A Computational Framework for Data Fusion in MEMS-Based Cardiac and Respiratory Gating
Mojtaba Jafari Tadi1, Eero Lehtonen2, Jarmo Teuho3
1Department of Future Technologies, University of Turku, 20500 Turku, Finland. mojtaba.jafaritadi@utu.fi.
Sensors (Basel, Switzerland)
|September 27, 2019
Summary
This study introduces a novel framework using microelectromechanical (MEMS) sensors for dual cardiac and respiratory gating in medical imaging. The new method effectively fuses sensor data to improve motion compensation, benefiting patients undergoing PET/CT scans.
Area of Science:
- Medical Imaging
- Biomedical Engineering
- Sensor Technology
Background:
- Dual cardiac and respiratory gating is crucial for motion compensation in nuclear medicine.
- Existing gating techniques can be complex and require specialized equipment.
- Accurate gating is essential for improving image quality in Positron Emission Tomography (PET) and Computed Tomography (CT).
Purpose of the Study:
- To develop and validate a novel data fusion framework for dual cardiac and respiratory gating using multidimensional microelectromechanical (MEMS) motion sensors.
- To enable robust estimation of chest vibrations, including precordial vibrations and respiratory movements.
- To facilitate prospective gating for enhanced accuracy in PET, CT, and radiotherapy applications.
Main Methods:
- Utilized a single dual sensor unit with accelerometer and gyroscope to capture chest movements in three orientations.
- Applied Principal Component Analysis (PCA) to fuse accelerometer and gyroscope signals for respiration estimation.
- Employed Independent Component Analysis (ICA) to extract cardiac motion from combined sensor data.
- Identified systolic and diastolic phases using an adaptive multi-scale peak detector and short-time autocorrelation function.
Main Results:
- Demonstrated a strong positive correlation (r = 0.73–0.87) between MEMS-derived respiration curves and reference sensors.
- Achieved a mean time offset of 0.23–0.3 ± 0.15–0.17 s for MEMS-driven triggers compared to optical camera triggers.
- Successfully identified systolic time intervals from MEMS signals, correlating them with total cardiac cycle length.
Conclusions:
- The combination of chest motion data using ICA and PCA provides a robust dual gating solution solely with MEMS sensors.
- This approach can significantly improve the prediction of cardiac and respiratory quiescent phases, particularly for clinical patients.
- The presented methods lay the foundation for future clinical PET/CT imaging advancements utilizing dual inertial sensors.
