TL;DR
This paper introduces a novel, cost-effective 3D ultrasound tomography ring system for breast cancer screening. By employing a rotatable linear array of piezoelectric micromachined ultrasonic transducers (PMUTs) and a circular scanning method, the system achieves 3D imaging with improved patient comfort and reduced complexity compared to existing technologies.
What Problem Does This Paper Address?
Breast cancer remains a leading cause of mortality in women, making early detection crucial. While current screening methods like mammography, CT, and MRI are effective, they often involve harmful radiation and patient discomfort (compression). Ultrasound Computer Tomography (USCT) offers a promising radiation-free and painless alternative, but existing USCT systems are often prohibitively expensive and complex to build, hindering their widespread adoption. This paper aims to address this gap by developing a more accessible and user-friendly 3D USCT system.
Methods
The core innovation lies in the design of a novel 3D ultrasound imaging ring system. Instead of complex cylindrical transducer arrays, this system utilizes four linear arrays of 1x128 Piezoelectric Micromachined Ultrasonic Transducers (PMUTs) arranged in a 90-degree cross-vertical configuration. These arrays are mounted on a rotatable platform within a water tank, allowing for circular scanning of the breast.
Key methodological aspects include:
- PMUT Linear Array: A 1x128 PMUT linear array operating at a 3.5 MHz center frequency is employed. The array's cross-sectional structure includes a shell, damping pad, transducer array, matching layer, and acoustical lens to enhance emissive power.
- Circular Scanning: The system uses an electric rotary table to perform a 360-degree scan. This allows for the acquisition of tomographic data by rotating the transducer assembly around the object.
- Ultrasonic Tomography Algorithm: A specific algorithm is designed to reconstruct horizontal slice images from the acquired pulse-echo signals. This algorithm focuses on detecting reflections at interfaces where refractive indices differ.
- System Integration: The system comprises a personal computer workstation, the transducer ring, an electric rotary table with a controller, a constant temperature water tank (maintained at 32°C for patient comfort), and ultrasonic signal transmitting/receiving circuits.
- Data Acquisition: A programmable 16-channel ultrasonic transmitting/receiving circuit based on the AD9222 core chip with a 12 MHz sampling frequency is used.
Datasets
The paper does not explicitly mention or use pre-existing datasets for training or evaluation. Instead, it relies on experimental data acquired using:
- Iron Models: Cylindrical and square iron models were used for initial testing to validate the system's ability to image objects with distinct acoustic properties and delineate their shapes.
- Breast Model: A simulated breast model with an embedded mass was used to demonstrate the system's capability in imaging soft tissue and detecting abnormalities. This model allowed for evaluation of imaging quality at different scanning intervals.
Key Insights and Ideas
- Cost-Effectiveness and Simplicity: The use of rotatable linear PMUT arrays instead of complex cylindrical arrays significantly reduces fabrication difficulties and costs, making 3D USCT more accessible.
- Improved Patient Comfort: The prone positioning of the patient with the breast suspended in a warm water bath enhances comfort during the examination.
- Feasibility of Circular Scanning: The study validates that circular scanning with a rotatable linear array can effectively achieve 3D imaging, allowing for the recognition of mass size and location.
- Algorithm's Efficiency: The designed tomography algorithm is efficient, requiring detection of only half the depth of the target, which reduces power consumption and potentially speeds up imaging.
- Resolution-Imaging Quality Trade-off: The experiments demonstrate a clear relationship between the scanning interval angle and image quality, with smaller intervals leading to improved resolution.
Key Results
- Iron Model Imaging: The system successfully imaged cylindrical and square iron models, accurately delineating their sizes and outlines with relatively low error (5% for cylindrical, 7.5% for square). The cylindrical iron was imaged with better clarity due to more favorable reflection angles.
- Breast Model Imaging:
- The system was able to detect a simulated mass within the breast model, estimating its size to be approximately 5.0-5.1 cm.
- The boundary of the breast model was also distinguishable, though not perfectly complete due to the model's uneven shape and the limitations of single-element transmit/receive.
- Imaging quality improved significantly with decreasing scanning interval angles (10°, 6°, and 2°), demonstrating the impact of angular resolution on image detail.
- 3D Reconstruction Potential: The paper demonstrates the acquisition of multiple horizontal slices, laying the groundwork for 3D breast image reconstruction through multilayer slice data.
Limitations and Future Work
- Water Disturbance and Acquisition Time: The mechanical rotation required for scanning can lead to water disturbance and increased data acquisition time.
- Incomplete Boundary Detection: The current system struggles with detecting the complete boundary of unevenly shaped objects like the breast model, especially at larger incident angles.
- Single-Element Transmit/Receive: The use of a single element for both transmitting and receiving limits image quality.
- Future Work: The authors suggest implementing focus technology to improve image quality and plan to conduct clinical studies to further validate the system's feasibility and effectiveness. The potential for multi-modal imaging (combining reflection and transmission signals) is also highlighted.
Summary updated