Needle detection using ultrasound B-mode and power Doppler analyses.

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Interventional Procedures

Background:

• Ultrasound is crucial for visualizing anatomy and guiding needles in medical interventions.
• Accurate needle detection in ultrasound images, especially at steep angles, remains a significant challenge.

Purpose of the Study:

• To present a new method for effective needle detection using combined B-mode and power Doppler ultrasound.
• To improve the accuracy and reliability of needle tracking during interventions.

Main Methods:

• Utilized a buzzer to excite the needle and acquired B-mode and power Doppler ultrasound images.
• Applied Radon transform, local-phase analysis, and alpha shape analysis for initial needle axis detection.
• Employed machine learning on extracted features to create a likelihood image, followed by Radon transform for refined axis detection.
• Developed a probabilistic approach to identify the needle tip in ex vivo animal tissues at various insertion angles.

Main Results:

• Achieved a 0% failure rate in needle detection across all tested angles (shallow, moderate, steep).
• Demonstrated high accuracy with mean errors of ≤0.7° for angle, ≤0.6 mm for axis, and ≤0.7 mm for tip.
• Outperformed two recently introduced needle detection methods in comparative evaluations.

Conclusions:

• The developed method shows significant potential for reliable and accurate needle detection in ultrasound imaging.
• This technique can enhance the safety and efficacy of ultrasound-guided needle interventions.