Daiwei Li

Urinary stone disease is a common urological condition with increasing incidence, particularly in developed countries. Laser lithotripsy (LL) has become a preferred minimally invasive treatment due to its high precision and low tissue damage. Recent studies suggest that cavitation plays a critical role in stone damage during LL, and three-dimensional passive cavitation mapping (3D-PCM) has emerged as a promising tool for detecting these events. However, clinical translation of 3D-PCM remains challenging due to limitations in imaging depth, field of view (FOV), and procedural compatibility. Here, we present a large-FOV dual-modality imaging system (3D-PCM and B-mode ultrasound) based on a large-aperture planar ultrasound array. Through array optimization and model-based reconstruction, our system achieves an expanded FOV of ~40*40mm^2 at a clinically relevant imaging depth of ~110mm, while maintaining high spatial resolution of ~0.6 mm laterally and ~0.4 mm axially. In vivo experiments in a porcine model demonstrate that the reconstructed cavitation distribution correlates well with stone damage. Our technology has the potential to provide real-time treatment feedback during LL without disrupting the standard workflow.

Kidney stone disease is a major public health issue. By breaking stones with repeated laser irradiation, laser lithotripsy (LL) has become the main treatment for kidney stone disease. Laser-induced cavitation is closely associated with the stone damage in LL. Monitoring the cavitation activities during LL is thus crucial to optimizing the stone damage and maximizing LL efficiency. In this study, we have developed three-dimensional super-resolution passive cavitation mapping (3D-SRPCM), in which the cavitation bubble positions can be localized with an accuracy of 40 μm, which is 1/10th of the acoustic diffraction limit. Moreover, the 3D-SRPCM reconstruction speed has been improved by 300 times by adopting a GPU-based sparse-matrix beamforming approach. Using 3D-SRPCM, we studied LL-induced cavitation activities on BegoStones, both in free space of water and confined space of a kidney phantom. The dose-dependence analysis provided by 3D-SRPCM revealed that accumulated impact pressure on the stone surface has the highest correlation with the stone damage. By providing high-resolution cavitation mapping during LL treatment, we expect that 3D-SRPCM may become a powerful tool to improve the clinical LL efficiency and patient outcome.

Recent studies indicate that cavitation may play a vital role in laser lithotripsy. However, the underlying bubble dynamics and associated damage mechanisms are largely unknown. In this study, we use ultra-high-speed shadowgraph imaging, hydrophone measurements, three-dimensional passive cavitation mapping (3D-PCM), and phantom test to investigate the transient dynamics of vapor bubbles induced by a holmium:yttrium aluminum garnet laser and their correlation with solid damage. We vary the standoff distance (SD) between the fiber tip and solid boundary under parallel fiber alignment and observe several distinctive features in bubble dynamics. First, long pulsed laser irradiation and solid boundary interaction create an elongated “pear-shaped” bubble that collapses asymmetrically and forms multiple jets in sequence. Second, unlike nanosecond laser-induced cavitation bubbles, jet impact on solid boundary generates negligible pressure transients and causes no direct damage. A non-circular toroidal bubble forms, particularly following the primary and secondary bubble collapses at SD = 1.0 and 3.0 mm, respectively. We observe three intensified bubble collapses with strong shock wave emissions: the intensified bubble collapse by shock wave, the ensuing reflected shock wave from the solid boundary, and self-intensified collapse of an inverted “triangle-shaped” or “horseshoe-shaped” bubble. Third, high-speed shadowgraph imaging and 3D-PCM confirm that the shock origins from the distinctive bubble collapse form either two discrete spots or a “smiling-face” shape. The spatial collapse pattern is consistent with the similar BegoStone surface damage, suggesting that the shockwave emissions during the intensified asymmetric collapse of the pear-shaped bubble are decisive for the solid damage.

To investigate the effects of fiber lateral scanning speed across the stone surface (v_fiber) and fiber standoff distance (SD) on dusting efficiency during short pulse holmium (Ho): YAG laser lithotripsy (LL), pre-soaked BegoStone samples were treated in water using 0.2 J/20 Hz at SD of 0.10~0.50 mm with v_fiber in the range of 0~10 mm/s. Bubble dynamics, pressure transients, and stone damage were analyzed. To differentiate photothermal ablation vs. cavitation damage, experiments were repeated in air, or in water with the fiber tip at 0.25 mm proximity from the ureteroscope end to mitigate cavitation damage. At SD = 0.10 mm, the maximum dusting efficiency was produced at v_fiber = 3.5 mm/s, resulting in long (17.5 mm), shallow (0.15 mm), and narrow (0.4 mm) troughs. In contrast, at SD = 0.50 mm, the maximum efficiency was produced at v_fiber = 0.5 mm/s, with much shorter (2.5 mm), yet deeper (0.35 mm) and wider (1.4 mm), troughs. With the ureteroscope end near the fiber tip, stone damage was significantly reduced in water compared to those produced without the ureteroscope. Under clinically relevant v_fiber (1~3 mm/s), dusting at SD = 0.5 mm that promotes cavitation damage may leverage the higher frequency of the laser (e.g., 40 to 120 Hz) and, thus, significantly reduces the procedure time, compared to at SD = 0.1 mm that promotes photothermal ablation. Dusting efficiency during short pulse Ho: YAG LL may be substantially improved by utilizing an optimal combination of v_fiber, SD, and frequency.

One primary technical challenge in photoacoustic microscopy (PAM) is the necessary compromise between spatial resolution and imaging speed. In this study, we propose a novel application of deep learning principles to reconstruct undersampled PAM images and transcend the trade-off between spatial resolution and imaging speed. We compared various convolutional neural network (CNN) architectures, and selected a Fully Dense U-net (FD U-net) model that produced the best results. To mimic various undersampling conditions in practice, we artificially downsampled fully-sampled PAM images of mouse brain vasculature at different ratios. This allowed us to not only definitively establish the ground truth, but also train and test our deep learning model at various imaging conditions. Our results and numerical analysis have collectively demonstrated the robust performance of our model to reconstruct PAM images with as few as 2% of the original pixels, which can effectively shorten the imaging time without substantially sacrificing the image quality.