Multi-pass Volume Reconstruction for Freehand Ultrasound: Demonstration in Simulation, Phantom, and In Vivo Kidney.

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Algorithm Development

Background:

• Freehand ultrasound systems offer portability but often suffer from noisy scans, impacting volume reconstruction accuracy.
• Accurate volume measurement is crucial for clinical diagnosis and monitoring, particularly in organs like the kidney.

Purpose of the Study:

• To develop and validate an algorithm for improving ultrasound volume reconstruction from noisy freehand scans.
• To enhance the clinical translatability and repeatability of freehand ultrasound systems.

Main Methods:

• Algorithm refinement through simulation using kidney models under various scan conditions.
• Pose tracking of an ultrasound probe using an Intel RealSense D435i camera.
• Validation using phantom scans with known volumes and in vivo scans of human kidneys.

Main Results:

• Phantom volume measurements achieved 2.5-17% accuracy (1-11.5% with automatic pass finding), outperforming traditional methods.
• In vivo kidney volume measurements showed 3.6-14.6% variation (7.7-23.3% with automatic pass finding) between scans.
• The algorithm demonstrated repeatable volume measurement results in both phantom and human subject scans.

Conclusions:

• The developed algorithm significantly improves the accuracy and repeatability of freehand ultrasound volume reconstruction.
• This advancement enhances the clinical utility and reliability of freehand ultrasound systems for volumetric measurements.