This conference presentation was prepared for SPIE BiOS, 2024.

To enable high-resolution imaging of melanoma vasculature in humans, we developed fast raster-scanning optoacoustic mesoscopy (FRSOM) using an illumination scheme that is co-axial with a high-sensitivity ultrasound detector path, yielding 15 second single-breath-hold scans that minimize motion artifacts. We apply this FRSOM to image 10 melanomas and 10 benign nevi in vivo, showing marked differences between malignant and benign lesions, supporting the possibility to use biomarkers extracted from RSOM imaging of vasculature for lesion characterization to improve diagnostics.

A longitudinal study on both 3D and 2D photoacoustic and Doppler ultrasound imaging of human hand rheumatoid arthritis progression has been performed using an automatic imaging system based on GEHC VividTM E95 with L8-18i-D probe and OPOTEK tunable laser system. Bi-weekly imaging has been performed starting from baseline (before patients start medication). Both photoacoustic and Doppler ultrasound can confirm the disease development, however, photoacoustic has higher correlation coefficients (with a median of 78.9%, p = 0.039) with patients’ PGA score.

Post-treatment colorectal cancer tissue often contains fibrosis and edema in tumor bed, photoacoustic elastography is a good tool for evaluating the elasticity change in post treatment colorectal cancer, especially in ultrasonically homogeneous and hypoechogenic regions. We implemented photoacoustic elastography for post-treatment colorectal cancer, developed elasticity phantoms to simulate colon scanning to evaluate elasticity measurement accuracy, applied it in ex vivo tissue scan to evaluate the elasticity change with ultrasound and photoacoustic signal. Results demonstrated good correspondence between ultrasound and photoacoustic elasticity phantom measurements, and showed potential of elasticity measurement with a normal ex vivo colon scan.

Biomedical optical imaging and sensing techniques are known to be confounded by skin tone, typically leading to worse outcomes in people with darker skin tones. Here, we present a healthy volunteer study to evaluate the effects of skin tone in photoacoustic imaging. We recruited 42 people, 6 from each Fitzpatrick skin type and 6 people with vitiligo. Our preliminary analysis shows increased reconstruction artefacts and changes in blood oxygen estimates in higher Fitzpatrick types. The results suggest that equitable application of quantitative photoacoustic imaging in the clinic will require improved methods to account for changing light fluence and acoustic artefacts.

Our previous investigations effectively employed an interstitial all-optical needle photoacoustic sensing probe on ex vivo tissue studies. In this study, our goal is to locate aggressive PCa within an intact prostate ex vivo using our latest version translational needle PA probe. Targeting specific tissue components, we utilized wavelengths of 1220nm, 1370nm, 800nm, and 266nm. Evans blue dye was injected at the measured positions for histopathology analysis. The acquired photoacoustic signals were analyzed using PASA, including spectrum slopes and midbandfits derived at all wavelengths. With the limited number of insertions, we were able to identify cancers in 3 out of 4 prostates. This non-invasive methodology holds considerable promise for future clinical applications.

Crohn's Disease often leads to obstructing intestinal strictures caused by inflammation, fibrosis, and muscular hypertrophy. In this work, we developed a translational ultrasound/photoacoustic catheter probe that is compatible with standard colonoscopy procedures to differentiate normal, acute inflammation and fibrotic conditions in intestinal obstruction in rabbits. The prototype catheter probe consists of a 600-um core optical fiber for photoacoustic illumination and a miniaturized ultrasound array with 48 elements operating at 9.1 MHz for photoacoustic signal reception. A hydrostatic balloon covers the optical fiber and ultrasound array for acoustic coupling. Initial validation in rabbits was successful, and human subject evaluation is ongoing.

Polymer micro-ring resonators (MRRs) have been proven to be one of the most versatile choices for ultrasonic detections, largely due to their high detection sensitivity over a wide frequency range, optical transparency, and at a significantly reduced form-factor. However, precisely controlling the resonance mode of individual MRRs became increasingly difficult due to the stringent requirements in their dimensional tolerance. Here we report a nanofabrication strategy to substantially improve the dimensional accuracy of the MRR to precisely control its resonance modes. We experimentally demonstrate polymer MRR in an array format, in enabling rapid and parallel ultrasound detection.

Photon Absorption Remote Sensing (PARS) captures optical absorption (radiative and non-radiative) and scattering contrast facilitating label-free imaging of biomolecules. Presented here, a new automated multiwavelength PARS system is developed. Concurrently, intelligent signal processing explores PARS time-domain features. These contrasts form unique PARS feature vectors which highlight numerous biomolecules label-free. Combined with innovations in AI, PARS can produce virtual histology images which are effectively indistinguishable from current histochemical staining as demonstrated by clinical diagnostic equivalence studies. This represents a significant step towards developing a clinical system for label-free pathology of tissues, with the potential to provide molecular diagnostics in the future.

The Fabry-Perot (FP) sensor is typically read out by scanning a focused laser beam over the sensor surface and measuring the reflected light. However, the acquisition time can be relatively long when a large number of spatial sampling points are required. An alternative approach is to use widefield illumination and measure the reflected light using a camera. This approach has enabled the acquisition of 3D photoacoustic images composed of 327,000 spatial sampling points in a couple of seconds or less. In this study, we investigate the possibility of further improving imaging speed by combining the widefield illumination approach with the principles of compressed sensing. Preliminary results suggest that current imaging times could be reduced by at least a factor of 4, enabling the acquisition of 3D photoacoustic images in sub-second timeframes.

We report a head-mounted photoacoustic fiberscope for cerebral imaging in a freely behaving mouse. The 4.5-gram imaging probe has a 9-µm lateral resolution and 0.2-Hz frame rate over a 1.2-mm wide area. The probe can continuously monitor cerebral oxygenation and hemodynamic responses at single-vessel resolution, showing significantly different cerebrovascular responses to external stimuli under anesthesia and in the freely moving state. When mice subjected to high-concentration CO₂ respiration, enhanced oxygenation to compensate for hypercapnia can be visualized in freely moving state. Comparative studies exhibit significantly weakened compensation capabilities in obese mice. This new imaging modality shows promise in neuroscience studies.

We developed a system for whole-body ultrasound tomography of the entire human cross-section, localizing biopsy needles, and monitoring microwave ablation progress. Our ultrasound approach enables 2D isotropic images of reflectivity, speed of sound, and attenuation profiles, where several organs and features are clearly observed in humans. For needle localization, we couple ultrasound into a commercial biopsy needle, resulting in scattered signals at the tip which are detected and used to recover ~7 images per second over a human-scale cross-section. In ablation monitoring, we apply a modulated microwave signal to a commercial ablation probe to obtain thermoacoustic images of thermal deposition. We validate this technique using bovine liver during ablation. This suite of techniques may provide a convenient tool for diagnosing and treating a variety of abdominal conditions.

This work presents hybrid photoacoustic and ultrasound tomography, which enables three-dimensional panoramic imaging of the human body’s morphological and angiographic information to provide dual-contrast images of representative parts of human body (i.e., head, breast, and hand) with a single system. Through in vivo human application, we present our hybrid tomography system as a powerful tool for high-speed, three-dimensional, dual-contrast imaging of the human body with potential for rapid clinical translation.

Exploiting the optical absorption of hemoglobin, photoacoustic microscopy (PAM) has achieved label-free imaging of the microvasculature in vivo and enabled simultaneous quantification of blood oxygenation and flow. However, the axial resolution of PAM is limited to the mesoscopic level due to the finite bandwidth of detected ultrasound signals. To address this limitation, we have developed a super-resolution functional PAM technique based on spatiotemporal tracking of red blood cells, which enables label-free functional microvascular imaging in 3D at the single-cell level. We have demonstrated the utility of this technique by imaging the mouse brain’s responses to a single-vessel stroke in 3D.

Endocavity ultrasound (US) imaging is commonly used in gynecology and urology to diagnose challenging genital diseases. The convergence of photoacoustic (PA) imaging with clinical US imaging is actively explored in research due to their compatibility. However, the physical size of the laser delivery unit limits its insertion into confined vaginal or rectal canals. In our study, we performed video-rate endocavity PAUS imaging using a miniaturized probe. This achieved a fine radial resolution of 277 μm and 341 μm, and angular resolution of 1.52° and 1.23°. Multispectral PA imaging allowed quantification of a methylene blue (MB)/indocyanine green (ICG) cocktail ratio. The fast temporal resolution captured ICG infusion kinetics in ex vivo porcine ovary. Finally, we noninvasively imaged the male rat's reproductive system, confirming prostate vessel pulsation via PA and US imaging.

A high resolution all optical 3D photoacoustic scanner for superficial vascular imaging has been employed in a pilot study on patients with Peripheral Arterial Disease (PAD). The system can reliably produce high quality in-vivo photoacoustic images and is well tolerated by patients. The scanner can visualise the irregular vascular patterns in patients. In addition, by comparing the microvasculature of healthy volunteers and participants with CLTI we have demonstrated significant differences in the tortuosity, vessel size and vascular density between the 2 groups. The ability to visualise the lower limb microvasculature in detail in this way could be used to study small vessel-PAD linked to diseases such as diabetes with a view to informing diagnosis and treatment decision making.

We present photoacoustic computed tomography through an ergodic relay (PACTER), a method for single-shot 3D imaging of hemodynamics using a single-element detector. Our approach allows for ultrafast volumetric imaging at kilohertz rates without the need for numerous detector elements. We demonstrate PACTER in both human and small animal subjects, emphasizing its potential utility in early detection and monitoring of peripheral vascular diseases. Our single-element detector design aims to offer a more convenient and potentially affordable option, while the concept could also be relevant to other imaging technologies, contributing to various applications in medical imaging.

This abstract describes a potential method to improve the lateral resolution of Phonon microscopy, a novel noninvasive elasticity imaging microscopy for 3D cell imaging by measuring the time-resolved Brillouin scattering signal. While this technique provides sub-optical axial resolution, the lateral resolution is limited by the optical system that generates the coherent phonon fields. To overcome this limitation, the authors suggest using novel optoacoustic lenses working in GHz frequencies to focus the laser generated coherent phonon fields and thus obtain true acoustic resolution in both axial and lateral dimensions. These lenses can be fabricated at the nanoscale and can also be compatible with ultrasonic endoscopic imaging systems in further applications.

This study illustrates the potential of non-invasive Photoacoustic Microscopy (PAM) to monitor functional changes in a squirrel monkey brain due to peripheral mechanical stimulation. Our unique approach employs a deep Fully Convolutional Neural Network (FCNN) to significantly enhance PAM image quality, improving signal-to-noise ratio and structural similarity index. Notably, functional changes induced by peripheral mechanical stimulation were effectively observed. The study showcases the potential of PAM in neurological applications, advancing our understanding of brain hemodynamics, and the transformative effect of machine learning techniques on PAM image quality, opening new possibilities for future neuroscientific research.

We present the developments on a rotating multiple-view dual-mode photoacoustic/ultrasound system for in-vivo, non-invasive, whole-body small-animal imaging and based on planar Fabry-Pérot sensor-based tomography to overcome present challenges. Single planar Fabry-Pérot sensors suffer from an incomplete view of the acoustic fields, which leads to blurring and artefacts in tissue sample images. Increasing the fields of view would relax this limitation. Another contribution to the degradation of the image quality are wavefront aberrations stemming from spatially-varying sound speeds in a tissue sample and which limit the imaging depth. These can however be corrected by carrying out ultrasound computed tomography.

Longitudinal characterisation of the tumour vascular response to radiotherapy is essential for understanding the role of oxygenation and microvascular disruption in response to therapy. Using multi-scale in vivo photoacoustic imaging (PAI), we assessed early response to two hypofractionated radiotherapy schemes in two human breast cancer models. Mesoscopic and multispectral tomographic photoacoustic imaging was performed 24h pre-, post-radiotherapy, and at endpoint. PAI biomarkers were validated ex vivo with multiplex immunofluorescence using a 20-plex panel developed specifically for vascular response assessment at sub-cellular resolution. PAI captured radiotherapy response, revealing the differential effect between radiotherapy schemes and models with different hypoxia phenotypes.

Alzheimer’s disease (AD) is a neurodegenerative disorder usually accompanied by buildup of amyloid-beta (Aβ) plaques and neurofibrillary tangles, cerebral vascular morphological and functional changes. Our study presents a combined photoacoustic microscopy and fluorescence microscopy dual-modality imaging system for in-vivo monitoring of biomarkers for AD via cranial window. Photoacoustic microscopy resolves vasculature, blood oxygenation and flow speed at capillary level, while a confocally designed fluorescence microscope, aided by an Aβ targeting dye, maps the distribution of Aβ plaques. Results with age-varied AD mice group showed reduced vessel density, decreased perfusion at distal capillaries, and increased plaque density compared to wild type mice. This technique provides guidance for longitudinal monitoring of AD progression.

We demonstrate an ultra-wideband silicon-photonics acoustic detector (SPADE) with high sensitivity and signal fidelity, enabling highly detailed vascular imaging with resolutions comparable to optoacoustic microscopy. Our design is based on a planarized micro-ring resonator coated with an elastomer cladding in which the sensing is performed. The design of the guided optical mode and the use of planarization lead to high signal fidelity and a tomographic point-spread function that is in excellent agreement with the theoretical prediction. The new SPADE platform is demonstrated in vivo for imaging the vasculature of a mouse ear in high definition.

Our research utilizes capacitive micromachined ultrasonic transducers (CMUTs) for advanced photoacoustic (PA) and ultrasound (US) imaging in mice, focusing on visual-evoked hemodynamic responses. CMUTs exhibit superior sensitivity and wider bandwidth than traditional PZT arrays, enabling improved spatial resolution and image quality. Preliminary tests of a 1-D CMUT array demonstrated its potential in capturing hemodynamic responses in the primary visual cortex (V1) and superior colliculus (SC) during retinal photostimulation. With its impressive performance and miniaturization potential, CMUTs show promise for future applications in studying visual responses in transgenic mice and disease models without the need for anesthesia or restraint.

This study uses a label-free photoacoustic computed tomography (PACT) system with a linear ultrasound transducer array to monitor the visual-evoked hemodynamic responses in the mouse brain. Acoustic signals were collected during retinal photostimulation, utilizing flickering white light. The observed hemodynamic responses occurred in the primary visual cortex (V1), superior colliculus (SC), lateral geniculate nucleus (LGd), olivary pretectal nucleus (OPN), and suprachiasmatic nucleus (SCN). Response magnitudes and latencies were compared between wild-type, retinal degeneration (rd1), and melanopsin knock-out mice, illustrating the potential of our PACT system in studying brain activities related to retinal diseases.

Optical fiber enables the implementation of flexible medical endoscopes. Here, we present the development of fiber-optic endoscopic ultrasound, which utilizes laser pulse absorption to generate ultrasound waves and a fiber-optic acoustic sensor to detect echo waves. Compared to its piezoelectric counterpart, the fiber-optic sensor has a significantly higher detection sensitivity and broader bandwidth. As a result, we were able to perform in vivo rotational-scanning (or B-mode) imaging of the gastrointestinal tract and extraluminal structures of a rat with an operating frequency of 20 MHz, an imaging depth of 2 cm, and a frame rate of 1 Hz.

The application of a capsular structure to the distal section of a catheter-based probe has been a subject of significant investigation in various optical endoscopy technologies, such as endoscopic optical coherence tomography and optical-resolution photoacoustic endoscopy, particularly concerning their gastrointestinal imaging applications. In this study, we developed a photoacoustic capsule endoscope with an 8-mm diameter to bring these benefits to our gastrointestinal endoscopic applications. We specifically chose the 8 mm diameter for the first prototype probe to investigate its feasibility and associated technical issues based on a rat colorectum model. The achieved lateral resolution was as low as 10 μm, sufficient for resolving capillaries, and we successfully conducted in vivo imaging at an A-line acquisition rate of 32 kHz, satisfying the Nyquist sampling theory. As a result, we obtained a clear in vivo 3D microvasculature image from a rat colorectum over a 360° full angular region, overcoming the blind spot issue that is often caused by imperfect acoustic coupling between a catheter probe and a target lumen.

Curvilinear endocavity ultrasound images capture a wide field of view with a miniature probe. In adapting photoacoustic imaging (PAI) to work with such ultrasound systems, light delivery is challenged by the tradeoff between image quality and laser safety concerns. Here, we present two novel designs based on cylindrical lenses that are optimized for transvaginal PAI B-scan imaging. Our simulation and experimental results demonstrate that, compared to conventional light delivery methods for PAI imaging, the proposed designs are safer for higher pulse energies and provide deeper imaging and a wider lateral field of view. The proposed designs could also improve the performance of endoscopic co-registered ultrasound/ photoacoustic imaging in other clinical applications.

The limited frequency bandwidth of the ultrasound transducer used as receivers in photoacoustic imaging (PAI) has a significant impact on the resolution and depth of the reconstructed image. Therefore, for clinical applications, PAI may require the use of a low or high-frequency transducer depending on the application. Here we propose an adaptive spectral PAI algorithm based on prior information of multiple linear array ultrasound transducers ranging from 1-50 MHz, to retrieve tissue molecular components from micro to macro scale. The method includes multiple modules to coherently unmix molecular components at multi-scale. Results on numerical simulations and human imaging demonstrate that the adaptive method can effectively merge multiple-scale molecular imaging and substantially improve PAI quality.

The generation of realistically simulated photoacoustic (PA) images with ground truth labels for optical and acoustic properties has become a critical method for training and validating neural networks for PA imaging. As state-of-the-art model-based simulations often suffer from various inaccuracies, unsupervised domain transfer methods have been recently proposed to enhance the quality of model-based simulations. The validation of these methods, however, is challenging as there are no reliable labels for absorption or oxygen saturation in vivo. In this work, we examine various domain shifts between simulations and real images such as simulating the wrong noise model, inaccuracies in modeling the digital device twin or erroneous assumptions on tissue composition. We show in silico how a Cycle GAN, unsupervised image-to-image translation networks (UNIT) and a conditional invertible neural network handle these domain shifts and what their consequences are for blood oxygen saturation estimation.

Liver fibrosis is a global health burden characterized by excessive collagen deposition, impairing liver function. Noninvasive techniques that automatically visualize and quantify collagen content are needed for early detection and monitoring of fibrosis progression. We explored the potential of spectral photoacoustic imaging (sPAI) in monitoring collagen development during liver fibrosis. Here a novel data-driven superpixel PA unmixing (SPAX) framework, has been implemented to differentiate collagen presence and assess its correlation with fibrosis progression non-invasively without any a priori information. Overall, the in-vivo findings highlight the potential of sPAI and SPAX in non-invasively monitoring collagen dynamics and assessing fibrosis severity.

Ionizing radiation acoustic imaging (iRAI) is a passive and non-invasive imaging technique that allows for real-time mapping of the radiation dose deposition deep in the body during clinical radiation therapy. In this study, working toward truly quantitative iRAI imaging of radiation dose delivery, correction factors were calibrated through simulation studies to address the special variances in both the detection sensitivity of the 2D matrix array and the reconstruction sensitivity resulting from the delay-and-sum image reconstruction algorithm. By incorporating these compensation factors, volumetric iRAI was employed to accurately map radiation dose deliveries in a soft-tissue phantom. The study demonstrates the potential of iRAI in quantifying the deposition of radiation doses during treatment and facilitating online adaptive radiation therapy.

Knowledge of the positions of ultrasound transducer elements in a photoacoustic computed tomography (PACT) system is essential for reconstructing high-quality images. Errors in these positions, typically due to manufacturing defects, can degrade the reconstructed image quality perceivably. To overcome this, we present a calibration method for the ultrasound transducer array geometry that is based on the times-of-arrival of point source signals at the array elements. We pose the problem in terms of the speed of sound, the transducer positions, and the point source positions. We reformulate the problem as a linear problem in the transducer coordinates by obtaining the other unknowns using surrogate methods. Finally, we estimate the transducer coordinates using the pseudoinverse solution and characterize the estimation error in the coordinates. We use our method for calibrating an experimental PACT system, which results in an improvement in the contrast-to-noise ratio and resolution of point source reconstructions. Additionally, we reconstruct the images of a healthy human breast and show that the calibrated image reveals vasculatures that were previously not visible.

Accurate dose definition is vital for ensuring optimal radiation therapy (RT) outcomes. The combination of ionizing radiation acoustic imaging (iRAI) and volumetric ultrasound imaging (US) holds the potential for real-time and precise determination of the radiation dose on anatomical structures. We developed an iRAI-US dual-modality system, utilizing a custom 2D matrix array transducer for iRAI and a commercial 2D MAT for US. The studies on phantoms quantified the system performance, and then the experiments using a rabbit liver model in vivo achieved online monitoring of dose on anatomy during RT in real time. These findings demonstrated the potential of iRAI-US combined imaging for personalized RT with improved efficacy and safety.

We show that a flat plate acoustic encoder can enable fast, single-sensor photoacoustic imaging. To demonstrate the concept, we used an off-the-shelf microscope slide with a sample consisting of black tape. The glass slide needs only to be calibrated once through raster scanning. Using a different glass slide, an image of a patterned black tape was obtained within 5 seconds. This represents a 480 times imaging time reduction compared to a raster-scan based system for a similar field of view . Our result shows that a simple, flat plate encoder can be a practical solution towards fast, single-sensor photoacoustic imaging.

The International Photoacoustic Standardisation Consortium (IPASC) hereby reports the results of a consensus-finding exercise undertaken to agree recommendations for properties of tissue-mimicking phantom materials and their characterization. Guidelines on material properties are given under defined environmental conditions and include, for example, recommendations on optical and acoustic properties, stability, and structural composition. Multi-center studies involving independent fabrication of phantom material batches are encouraged for reproducibility and verification of properties within acceptable ranges. The recommendations aim to support researchers and manufacturers to develop phantoms that facilitate system performance assessment, inter-device comparisons, and system optimization, ultimately advancing photoacoustic technology.

Non-invasive assessment of aggressive tumors with high spatial resolution and specificity has been challenging. This study presents an application of multispectral photoacoustic tomography (MSPT) in tumor imaging, using a specially designed hypoxia-responding agent, NOx-JS013. The developed hypoxia probe, NOx-JS013, is reduced to JS013 under hypoxic conditions common in aggressive tumors, with a shifted absorption peak. Additionally, NOx-JS013 specifically binds with NCEH1, a characteristic enzyme highly expressed in aggressive tumor cells. The results from In-vivo experiments on aggressive and non-aggressive tumors demonstrate the feasibility of the MSPT hypoxia imaging using NOx-JS013 hypoxia indicator enhanced with aggressive cancer cell targeting ability.

Bio-compatible contrast agents based on clinically-approved indocyanine green (ICG) can greatly enhance the deep-tissue imaging performance of optoacoustic (OA) imaging systems and facilitate the clinical translation of this modality. In this work, we propose an inverse emulsion approach to synthetize bio-based, bio-degradable nano- and micro-capsules consisting of an aqueous core of ICG and a cross-linked casein shell. The feasibility to visualize and track individual capsules with a diameter of 4-5 micrometers is demonstrated in vitro and in vivo. This can pave the way towards clinical approval of contrast agents capable of being detected at a single-particle level.

The feasibility to individually localize and manipulate individual particles flowing in blood can lead to important advances in super-resolution imaging, targeted drug delivery, and other fields. State-of-the-art optoacoustic (OA) tomographic imaging systems provide a unique high frame rate imaging capability in three dimensions, which can be exploited for this purpose if particles are sufficiently absorbing. In this work, we introduce silica-core microparticles with a polypyrrole-gold composite shell deposited with a layer-by-layer approach. Microparticles as small as 2 microns could be individually detected. Laser-induced motion of the particles was also observed, which provides a new means for motion control.

Validating processing algorithms for photoacoustic images is complex due to a gap between simulated and experimental data. To address this challenge, we present a multi-device dataset of well-characterised phantoms and investigate the simulation gap using a supervised calibration of the forward model. We use N=15 phantoms for calibration and systematically compare simulated and experimental data from the remaining N=15 phantoms. Our results highlight the importance of the device geometry, impulse response, and noise for accurate simulation. By reducing the simulation gap and providing an open dataset, our work will contribute to advancing data-driven photoacoustic image processing techniques.

Photoacoustic imaging (PAI) is rapidly emerging for many clinical applications involving vascular mapping and tissue oximetry. Phantom-based evaluation of oximetry accuracy is invaluable for device evaluation but requires specialized equipment and use of potentially biohazardous blood. We developed stable, tunable blood-mimicking dye solutions that can simulate photoacoustic signals of blood at 750 nm and 850 nm for 40-100% SO2. We filled breast-mimicking polyacrylamide hydrogel phantoms with either dye solutions or deoxygenated bovine blood to test a custom PAI system. Results showed that the blood-mimicking dyes produced similar levels of SO2 accuracy in phantoms compared to blood samples, although some differences in sensitivity and root-mean-square difference were observed. These blood-mimicking dyes may offer a simpler, safer approach to evaluating oximetry measurement accuracy in PAI devices that could accelerate device evaluation and clinical translation.

Photoacoustic imaging (PAI) has started to emerge in clinics for cancer care from early detection to therapy monitoring. Herein, we encapsulate indocyanine green J-aggregates (ICGJ), an FDA-approved NIR contrast agent, into biodegradable polymersomes (ICGJ-Ps) with high loading for strong photoacoustic signals. The small Ps hydrodynamic diameter of 80-90nm is beneficial for the ability to penetrate tumors and for uptake in cancer cells. The high ICGJ loadings up to ~40% produce strong photoacoustic (PA) signals. Conjugation of ICGJ-Ps with cetuximab antibodies produced PA signals in breast cancer cells that correlated well with their inherent EGFR expression levels indicating exceptional specificity.

Members of IPASC have created an open-source library for image reconstruction algorithms that are compatible with the IPASC data format. Within this project, we create a testing framework for the evaluation of image reconstruction algorithms to identify their context-dependent strengths and weaknesses. We develop an open-access dataset comprising both simulated and experimental data to facilitate collaboration among all stakeholders associated with photoacoustic imaging and lower the barrier of entry for new researchers in the field by making the project deliverables available open-source.

This work develops a thorough and user-friendly methodology for fluence compensation of photoacoustic (PA) data from the commercial Visualsonics (VS) VevoLAZR-X System. PHANTOM (PHotoacoustic ANnotation TOolkit) has been created to help fluence compensate large PA datasets. PHANTOM provides an easy-to-use graphical user interface to assist users in segmenting US-coupled PA images into three-dimensional labeled tissue structures with assignable optical characteristics. We also modeled an MCX-compatible light source configuration that replicates the VS system's optical fiber illumination. PHANTOM exports all Monte Carlo configuration parameters which are compatible with Monte Carlo eXtreme (MCX) allowing fluence compensation of PA data.

In this work, a novel PAQUS apparatus was developed for in vivo assessment of the calcaneus. We first made lattice-shaped cubic phantoms of varying density, resulting in different BVF. PA absorbance at 740 nm (hemoglobin) and 930 nm (lipid) were significantly different across phantoms. Next, we created two anatomic phantoms based on a human calcaneus microCT volumetric image, and artificially eroded the trabecular structure on one to reduce BVF by 32.5%, mimicking osteoporosis-mediated bone loss. Phantoms were measured repeatedly and also significantly differed. Lastly, in vivo bilateral PA measurement was conducted at the calcaneus on 32 female Caucasian subjects (20-74 yrs). A significant difference in PA absorption was observed at 740 nm between younger (<50 yrs) and older (>50 yrs) subjects.

We report on our clinical trial findings for using quantitative ultrasound and photoacoustic imaging for assessing kidney transplant quality.

Longitudinal mesoscopic photoacoustic imaging of vascular networks requires accurate image co-registration to assess local changes in growing tumours, but remains challenging due to sparsity of data and scan-to-scan variability. Here, we compared a set of 5 curated co-registration methods applied to 49 pairs of vascular images of mouse ears and breast cancer xenografts. Images were segmented using a generative adversarial network and pairs of images and/or segmentations were fed into the 5 tested algorithms. We show the feasibility of co-registering vascular networks accurately using a range of quality metrics, taking a step towards longitudinal characterization of those complex structures.

Multispectral photoacoustic tomography (PAT) offers high-resolution images of deep tissue oxygen saturation (sO2), but the complexity of photon absorption and scattering affects sO2 accuracy. This study applied a rigorous light transport model, revealing that PA spectra within biological tissue can be represented as convex cones (CCs) in a high-dimensional space. Using the CC model, sO2 can be estimated by finding the nearest CC to measured data, even in noisy conditions. This method combines a physical model with machine learning, demonstrating practicality and robustness in numerical, phantom, and in vivo imaging experiments, with an average sO2 estimation error of just 3% in human trials. Additionally, it outperforms clinical practices like linear spectral unmixing, suggesting broader applications in PA molecular imaging and diffuse optical imaging.

Glaucoma is a leading cause of blindness. Previously, we quantified the deformations in scleral tissue components through a PAM-FEA methodology. This study furthers our examination of perilimbal sclera (PLS) and aqueous veins in intraocular pressure (IOP) regulation using ex vivo porcine eyes. Our results reveal that the cross-sectional area change of the aqueous veins and strain of PLS are strongly and positively correlated with the steady-state IOP (n=8, R2=0.90, R2=0.57, respectively) and the IOP elevation rate at a constantly increasing flow rate (n=8, R2=0.89, R2=0.58, respectively). These insights bolster our supposition that a stiffer PLS can lead to heightened IOP and reduced ocular adaptability to aqueous outflow, primarily due to restrictions on aqueous vein dilation.

Although several optical imaging modalities can acquire high-resolution images, the light penetration depth is limited because optical scattering occurs in a tissue, making it difficult to acquire deep-depth images. For this reason, we previously suggested and verified that ultrasound-induced optical clearing microscopy (US-OCM) can suppress optical scattering and improve the depth of light penetration. However, it is challenging to accurately predict the nucleation position of a bubble cloud due to randomly generated within the depth of focus. In this study, we proposed a combined acoustic and optical cavitation-induced gas bubbles method that can generate a bubble cloud at an accurately predicted location, and demonstrate image performance through various experiments.

Pathologist evaluation of brightfield microscopy images of H&E-stained tissues is time-consuming and labor-intensive, preventing intraoperative feedback to surgeons. Combined ultraviolet reflectance and photoacoustic remote sensing (PARS) microscopy was recently introduced, providing a label-free way to obtain cell nuclei and cytoplasm images. Additionally, utilizing deep learning and fast voice-coil scanning, maximally realistic virtual histology can be rapidly generated with this technique. Here, we introduce combined ultraviolet confocal reflectance, autofluorescence, and PARS microscopy, utilizing a single 266 nm excitation beam. This system now allows us to generate sectioned virtual histology in thick tissues well suited to the intraoperative setting.

We present a novel implementation of high-speed photoacoustic microscopy with an ultrawide field of view (FOV). Existing photoacoustic microscopy systems have limitations in resolution, FOV, and speed. To address these challenges, we improved our previous immersible-polygon-scanner-based photoacoustic microscopy and developed a dual-view photoacoustic microscopy system using two facets of the same scanner. The system has achieved high resolutions of ~10 µm (lateral) and 35 µm (axial) with a B-scan rate of 500 Hz. We have demonstrated the imaging performance on freely-moving zebrafish and hypoxia challenge of a pair of mice.

High-resolution optoacoustic imaging is conventionally performed using a microscopy setup where a strongly focused ultrasound transducer samples the image object point-by-point. Although recent advancements in miniaturized ultrasound detectors enables to achieve microscopic resolution with an unfocused detector in a tomographic configuration, such an approach requires illuminating the entire object, leading to an inefficient use of the optical power, and imposing a trans-illumination configuration that is limited to thin objects. We developed an optoacoustic micro-tomography system in an epi-illumination configuration, in which the illumination is scanned with the detector. The system is demonstrated in phantoms for imaging depths of up to 5 mm and in vivo for imaging the vasculature of a mouse ear. Although image-formation in optoacoustic tomography generally requires static illumination, our numerical simulations and experimental measurements show that this requirement is relaxed in practice due to light diffusion, which homogenizes the fluence in deep tissue layers.

Ultraviolet photacoustic remote sensing (UV-PARS) microscopy is a non-contact, label-free imaging modality which has demonstrated the ability to generate virtual histology images that show good concordance to gold-standard hematoxylin and eosin (H&E) stains of fixed tissue sections. However, UV-PARS microscopy requires time-consuming and complex coalignment of multiple beams for imaging. In this work, we demonstrate UV transmittance and scattering microscopy for virtual histology and compare with UV-PARS microscopy. Maximally realistic virtual histology images are generated for both UV-PARS and UV-transmittance microscopy techniques using a cycle-consistent generative adversarial network (CycleGAN) and compared to one another alongside the gold-standard brightfield microscopy of H&E stained fixed tissues.

In this work, a PAM and OCT combined dual-modality imaging system was applied for non-invasive in vivo imaging of angiogenesis. Two groups of fibrin scaffolds, with and without basic fibroblast growth factor (bFGF), were implanted and covered by transparent imaging window for longitudinal observation. Imaging experiments were conducted on days 3, 5, 7, and 9 to monitor vascularization. Several vasculature related parameters were statistically compared and significant differences were discovered, indicating the effect of bFGF in promoting angiogenesis. By counting the vessel density in histological sections, similar statistical results consistent with the imaging was acquired, further validating the imaging accuracy.

Optical imaging of human midbrain organoids in the field of Parkinson research is challenging using current established technologies since they require lengthy and invasive clearing methods, which preclude live organoid observations and longitudinal studies.

Raster-Scanning Optoacoustic Mesoscopy (RSOM) overcomes this limitation by utilizing the intrinsic optical absorption contrast based on the optoacoustic effect. We show that RSOM is able to provide fast quantitative visualization of the neuromelanin content in intact large organoids at single cell resolution.

We developed a flexible and stretchable photoacoustic (PA) transmitter for finger-bending detection. The thin-film PA transmitter consists of dual-layer light-absorbing films, whose thickness and PA signal can change with a strain applied onto the transmitter. The transmitter can be fabricated with a wide area and conformally attached on a curved surface. As an example to demonstrate this flexible and stretchable transmitter performance, we utilized it for finger-bending detection. This exhibited its capability of detecting a minute change of force-induced finger motion (variation in tens of micrometers) with negligible hysteresis. Leveraging these advantageous properties, we characterized the output performance of PA transmitter according to finger-bending angles ranging from 0 to 90 degree. This shows a great potential for various types of motion detection with precise and reliable strain sensing capabilities.

Photoacoustic Imaging faces limitations such as limited acoustic detection and low optical propagation depth, resulting in poor image quality, low signal-to-noise ratio and resultant shallow imaging depth. The research evaluates the performance of various deep learning techniques such as convolutional layers, residual layers, vision transformers when combined with generative adversarial models help in enhancing the quality of photoacoustic images. The evaluation shows promising results for deep learning to improve photoacoustic images to improve the signal-to-noise ratio and enhance the imaging depth in the form of deeper vascular structure in the model outputs.

Transcranial photoacoustic imaging shows promise for non-invasive evaluation of brain injury and blood-brain barrier disruption (BBB-D). Our study used neonatal rats with immature cerebral blood vessels, making them more susceptible to brain injury. Neuroinflammation was induced using lipopolysaccharide (LPS), leading to BBB-D. We employed a small-animal photoacoustic imaging system that integrated a wavelength-tunable laser (680-970 nm) and a high-frequency ultrasound transducer to obtain transcranial ultrasound and photoacoustic (PA) images. BBB-D was visualized by the migration and accumulation of indocyanine green (ICG) J-aggregate nanoprobes in the brain, resulting in enhanced PA signal. Following LPS injection, a two-fold increase in PA signal intensity was observed at 2 hours, peaking at a four-fold increase at 4 hours. The enhanced PA signal persisted up to 24 hours and remained within 30% of the baseline at 48 hours. These findings have significant implications for early detection of BBB-D using transcranial photoacoustic imaging, made possible by the use of neonatal rats with thin skulls and photoacoustic contrast agents with distinct spectral signatures in vivo.

Photoacoustic imaging holds promise in clinical applications, but lacks standardized testing methods. To overcome this, the International Photoacoustic Standardization Consortium (IPASC) assessed the fabrication reproducibility of a stable tissue-mimicking phantom material in an international multicenter study (n>15 centers). The material consists of mineral oil, polymer, ink, and titanium dioxide. Participating centers followed a recipe set up by the main site (Cambridge, UK) and returned samples for characterization. The results demonstrate promising reproducibility for acoustic, photoacoustic and optical properties. By performing this study, IPASC hopes to broaden the uptake of a stable phantom material, supporting system validation and testing.

Near-infrared Photoacoustic imaging (NIR-PA) can enable deep-tissue imaging, yet clinical translation has been hindered by a lack of suitable NIR-PA contrast agents. The FDA-approved Indocyanine green (ICG) dye is a promising candidate, but it offers limited targeting ability and poor stability. To address this unmet need, we examined three novel ICG-based platforms in the form of DNA scaffolds, J-aggregates, and nanobubbles. We demonstrate that all three platforms yield a PA signal stronger than whole blood at concentrations as low as 45 µM and are amenable to molecular targeting.

Histopathology is the fundamental microscopic examination tool in many domains including cancer diagnosis and surgical oncology, and various fields. Recently, the UV-PAM is powerful imaging system for label-free histology-like imaging. In histopathology, the nuclei in cancer cells have typical morphological features, such as irregular shapes and large sizes. Alternatively, UV-PAM can obtain unstrained cell nuclei with specific, positive, and high-contrast images. UV-PAM is useful histological images of unstained cell nuclei with specific, positive, and high-image contrast. By using the 266 nm light, high contrast photoacoustic images can be obtained. In the recent UV-PAM, the hardware has been faced a fundamental limitation to laser source especially. There is no available commercial nanosecond (ns) laser that has a high repetition rate operation of hundreds of kHz at 266 nm. Here, we develop a stable high-repetition rate 266 nm source using the basic second harmonic generation (SHG) phenomenon. This stable 266 nm source is based on a simplified architecture using single-pass SHG with BBO crystal.

Synaptic dysfunction driven by toxic amyloid Aβ oligomers (Aβo) is one of the early events in Alzheimer’s disease (AD) leading to cognitive decline, disability, and death. We have shown that optoacoustic therapy/NPLT can be used for traumatic brain injury (TBI) treatment. We have also shown that mature neurons generated from the NPLT-treated rat hippocampal neural stem cells (NSCs) had significantly lower Aβo levels. Here we used a custom-made 808-nm optoacoustic system to study autophagy pathways changes induced by NPLT. Our results suggest that autophagy pathways are upregulated in NPLT-treated NSCs compared to sham which could lead to enhanced Aβo clearance.

Colorectal cancer ranks as the second most prevalent global malignancy and stands as the fourth leading cause of cancer-related deaths in the United States. The ability to accurately monitor rectal cancer treatment responses poses a significant challenge given the limitations of existing imaging modalities in confirming pathological complete response after chemoradiation. Non-invasive confirmation of complete response can offer improved quality of life, reduced medical costs, and decreased strain on the healthcare system for patients. Previous findings from our research group highlighted the potential of co-registered acoustic-resolution Photoacoustic Microscopy (ARPAM) and ultrasound (ARPAM/US) in monitoring treatment response, revealing the recovery of regular microvascular patterns in the tumor bed through photoacoustic microscopy in treatment responders. In this presentation, we introduce a second-generation compact and robust ARPAM/US system designed for monitoring rectal cancer treatment responses in an endoscopy unit suitable for repeated imaging. We will present a comparative analysis between normal tissue and tumor bed with and without residual tumor after chemoradiation.

Photoacoustic tomography (PAT) is a promising hybrid imaging technique with clinical potential, but it faces challenges due to limited-view reconstruction. This research develops a deep learning-based approach using a multi-view imaging system and a Uformer network to reconstruct high-resolution images from limited-angle input data. The results show state-of-the-art performance compared to conventional restoration models, highlighting the potential of this method for improving PAT in clinical settings. This novel strategy helps overcome limited-data challenges and contributes to the development of innovative imaging solutions for clinical applications.

In this work, we demonstrated the potential that multimodality imaging of angiogenesis in biomaterial scaffolds can detect early-stage breast cancer in a murine model. By conducting in vivo photoacoustic microscopy (PAM) imaging of vascularization on the tumor-bearing and tumor-free groups, significant differences of several vasculature related parameters were found after tumor inoculation, and more neovascularization was discovered in tumor-bearing group. Additionally, optical coherence tomography (OCT) imaging was performed to provide 3D morphology information of the scaffold which offered similar microenvironment for tumor growth. Imaging results were also validated by IHC histology analysis, further suggesting the accuracy of the imaging.

LED-based photoacoustic systems have potential to diagnose diseases and tumors with high sensitivity and specificity at a cost that is affordable for all clinics. However, the expensive high-power pulsed Q-switch lasers still provide better image quality than LEDs. They also use piezoelectric transducers that are limited in sensitivity, and noise when miniaturized. Here, we present a low-cost LED-based photoacoustic imaging system with our highly sensitive optomechanical ultrasound sensor (OMUS), which is only limited by thermomechanical noise. In future, the cost of the OMUS read-out and multiplexing will enable clinical translation of in vivo small animal studies.

Photoacoustic microscopy (PAM) allows for the visualization of microscale structures and functionalities by detecting various chromophores in biological tissues based on their absorption spectra at specific laser wavelengths. PAM differentiates oxy and deoxy blood, water, collagen, and lipid with unique absorption spectra in NIR. To enhance the functionality of PAM, integrating multi-wavelength laser sources, particularly Ti:Sapphire lasers, is gaining significant attention. Ti:Sapphire lasers are considered an advanced solution due to their high energy efficiency and wavelength tunability. This presentation introduces a single-shot, wavelength-tunable Ti:Sapphire-based multispectral PAM system capable of rapidly performing functional imaging of blood concentrations and ICG-lymphography.

Accurate characterization of carotid plaque composition is essential to identify vulnerable plaque that often leads to stroke. Photoacoustic imaging, which combines optical absorption contrast with ultrasonic imaging depth, shows promise for quantitative examination of carotid plaque. However, unknown light fluence in the tissue makes quantitative photoacoustic imaging challenging. We propose utilizing a known chromophore as a light fluence marker. The feasibility of the approach was tested using simulations on digital phantoms and experiments using tissue-mimicking phantoms. The results show agreement with the actual concentrations, supporting our hypothesis. We intend to extend the approach to ex vivo plaque imaging.

Photoacoustic imaging is a powerful biomedical imaging modality with the unique capability to perform many challenging imaging tasks due to its molecular contrast based on optical absorbers in tissue. This makes it ideal for applications such as functional brain imaging, and early detection of diseases such as cancer. By combining the optical detection method of a Fabry-Pérot Etalon with ultrafast single-shot compressed sensing techniques, this work aims to increase the temporal resolution of photoacoustic imaging by up to three orders of magnitude. This would allow real-time volumetric photoacoustic imaging while reducing complexity and cost, enabling difficult imaging tasks and expanding the clinical applicability.

Endometriosis is a chronic inflammatory disease in women aged from 20 to 40. While laparoscopy and MRI are commonly used methods for detecting endometriosis, its diagnosis remains a challenge due to variations in location of endometrial tissue and symptoms which can vary between individuals. We are designing a low-cost photoacoustic imaging system to detect endometriosis in vivo without the insertion of a transvaginal probe or fluorescence. In our talk, we will present the design of our system, compare and analyze our preliminary results, and discuss the feasibility of our approach in non-invasive in vivo detection of endometriosis.

This study delves into the largely uncharted domain of biases in photoacoustic imaging, spotlighting potential shortcut learning as a key issue in reliable machine learning. Our focus is on hardware variation biases. We identify device-specific traits that create detectable fingerprints in photoacoustic images, demonstrate machine learning's capability to use these discrepancies to determine the device that acquired the image, and highlight their potential impact on machine learning model predictions in downstream tasks, such as disease classification.

Data-driven approaches to the quantification problem in photoacoustic imaging have shown great potential in silico, but the inherent lack of labelled ground truth data in vivo currently restricts their application and translation into clinics. In this study we leverage Fourier Neural Operator networks as surrogate models to synthesize multispectral photoacoustic human forearm images in order to replace time-consuming and not inherently differentiable state-of-the-art Monte Carlo and k-Wave simulations. We investigate the accuracy and efficiency of these surrogate models for the optical and acoustic simulation step.

Deep learning (DL) is a powerful reconstruction approach now applied broadly in photoacoustic (PA) imaging. However, in practical applications where the ground-truth is unknown, the reliability of predicted PA images from the trained DL network cannot be quantified. Here, we present a new DL approach to simultaneously estimate segmentation (source location) and PA images with uncertainty quantification based on the Bayesian convolutional neural network (BCNN). The BCNN was trained on simulated PA images and tested on both simulated and experimental PA images. The results show accurate segmentation and PA predictions as well as reliable PA uncertainty predictions.

This work introduces a total-peptide-based contrast agent for deep-tissue photoacoustic imaging, addressing the limitations of conventional imaging techniques. Chlorophyll a (Chla) was selected for its light-harvesting ability and solubilized using the Leipidium virginicum water-soluble chlorophyll-binding protein (LvP). LvP-chla enables clear visualization in vivo, discernible from the blood via spectroscopic PA imaging. LvP-chla showed promise as a favorable candidate for clinical photoacoustic imaging applications.

Unstable plaques in the carotid artery are a leading cause of ischemic stroke, resulting in numerous deaths globally. A reliable predictor of rupture risk is plaque composition. Photoacoustic Imaging is a promising imaging modality suitable for this task. Quantifying Photoacoustic Imaging poses challenges as it requires knowledge of complex light propagation in tissue. Our approach utilizes arterial blood as a fluence marker to determine the optical properties of the plaque. This proposed method involves optical and acoustic modeling and iterative updating of the optical properties. We characterized the proposed technique by varying different parameters in digital phantoms.

Optoacoustic tomography (OAT) holds promise for enhancing neuroimaging in the mouse brain. It provides the opportunity for multispectral, mesoscopic, high temporal resolution imaging of neural activity using exogenous and endogenous contrasts. OAT data also presents challenges, such as the need to optimize reconstruction and spectral unmixing parameters as well as the need to overcome light attenuation and nonuniformity. We demonstrate methods to optimize OAT acquisition and analysis for functional neuroimaging, including using a hybrid fluorescence-OAT approach.

Solid tumors represent abnormal tissue masses where surgical resection is often the primary treatment. However, positive margins increase local recurrence. Near-infrared fluorescence (NIRF) can intraoperatively identify solid tumors and assess margins. However, fluorescent light is diffused at depth, resulting in poor localization. Here, we test two lab-built systems (fast-sweep spectroscopic photoacoustic-ultrasound (PAUS) and NIRF multimodal scanning fiber endoscope) in a pilot ex vivo study to localize FDA-approved Tumor Paint^® tozuleristide (BLZ-100, Blaze Bioscience, Inc.) tumor-targeted contrast agent in muscle for real-time guidance (tumor excision) by combining them for pre- (PAUS) and intra- (NIRF) operative solid tumor detection.

We introduce an innovative photoacoustic tomography approach for volumetric imaging that addresses the limitations of traditional piezoelectric ultrasound transducers. By using non-contact optical detection via digital holography, coupled with a flexible PDMS cover layer, our method maps surface displacements induced by photoacoustic waves on rough biological tissues with sub-nanometer precision. We also develop a two-step backpropagation algorithm that compensates for wave component attenuation and adjusts acoustic velocity. Our method successfully records vascular images of a mouse's hind leg and a chicken embryo's chorioallantoic membrane with a lateral resolution of 140 μm.