Tumor Identification

Objective: In US-guided liver tumor ablations, tumor identification remains challenging in some scenarios, such as similar acoustic characteristics between the tumor and surrounding tissue, irregular tumor locations (e.g., high liver dome), and tumor mimics (e.g., cirrhotic regenerative nodules, prior ablation sites). While multimodal registration techniques have been extensively studied to address these limitations, they have not yet become a standard clinical component due to the high complexity and computational demands. Therefore, this work aims to develop a clinically feasible 2D US-CT/MRI registration approach to reduce inter-physician variability in interpreting structural details, thereby facilitating applicator guidance during procedures.

Reference:

[1] Deep Regression 2D-3D Ultrasound Registration for Liver Motion Correction. (see detail)

[2] 3D US-CT/MRI registration for percutaneous focal liver tumor ablations. (see detail)

[3] Development of 3D US liver system. (see detail)