Real-time non-rigid target tracking for ultrasound-guided clinical interventions

Area of Science:

• Medical Imaging
• Image Registration
• Interventional Ultrasound

Background:

• Biological motion in deformable organs complicates minimally invasive interventions, leading to potential treatment failures.
• Real-time tracking of target anatomy is crucial for improving accuracy and efficacy in ultrasound-guided procedures.
• Speckle noise in ultrasound images challenges traditional intensity-based image registration methods.

Purpose of the Study:

• To develop and validate a robust ultrasound-based target tracking method for real-time guidance in non- or mini-invasive interventions.
• To address the limitations of existing registration techniques in handling speckle noise and motion artifacts in ultrasound imaging.
• To evaluate the performance of the EVolution registration algorithm for tracking soft tissues in mobile organs.

Main Methods:

• Employed the EVolution registration algorithm, which aligns image contrast patterns rather than direct intensities.
• Utilized sub-region matching and a global functional for estimating dense, elastic deformations, overcoming rigid displacement assumptions.
• Validated the approach using contact and standoff echography on healthy volunteers under free-breathing conditions.

Main Results:

• The EVolution algorithm demonstrated an average accuracy of approximately 1.5 mm with submillimeter precision.
• Achieved a computational performance of 20 images per second, enabling real-time tracking capabilities.
• Successfully validated for soft tissue tracking in the abdomen during free-breathing conditions.

Conclusions:

• The EVolution algorithm offers a robust and accurate solution for real-time target tracking in ultrasound-guided interventions.
• Its ability to handle speckle noise and deformable motion makes it suitable for complex anatomical targets.
• The method presents an attractive advancement for improving the safety and effectiveness of minimally invasive procedures.