Spatio-spectral regularization to improve magnetic resonance spectroscopic imaging quantification.

Area of Science:

• Biomedical Imaging
• Medical Physics
• Spectroscopy

Background:

• Magnetic Resonance Spectroscopic Imaging (MRSI) provides spatial biochemical distribution for disease diagnosis.
• Accurate metabolite quantification is essential for clinical applications of MRSI.
• Current MRSI quantification faces challenges like low signal-to-noise ratio and spectral overlap.

Purpose of the Study:

• To develop a novel quantification method for in vivo MRSI data.
• To address limitations of existing methods, including low signal-to-noise ratio and spectral overlap.
• To improve the accuracy and clinical utility of MRSI metabolite concentration measurements.

Main Methods:

• A novel spatio-spectral regularization scheme is proposed for MRSI quantification.
• Regularization terms are applied in transformed domains, not directly on parameters.
• A fast proximal optimization algorithm is used for simultaneous signal quantification and prior information incorporation.

Main Results:

• The proposed method reduced estimation error in metabolite concentrations by 41% on synthetic data.
• Experiments on in vivo brain MRSI data demonstrated improved fitting of overlapping peaks.
• The approach successfully identified low-concentration metabolites missed by single-voxel methods.

Conclusions:

• The novel spatio-spectral regularization method significantly enhances MRSI quantification accuracy.
• This technique offers improved diagnostic and monitoring capabilities for diseases like cancer and neurological disorders.
• The proposed algorithm provides a robust solution for challenging in vivo MRSI data.