검색 상세

Advanced Beamforming Methods Based on Virtual and Passive Arrays for Ultrasound Theranostics

  • 박재범 (서강대학교 일반대학원)

초록(요약문)

의료 초음파는 질병 진단을 위한 영상화 기술이면서 동시에 조직에 음향 에너지를 전달하는 치료 기술로 활용된다. 진단 초음파에서는 제한된 변환자 구경, 채널 수, 메모리, 획득 시간 조건에서도 충분한 해상도와 대조도를 확보해야 하며, 치료 초음파에서는 고강도 집속 초음파가 표적 위치에 정확히 전달되어 주변 정상 조직 손상을 최소화해야 한다. 그러나 초음파 테라노스틱 시스템에서는 치료용 변환자와 진단용 프로브 사이의 정렬 오차, 치료 aperture 제한, 영상 품질 저하, 하드웨어 및 채널 수 제약과 같은 한계가 존재한다. 본 학위논문은 이러한 한계를 보완하기 위해 수동 빔포밍, 가상 송신원, 가상 수신 센서를 이용한 초음파 빔포밍 및 가상 aperture 기법을 제안한다. 첫 번째 연구에서는 HIFU 치료 전 안전한 초점 위치 확인을 위한 수동 빔포밍 기반 focal spot estimation 방법을 제안하였다. 실제 조직에서는 음속 및 감쇠의 불균질성, phase aberration, 산란, 반사, coupling 변화, 미세한 기계적 오차로 인해 실제 초점이 예상 위치와 달라질 수 있다. 본 연구에서는 조직 손상을 유발하지 않는 저강도 HIFU burst를 송신하여 초점 부근의 산란 신호를 생성하고, 진단용 ultrasound probe에서 동기화된 RF channel data를 획득하였다. 이후 HIFU 송신 경로를 명시적으로 모델링하지 않고 receive delay만으로 source-power image를 재구성하였으며, 강한 반사체에 의한 false peak를 줄이기 위해 B-mode reflectivity compensation을 적용하였다. Acoustic safety measurement, Field II simulation, BSA phantom, ex-vivo porcine tissue 실험을 통해 제안한 방법이 치료 전 안전한 초점 위치 추정에 활용될 수 있음을 확인하였다. 두 번째 연구에서는 가상 송신원을 이용한 diverging-wave imaging 기반 진단 영상 화질 개선 방법을 다루었다. Diverging-wave imaging은 넓은 시야각과 높은 프레임율을 제공하지만, virtual transmit source의 위치, 간격, 개수 및 receive aperture 선택 방식에 따라 해상도, grating lobe, SNR, contrast, frame rate 사이의 trade-off가 달라진다. 본 연구에서는 virtual-source depth, virtual-source pitch, virtual aperture size, receive strategy를 포함하는 unified beam-pattern model을 구성하고, DW-SAF와 DW-SF를 비교하였다. 이론 분석, numerical beam-field analysis, Field II simulation, phantom experiment, in-vivo liver B-mode 및 power Doppler imaging 결과를 통해 DW-SAF는 높은 공간 해상도와 경계 표현에, DW-SF는 grating-lobe 억제와 구현 안정성에 유리함을 보였다. 세 번째 연구에서는 추가 데이터 획득 없이 plane-wave imaging의 해상도와 artifact 특성을 개선하기 위한 virtual receive sensor 기법을 제안하였다. 일반적인 linear array imaging에서는 aperture size와 element spacing에 의해 lateral resolution 및 grating-lobe 특성이 제한된다. 본 연구에서는 physical transmit/receive array로 획득한 RF data를 일정 거리만큼 떨어진 virtual receive sensor에서 수신한 것처럼 재구성하여 effective aperture를 증가시키고 effective element spacing을 감소시켰다. 또한 extended receive path가 실제 physical element를 통과하지 않는 경우에는 인접한 physical channel data를 이용하여 virtual RF data를 근사적으로 재생성하였다. Copper wire phantom 및 ATS 549 commercial phantom 실험을 통해 제안한 방법이 point-target artifact를 줄이고 cyst contrast를 개선할 수 있음을 확인하였다. 종합하면, 본 학위논문은 치료 초음파에서는 안전한 에너지 전달을 위한 초점 위치 추정 방법을, 진단 초음파에서는 제한된 하드웨어 및 획득 조건에서 virtual aperture 개념을 이용한 영상 화질 개선 방법을 제시하였다. 세 연구는 모두 획득된 acoustic data를 새로운 공간 모델로 재해석하여 초음파 테라노스틱 시스템의 한계를 보완한다는 공통점을 가진다. 향후 연구에서는 3차원 초점 위치 추정, heterogeneous tissue correction, 실시간 구현, reduced-channel system 최적화, 다양한 in-vivo환경에서의 검증이 필요하다.

more

초록(요약문)

Medical ultrasound is used not only as an imaging technique for disease diagnosis but also as a therapeutic technique for delivering acoustic energy to tissue. In diagnostic ultrasound, sufficient resolution and contrast must be achieved under limited transducer aperture, channel count, memory, and acquisition-time conditions. In therapeutic ultrasound, high-intensity focused ultrasound should be accurately delivered to the target region while minimizing damage to surrounding normal tissue. However, ultrasound theranostic systems still have limitations, including alignment errors between the therapeutic transducer and diagnostic probe, restricted therapeutic aperture, degraded image quality, and hardware or channel-count constraints. These limitations make it difficult to simultaneously achieve accurate treatment targeting and high-quality image guidance in practical systems. This thesis proposes ultrasound beamforming and virtual-aperture methods using passive beamforming, virtual transmit sources, and virtual receive sensors to compensate for these limitations. The first study proposed a passive beamforming-based focal spot estimation method for safe pre-treatment monitoring in HIFU therapy. In actual tissue, the focal spot may deviate from the expected position because of sound-speed and attenuation heterogeneity, phase aberration, scattering, reflection, coupling variation, and small mechanical errors. In this study, a low-intensity HIFU burst that does not induce tissue damage was transmitted to generate scattered signals near the focal region, and synchronized RF channel data were acquired using a diagnostic ultrasound probe. The source-power image was then reconstructed using receive delays only, without explicitly modeling the HIFU transmit path. B-mode reflectivity compensation was also applied to reduce false peaks caused by strong reflectors. Acoustic safety measurements, Field II simulations, BSA phantom experiments, and ex-vivo porcine tissue experiments confirmed that the proposed method can be used for safe focal spot estimation before treatment. The second study investigated diverging-wave imaging using virtual transmit sources for diagnostic image enhancement. Diverging-wave imaging provides a wide field of view and a high frame rate, but the trade-off among resolution, grating lobe, SNR, contrast, and frame rate depends on the position, pitch, number, and receive-aperture strategy of the virtual transmit sources. In this study, a unified beam-pattern model including virtual-source depth, virtual-source pitch, virtual aperture size, and receive strategy was developed, and two methods were compared: DW-SAF, in which the receive aperture follows the virtual source, and DW-SF, in which the receive aperture remains fixed. Theoretical analysis, numerical beam-field analysis, Field II simulation, phantom experiments, and in-vivo liver B-mode and power Doppler imaging showed that DW-SAF is advantageous for high spatial resolution and boundary visualization, whereas DW-SF is advantageous for grating-lobe suppression and implementation stability. The third study proposed a virtual receive sensor method to improve the resolution and artifact characteristics of plane-wave imaging without additional data acquisition. In conventional linear-array imaging, lateral resolution and grating-lobe behavior are limited by the aperture size and element spacing. In this study, RF data acquired from a physical transmit/receive array were reconstructed as if they had been received by virtual receive sensors located at an offset distance from the physical array. This increased the effective aperture and reduced the effective element spacing in the receive model, thereby improving resolution and suppressing grating lobes. In addition, when an extended receive path did not pass through an actual physical element, virtual RF data were approximately regenerated using neighboring physical-channel data to reduce data loss. Copper wire phantom and ATS 549 commercial phantom experiments demonstrated that the proposed method reduced point-target artifacts and improved cyst contrast. In summary, this thesis presents a focal spot estimation method for safe energy delivery in therapeutic ultrasound and virtual-aperture-based image enhancement methods for diagnostic ultrasound under limited hardware and acquisition conditions. The three studies share the common concept of reinterpreting acquired acoustic data using new spatial models without physically expanding the ultrasound system. By using passive and virtual array concepts, the proposed methods provide computational strategies for improving both treatment safety and diagnostic image quality in ultrasound theranostic systems. Future work should include three-dimensional focal spot estimation, heterogeneous tissue correction, real-time implementation, reduced-channel system optimization, and validation in various in-vivo environments.

more

목차

Chapter I. Introduction 13
Chapter II. Safe Treatment: Focal Spot Estimation with Passive Beamforming 19
2.1 Introduction 19
2.2 Theory 22
2.2.1 Transmit Waveform Design for Pre-Treatment Focal Monitoring 22
2.2.2 RF Data Acquisition for Passive Reception 24
2.2.3 Passive Beamforming Based Focal Spot Estimation 25
2.2.4 PBF Image Interpretation and Corrected Focal Spot Estimation 26
2.3 Materials and Methods 29
2.3.1 HIFU System and Acoustic Safety Measurements 29
2.3.2 Field II Simulation 32
2.3.3 Phantom and ex-vivo Experiment 34
2.3.4 Evaluation Metrics for Focal Spot Estimation 37
2.4 Results 39
2.4.1 Safety Measurement 39
2.4.2 Field II Simulation 43
2.4.3 Phantom and Ex-vivo Experiment 49
2.5 Discussion 54
2.6 Conclusion 59
Chapter III. Diagnostic Image Enhancement 1: Diverging-wave Imaging Using Virtual Transmit Sources 61
3.1 Introduction 61
3.2 Materials and Methods 63
3.2.1 Geometric Model for Diverging-Wave Generation 64
3.2.2 Receive Strategies for Diverging-Wave Imaging 65
3.2.3 Continuous Aperture Transmit Model for Diverging Waves 67
3.2.4 Continuous Aperture Receive Model for Diverging-wave Imaging 68
3.2.5 Two-Way Beam-Pattern Formulation 69
3.2.6 Discrete Array Model for Diverging-Wave Imaging 71
3.2.7 Simulation, Phantom, and in-vivo Experimental Setup 74
3.3 Results 78
3.3.1 Numerical Simulation of Synthetic-Aperture Beam Fields 78
3.3.2 Field II Simulation Results 81
3.3.3 Phantom Experiment Results 84
3.3.4 In-vivo Experiment Results 87
3.4 Discussion 90
3.5 Conclusion 93
Chapter IV. Diagnostic Image Enhancement 2: Virtual Receive Sensors for Plane-wave Imaging 94
4.1 Introduction 94
4.2 Theory 96
4.2.1 Principle of an Ultrasound Imaging System 96
4.2.2 Plane-Wave Transmission and Reception Using Transducer 97
4.2.3 Plane-Wave Acoustic Field Analysis of Array Method 98
4.3 Materials and Methods 103
4.3.1 Imaging Technique Using a Virtual Receive Sensor 103
4.3.2 Parameter Setting for Physical and Virtual Arrays 104
4.3.3 Method for Generating RF Data of Physical Tx/Rx Sensors 105
4.3.4 RF Data of a Physical Transmit and Virtual Receive Sensors 107
4.4 Experimental Results 111
4.4.1 Verification of the Virtual Receive Sensor Using Wire Phantom 111
4.4.2 Verification of the Virtual Sensor Using a Commercial Phantom 116
4.5 Discussion 119
4.6 Conclusion 123
Chapter V. Conclusions and Further Work 126

more
반출 Meta View 목록