Vastly accelerated linear least-squares fitting with numerical optimization for dual-input delay-compensated...

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Radiology

Background:

• Liver perfusion imaging is crucial for diagnosing and monitoring liver diseases.
• Conventional methods for generating liver perfusion maps are computationally intensive and time-consuming.

Purpose of the Study:

• To propose an efficient algorithm for dual-input compartment modeling to generate liver perfusion maps.
• To evaluate the accuracy, performance, and computation time of the proposed method compared to conventional techniques.

Main Methods:

• Implemented whole field-of-view linear least squares (LLS) to fit a delay-compensated dual-input single-compartment model.
• Used high temporal resolution (4 frames/sec) contrast-enhanced 3D liver data.
• Compared whole-field LLS with voxel-wise nonlinear least-squares (NLLS) using simulated and experimental data.

Main Results:

• Simulations demonstrated good agreement between LLS and NLLS for various kinetic parameters.
• The whole-field LLS method generated liver perfusion maps approximately 160-fold faster than voxel-wise NLLS.
• Similar perfusion parameters were obtained with both methods.

Conclusions:

• Whole-field LLS offers a considerable speedup for liver perfusion analysis.
• This method enables efficient generation of perfusion maps using dual-input modeling.
• The findings suggest LLS is a viable and faster alternative to NLLS for liver perfusion imaging.