Hemoglobin Microbubbles and the Prediction of Different Oxygen Levels Using RF Data and Deep Learning
Teja Pathour1,2, Sugandha Chaudhary2, Shashank R Sirsi1,2
1Center for Imaging and Surgical Innovation, University of Texas at Dallas, Richardson, TX.
Summary
Researchers developed novel oxygen-sensitive hemoglobin microbubbles for ultrasound imaging. These microbubbles can acoustically detect blood oxygen levels, achieving 90.15% accuracy in distinguishing oxygenated from deoxygenated states.
Area of Science:
- Biomedical Engineering
- Acoustics
- Medical Imaging
Background:
- Ultrasound contrast agents (UCAs) are gas-encapsulated microspheres used for enhanced ultrasound imaging and drug delivery.
- Current methods for assessing blood oxygen levels can be invasive or lack spatial resolution.
Purpose of the Study:
- To develop a novel oxygen-sensitive hemoglobin-shell microbubble for non-invasive blood oxygen level detection.
- To investigate the potential of acoustic signals from these microbubbles to differentiate between oxygenated and deoxygenated hemoglobin.
Main Methods:
- Development of hemoglobin-shell microbubbles sensitive to oxygen levels.
- Utilizing radiofrequency (RF) data from microbubble oscillations under ultrasound.
- Application of a 1D convolutional neural network for classification of acoustic signals.
Main Results:
- Successfully differentiated between oxygenated and deoxygenated hemoglobin microbubbles using RF data.
- Achieved a testing accuracy of 90.15% in classifying the oxygenation state.
- Demonstrated that hemoglobin oxygen content influences the acoustic response of the microbubbles.
Conclusions:
- Hemoglobin-shell microbubbles show promise for acoustically measuring blood oxygen levels.
- This technology could enable new applications in blood-oxygen-level-dependent imaging, such as evaluating hypoxic regions in tumors and the brain.
