使用行列式陣列之三維光聲成像

Abstract

使用光學解析度光聲顯微(optical-resolution photoacoustic microscopy)進行細胞等級之三維光聲成像時，會因取樣速度過慢而無法準確呈現物體動態變化。然而，當應用於三維腫瘤細胞培養系統時，可以放寬對於空間解析度之要求，以提升成像速度。因此可使用聲學解析度光聲顯微(acoustic-resolution photoacoustic microscopy)來即時擷取三維光聲影像。本研究以脈衝雷射透過光纖激發目標物體發生光聲效應，並使用中心頻率為11 MHz之行列式陣列(row-column addressed array)超音波探頭接收光聲訊號。透過超音波波束成型方法形成三維光聲影像。但在使用行列式陣列進行三維光聲成像時，會遇到影像不確定性(ambiguity)的問題，導致無法正確成像物體。行列式陣列三維光聲成像方法是分別對行陣列以及列陣列成像後，將兩者三維影像相乘而得。然而，由於行陣列和列陣列是由線型傳感器元件組成的一維陣列，所以分別在仰角方向和方位角方向上缺乏影像解析度。為了克服行列式陣列三維光聲成像中的影像不確定性，本研究提出了兩種方法。第一種方法為K空間濾波(k-space filtering)重建全採樣陣列通道資料。第二種方法為多角度接收複合方法，該方法透過將行列式陣列沿軸向旋轉，以不同角度來接收光聲訊號，分別進行成像後再依照旋轉的角度將影像回正，最後進行複合，以改善行列式陣列三維光聲成像時的影像不確定性。經過64次複合可以達到與全採樣陣列相似的結果，但成像所需的時間也會增加。因此需要在成像時間以及影像品質之間權衡。經模擬實驗結果評估，16次複合可以將影像不確定性造成的假影強度降低至-40 dB 以下，所需的成像時間也較少。在k-Wave模擬實驗中以點仿體以及網狀仿體來驗證效果。以波長532 nm雷射光架設實驗系統，照射碳粉薄膜、黑色裁縫線仿體。

Optical-resolution photoacoustic microscopy (OR-PAM) has been utilized for cellular-level 3-D photoacoustic imaging. However, its slow volume rate challenges capturing dynamic changes in real-time. In the context of 3-D tumor cell culture systems, where spatial resolution requirements can be relaxed, acoustic-resolution photoacoustic microscopy (AR-PAM) emerges as a viable solution for real-time imaging. This study utilizes an 11 MHz row-column addressed array (RCA) ultrasound transducer for the implementation of a 3D AR-PAM system. Specifically, the RCA encounters ambiguity issues leading to incorrect imaging. This study proposes a multi-angle reception compounding method to overcome the ambiguity problem. This method rotates the RCA along the central axis to receive photoacoustic signals at different angles. After individual imaging at each angle, the images are realigned according to the rotation angle and then combined to reduce ambiguity. Results show that after compounding 64 times, similar results to a fully sampled array can be achieved at the cost of the imaging time. On the other hand, compounding 16 times can reduce the artifact by ambiguity to below -40 dB. The effectiveness of the proposed method was verified through k-Wave simulation using point phantoms and mesh phantoms. In actual experiments, the proposed method was tested using a 532 nm wavelength laser and carbon films, black wire phantom. Finally, a k-space filtering approach previously proposed for ultrasound imaging (i.e., two-way focusing) was also tested in this study for photoacoustic imaging. The results were discussed, and it was concluded that the k-space approach is inadequate to solve the ambiguity problem.