Detection of mouse liver cancer via a parallel iterative shrinkage method in hybrid optical/microcomputed tomography...

Area of Science:

• Biomedical Imaging
• Medical Physics
• Oncology

Background:

• Liver cancer is a prevalent global malignancy.
• Noninvasive detection of small liver tumors is crucial for early diagnosis and treatment.
• Current imaging techniques may have limitations in detecting minute lesions.

Purpose of the Study:

• To develop and validate a novel algorithm for noninvasive, high-resolution detection of small liver tumors in mice.
• To improve the accuracy and efficiency of tumor localization using dual-modality imaging.
• To assess the potential of the algorithm for practical, whole-body cancer screening applications.

Main Methods:

• Development of a parallel iterative shrinkage (PIS) algorithm for dual-modality tomography.
• Integration of microcomputed tomography (micro-CT) for anatomical information and multiview bioluminescence imaging (BLI) for signal intensity.
• Incorporation of mathematical strategies including signal sparsity, smooth convex approximation, iterative shrinkage, and affine subspace for 3D reconstruction.
• Validation through in vivo experiments on bead-implanted mice and an HCCLM3 orthotopic xenograft mouse model.

Main Results:

• The PIS algorithm successfully localized tiny lesions (<3 mm) with a position bias ≤1 mm.
• Achieved computational efficiency 100-1000 times faster than existing algorithms.
• Demonstrated robustness to varying regularization parameters and lp norms.
• Successfully applied to an orthotopic xenograft mouse model for liver cancer detection.

Conclusions:

• The PIS algorithm offers accurate, efficient, and reliable 3D reconstruction for dual-modality tomography.
• This method significantly advances the noninvasive detection of small liver tumors in preclinical models.
• The PIS algorithm shows strong potential for practical applications in whole-body cancer detection and screening.