Pilot Lightweight Denoising Algorithm for Multiple Sclerosis on Spine MRI

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Neurology

Background:

• Multiple sclerosis (MS) diagnosis relies heavily on MRI, which is prone to noise and artifacts.
• Existing denoising algorithms for medical imaging are often complex and computationally intensive.
• Accurate and timely MS diagnosis is crucial for effective patient management.

Purpose of the Study:

• To develop and evaluate a lightweight denoising algorithm for MS MRI.
• To improve image quality and diagnostic accuracy using dictionary learning and orthogonal matching pursuit (OMP).
• To compare the proposed algorithm's performance against traditional denoising methods.

Main Methods:

• A retrospective analysis of 50 MS patients' spinal MRI scans (T2 weighted, 1.5T) was conducted.
• A novel OMP-based dictionary learning algorithm was applied to denoise images.
• Performance was evaluated using Peak Signal to Noise Ratio (PSNR) and Structural Similarity Index (SSIM) metrics.
• Compared results with traditional denoising techniques like wavelet denoising and non-local means (NLM) filtering.

Main Results:

• The OMP denoising algorithm achieved superior Structural Similarity Index (SSIM) (0.99 ± 0.01) with high consistency.
• PSNR values for OMP (37.6 ± 2.2) were comparable to non-local means (NLM) filtering (38.0 ± 1.8).
• Wavelet denoising showed higher PSNR but variable and lower SSIM compared to OMP and NLM.

Conclusions:

• The proposed OMP denoising algorithm demonstrates promising performance for clinical utility in MS diagnosis.
• Its lightweight and individualized approach facilitates easier integration into Picture Archiving and Communication Systems (PACS).
• This technology has the potential to enhance diagnostic accuracy, optimize radiologist workflow, and improve patient outcomes.