Evaluation of the OSC-TV iterative reconstruction algorithm for cone-beam optical CT

Area of Science:

• Medical Physics
• Image Reconstruction
• Radiotherapy Dosimetry

Background:

• Cone-beam optical computed tomography (CT) is utilized for 3D dosimetry in radiation therapy, mapping dose distributions in radiosensitive gels.
• Model-based iterative reconstruction algorithms have the potential to enhance optical CT image quality and expand its clinical applications.

Purpose of the Study:

• To evaluate an iterative reconstruction approach, specifically the ordered subsets convex (OSC) algorithm with total variation (TV) minimization, for cone-beam optical CT.
• To compare the performance of OSC-TV against conventional filtered backprojection (FBP) in optical CT image reconstruction.

Main Methods:

• The OSC-TV algorithm was implemented on a GPU for fast reconstruction, achieving times comparable to FBP.
• Experimental data from a cone-beam optical CT system and numerical simulations were used for evaluation.
• Image quality was assessed using spatial resolution and uniformity phantoms, analyzing modulation transfer function, uniformity, and accuracy. Artifacts from refraction and signal loss were also studied.

Main Results:

• OSC-TV demonstrated superior image quality compared to FBP, attributed to model-based simulation of photon attenuation.
• Significant improvements in spatial resolution and noise reduction were observed with OSC-TV.
• Accuracy of linear attenuation coefficient estimation was comparable to FBP, and certain artifacts were reduced, though container wall artifacts persisted.

Conclusions:

• Iterative reconstruction, particularly OSC-TV, enhances cone-beam optical CT image quality.
• OSC-TV shows potential for improved rendering of spatial features and reduction of artifacts in optical CT, benefiting clinical gel dosimetry.