Photoacoustic Tomography with Temporal Encoding Reconstruction (PATTERN) for cross-modal individual analysis of the whole brain.

Cross-modal analysis of the same whole brain is an ideal strategy to uncover brain function and dysfunction. However, it remains challenging due to the slow speed and destructiveness of traditional whole-brain optical imaging techniques. Here we develop a new platform, termed Photoacoustic Tomography with Temporal Encoding Reconstruction (PATTERN), for non-destructive, high-speed, 3D imaging of ex vivo rodent, ferret, and non-human primate brains. Using an optimally designed image acquisition scheme and an accompanying machine-learning algorithm, PATTERN extracts signals of genetically-encoded probes from photobleaching-based temporal modulation and enables reliable visualization of neural projection in the whole central nervous system with 3D isotropic resolution. Without structural and biological perturbation to the sample, PATTERN can be combined with other whole-brain imaging modalities to acquire the whole-brain image with both high resolution and morphological fidelity. Furthermore, cross-modal transcriptome analysis of an individual brain is achieved by PATTERN imaging. Together, PATTERN provides a compatible and versatile strategy for brain-wide cross-modal analysis at the individual level.

Photoacoustic/ultrasound dual-modality imaging for marker clip localization in neoadjuvant chemotherapy of breast cancer.

Significance: To ensure precise tumor localization and subsequent pathological examination, a metal marker clip (MC) is placed within the tumor or lymph node prior to neoadjuvant chemotherapy for breast cancer. However, as tumors decrease in size following treatment, detecting the MC using ultrasound imaging becomes challenging in some patients. Consequently, a mammogram is often required to pinpoint the MC, resulting in additional radiation exposure, time expenditure, and increased costs. Dual-modality imaging, combining photoacoustic (PA) and ultrasound (US), offers a promising solution to this issue. Aim: Our objective is to localize the MC without radiation exposure using PA/US dual-modality imaging. Approach: A PA/US dual-modality imaging system was developed. Utilizing this system, both phantom and clinical experiments were conducted to demonstrate that PA/US dual-modality imaging can effectively localize the MC. Results: The PA/US dual-modality imaging can identify and localize the MC. In clinical trials encompassing four patients and five MCs, the recognition rate was ∼80%. Three experiments to verify the accuracy of marker position recognition were successful. Conclusions: We effectively localized the MC in real time using PA/US dual-modality imaging. Unlike other techniques, the new method enables surgeons to pinpoint nodules both preoperatively and intraoperatively. In addition, it boasts non-radioactivity and is comparatively cost-effective.

Mitigating Under-Sampling Artifacts in 3D Photoacoustic Imaging Using Res-UNet Based on Digital Breast Phantom.

In recent years, photoacoustic (PA) imaging has rapidly grown as a non-invasive screening technique for breast cancer detection using three-dimensional (3D) hemispherical arrays due to their large field of view. However, the development of breast imaging systems is hindered by a lack of patients and ground truth samples, as well as under-sampling problems caused by high costs. Most research related to solving these problems in the PA field were based on 2D transducer arrays or simple regular shape phantoms for 3D transducer arrays or images from other modalities. Therefore, we demonstrate an effective method for removing under-sampling artifacts based on deep neural network (DNN) to reconstruct high-quality PA images using numerical digital breast simulations. We constructed 3D digital breast phantoms based on human anatomical structures and physical properties, which were then subjected to 3D Monte-Carlo and K-wave acoustic simulations to mimic acoustic propagation for hemispherical transducer arrays. Finally, we applied a 3D delay-and-sum reconstruction algorithm and a Res-UNet network to achieve higher resolution on sparsely-sampled data. Our results indicate that when using a 757 nm laser with uniform intensity distribution illuminated on a numerical digital breast, the imaging depth can reach 3 cm with 0.25 mm spatial resolution. In addition, the proposed DNN can significantly enhance image quality by up to 78.4%, as measured by MS-SSIM, and reduce background artifacts by up to 19.0%, as measured by PSNR, even at an under-sampling ratio of 10%. The post-processing time for these improvements is only 0.6 s. This paper suggests a new 3D real time DNN method addressing the sparse sampling problem based on numerical digital breast simulations, this approach can also be applied to clinical data and accelerate the development of 3D photoacoustic hemispherical transducer arrays for early breast cancer diagnosis.

Sentinel lymph node mapping in patients with breast cancer using a photoacoustic/ultrasound dual-modality imaging system with carbon nanoparticles as the contrast agent: a pilot study.

Assessing the metastatic status of axillary lymph nodes is a common clinical practice in the staging of early breast cancers. Yet sentinel lymph nodes (SLNs) are the regional lymph nodes believed to be the first stop along the lymphatic drainage path of the metastasizing cancer cells. Compared to axillary lymph node dissection, sentinel lymph node biopsy (SLNB) helps reduce morbidity and side effects. Current SLNB methods, however, still have suboptimum properties, such as restrictions due to nuclide accessibility and a relatively low therapeutic efficacy when only a single contrast agent is used. To overcome these limitations, researchers have been motivated to develop a non-radioactive SLN mapping method to replace or supplement radionuclide mapping. We proposed and demonstrated a clinical procedure using a dual-modality photoacoustic (PA)/ultrasound (US) imaging system to locate the SLNs to offer surgical guidance. In our work, the high contrast of PA imaging and its specificity to SLNs were based on the accumulation of carbon nanoparticles (CNPs) in the SLNs. A machine-learning model was also trained and validated to distinguish stained SLNs based on single-wavelength PA images. In the pilot study, we imaged 11 patients in vivo, and the specimens from 13 patients were studied ex vivo. PA/US imaging identified stained SLNs in vivo without a single false positive (23 SLNs), yielding 100% specificity and 52.6% sensitivity based on the current PA imaging system. Our machine-learning model can automatically detect SLNs in real time. In the new procedure, single-wavelength PA/US imaging uses CNPs as the contrast agent. The new system can, with that contrast agent, noninvasively image SLNs with high specificity in real time based on the unique features of the SLNs in the PA images. Ultimately, we aim to use our systems and approach to substitute or supplement nuclide tracers for a non-radioactive, less invasive SLN mapping method in SLNB for the axillary staging of breast cancer.

Development of a digital breast phantom for photoacoustic computed tomography.

Photoacoustic (PA) imaging provides morphological and functional information about angiogenesis and thus is potentially suitable for breast cancer diagnosis. However, the development of PA breast imaging has been hindered by inadequate patients and a lack of ground truth images. Here, we report a digital breast phantom with realistic acoustic and optical properties, with which a digital PA-ultrasound imaging pipeline is developed to create a diverse pool of virtual patients with three types of masses: ductal carcinoma in situ, invasive breast cancer, and fibroadenoma. The experimental results demonstrate that our model is realistic, flexible, and can be potentially useful for accelerating the development of PA breast imaging technology.

Deep learning in photoacoustic imaging: a review.

SIGNIFICANCE: Photoacoustic (PA) imaging can provide structural, functional, and molecular information for preclinical and clinical studies. For PA imaging (PAI), non-ideal signal detection deteriorates image quality, and quantitative PAI (QPAI) remains challenging due to the unknown light fluence spectra in deep tissue. In recent years, deep learning (DL) has shown outstanding performance when implemented in PAI, with applications in image reconstruction, quantification, and understanding. AIM: We provide (i) a comprehensive overview of the DL techniques that have been applied in PAI, (ii) references for designing DL models for various PAI tasks, and (iii) a summary of the future challenges and opportunities. APPROACH: Papers published before November 2020 in the area of applying DL in PAI were reviewed. We categorized them into three types: image understanding, reconstruction of the initial pressure distribution, and QPAI. RESULTS: When applied in PAI, DL can effectively process images, improve reconstruction quality, fuse information, and assist quantitative analysis. CONCLUSION: DL has become a powerful tool in PAI. With the development of DL theory and technology, it will continue to boost the performance and facilitate the clinical translation of PAI.