Enabling quantitative robot-assisted compressional elastography via the extended Kalman filter.

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Robotics

Background:

• Compressional elastography can detect occult cancers but typically yields qualitative results.
• Quantitative elastography, like shear-wave elastography, offers absolute elasticity measures crucial for histopathological classification.
• Existing methods lack the precision needed for reliable quantitative analysis in diverse tissue types.

Purpose of the Study:

• To develop a stochastic method for quantitative elastography using robot-assistance.
• To overcome the qualitative limitations of traditional compressional elastography.
• To improve the accuracy and resilience of elasticity measurements in medical imaging.

Main Methods:

• A probabilistic framework utilizing an extended Kalman filter was employed.
• Robot-assistance integrated data from joint encoders and force/torque sensors.
• Multiple ultrasound acquisitions from robotic palpations were fused for inverse reconstruction.

Main Results:

• Quantitative elastograms were generated with high accuracy (within 5 kPa of mechanical testing).
• The method demonstrated resilience to uncertain initial conditions and measurement noise.
• Inclusions were clearly identified in elastograms, even with artifacts in displacement fields.

Conclusions:

• The stochastic, robot-assisted approach enables quantitative compressional elastography.
• This method enhances accuracy and stability in elasticity imaging.
• It provides a framework for integrating sensor data to advance elastography capabilities.