A novel high order directional total variation algorithm of EPR imaging for fast scan

Area of Science:

• Medical Imaging
• Biophysics
• Radiotherapy Technology

Background:

• Pulsed Electron Paramagnetic Resonance (EPR) Imaging (EPRI) is crucial for oxygen imaging in precision radiotherapy.
• Current EPRI methods average projections to improve signal-to-noise ratio (SNR), leading to slow scan speeds.
• Reducing scan repetitions (shots) causes noisy projections, hindering image reconstruction.

Purpose of the Study:

• To develop a novel High Order Directional Total Variation (HODTV) algorithm for accelerated 3D pulsed EPRI.
• To address staircase artifacts common in existing Directional Total Variation (DTV) algorithms.
• To improve both image quality and scan speed in EPRI.

Main Methods:

• The HODTV algorithm incorporates regularization of high-order derivatives for data fidelity and detail recovery.
• A Chambolle-Pock (CP) algorithm was derived and verified for solving the HODTV model.
• The HODTV algorithm's performance was evaluated using real-world data, comparing it to FBP, TV, and DTV.

Main Results:

• HODTV demonstrated superior reconstruction accuracy compared to FBP, TV, and DTV algorithms.
• The HODTV algorithm achieved reconstructions with only 100 shots in approximately 6 seconds.
• This contrasts with the FBP algorithm, which required 2000 shots and approximately 120 seconds for similar quality.

Conclusions:

• The proposed HODTV algorithm enables significantly faster 3D pulsed EPRI with high accuracy.
• This advancement has the potential to enhance clinical imaging workflows, particularly for fast 3D EPRI.
• The HODTV method offers a promising solution for improving the efficiency and quality of EPRI in radiotherapy.