Photoacoustic and Ultrasound Imaging with Diffractive Resolution Enhancing Aperture Mode (DREAM)

Introduction:: Photoacoustic (PA) and ultrasound (US) imaging systems have many options for the method of ultrasonic detection. Linear-array transducers provide the benefits of low cost, large field-of-view, and fast scanning. However, a traditional scanning linear-array integrated PA/US system achieves low elevational resolution and sensitivity due to the fixed acoustic lens. The use of single-slit diffraction in linear-array based PA imaging systems has been shown to enhance elevational resolution and sensitivity. However, these systems have the downsides of slow and computationally expensive 3D image reconstructions and an inability to generate co-registered US images. In this study, we present a linear-array integrated PA/US imaging system with diffractive resolution enhancing aperture mode (DREAM). DREAM uses single-slit diffraction to enhance US as well as PA imaging. DREAM relies on a dynamic slit allowing for accurate calibration and adjustment of the slit width and position, and further allowing for a thorough characterization of the relationship between slit width and image quality. Furthermore, we present a novel fast focal line (FFL) reconstruction technique, reducing the complete 3D reconstruction time by over two orders of magnitude. We then present life science applications of PA/US imaging with DREAM in the preclinical noninvasive study of frogs and mice.


Materials and Methods:: Dynamic slit design: The system uses a scanning L7-4 linear array transducer coupled with four multimode fiber bundles on each side for PA imaging. A clear thin polyester bag is filled with water and placed on the transducer surface to act as an acoustic coupling medium while keeping all the system components out of water. Two borosilicate cover glasses are used to create a slit at the elevational focus of the transducer. The cover glasses are each individually attached to a separate motorized stage, allowing for precise control of the slit width. In PA mode, diffuse illumination is used for the generation of ultrasound waves in the target. In US mode, plane waves are used to generate approximately cylindrical waves through the slit by diffraction, which allows for a synthetic aperture to be created in the image reconstruction.

Fast focal line (FFL) reconstruction: The FFL reconstruction is geometrically based on the previously described 3D focal line (3DFL) reconstruction. However, to increase computational efficiency, the algorithm applies a single matrix multiplication operation for each reconstructed frame, as shown below:

Sr = f

Here, S is a sparse matrix of size n x m where n is the number of voxels in the reconstruction plane and m is the number of transducer elements times the number of data samples times the number of reconstruction frames. The vector r is the corresponding radio frequency data of size m x 1. The vector f is the resultant vectorized image at a given reconstruction plane.


Results, Conclusions, and Discussions:: Phantom characterization: Two phantom studies are presented in Fig. 1. The first shows the effect of slit width on image resolution and signal-to-noise ratio (SNR). A crosshair phantom was imaged at slit widths ranging from 0.08 to 2.23 mm, and the lateral and elevational resolutions and SNRs are plotted as a function of slit width. The second phantom is composed of four pieces of graphite (diameter: 300 µm) embedded at various depths in agarose gel mixed with intralipid. We can see that DREAM is capable of deep imaging ( > 2 cm) in both photoacoustic and ultrasound modes.

Frog imaging: In Fig. 2, we show the clear improvement of DREAM imaging compared to traditional linear-array imaging. Vessels and structures that were previously undetectable or unresolvable in traditional PA and US become visible in DREAM.

Gold nanoparticle tracking in tumor mouse: We applied DREAM to track the distribution of gold nanoparticles in a tumor mouse over 72 hours. Tumors are known to generate leaky vessels, and particles that are within a particular size window will accumulate in the tumor. Using DREAM, we were able to track the movement of gold nanoparticles as they were filtered from the circulatory system into the liver, and as they accumulated in the tumor due to the poorly constructed vessels.

Discussion and Conclusion: Linear-array based PA and US imaging are powerful tools for preclinical studies in the life sciences, through generating deep-tissue structural and functional data. Here, we have shown that DREAM can significantly improve image resolution and sensitivity in both PA and US modes. The FFL reconstruction improves the feasibility of DREAM by reducing computational expense and processing time. DREAM has many applications in biological preclinical studies and has high potential for clinical translation.


Acknowledgements (Optional): :

References (Optional): :