9% and a false-positive rate of 17.8%.
- A strong statistically significant correlation (r = 0.81) was observed between automated and expert visual detection.
Conclusions:
- The deformation field-based approach effectively detects new/enlarging T2 lesions in MS.
- This method significantly reduces false positives while preserving true positives.
- The approach shows good correlation with visual detection and has the potential to decrease user interaction and observer variability.
