Adaptive machine learning method for photoacoustic computed tomography based on sparse array sensor data.

BACKGROUND AND OBJECTIVE: Photoacoustic computed tomography (PACT) is a non-invasive biomedical imaging technology that has developed rapidly in recent decades, especially has shown potential for small animal studies and early diagnosis of human diseases. To obtain high-quality images, the photoacoustic imaging system needs a high-element-density detector array. However, in practical applications, due to the cost limitation, manufacturing technology, and the system requirement in miniaturization and robustness, it is challenging to achieve sufficient elements and high-quality reconstructed images, which may even suffer from artifacts. Different from the latest machine learning methods based on removing distortions and artifacts to recover high-quality images, this paper proposes an adaptive machine learning method to firstly predict and complement the photoacoustic sensor channel data from sparse array sampling and then reconstruct images through conventional reconstruction algorithms. METHODS: We develop an adaptive machine learning method to predict and complement the photoacoustic sensor channel data. The model consists of XGBoost and a neural network named SS-net. To handle data sets of different sizes and improve the generalization, a tunable parameter is used to control the weights of XGBoost and SS-net outputs. RESULTS: The proposed method achieved superior performance as demonstrated by simulation, phantom experiments, and in vivo experiment results. Compared with linear interpolation, XGBoost, CAE, and U-net, the simulation results show that the SSIM value is increased by 12.83%, 6.78%, 21.46%, and 12.33%. Moreover, the median R2 is increased by 34.4%, 8.1%, 28.6%, and 84.1% with the in vivo data. CONCLUSIONS: This model provides a framework to predict the missed photoacoustic sensor data on a sparse ring-shaped array for PACT imaging and has achieved considerable improvements in reconstructing the objects. Compared with linear interpolation and other deep learning methods qualitatively and quantitatively, our proposed methods can effectively suppress artifacts and improve image quality. The advantage of our methods is that there is no need for preparing a large number of images as the training dataset, and the data for training is directly from the sensors. It has the potential to be applied to a wide range of photoacoustic imaging detector arrays for low-cost and user-friendly clinical applications.

Photoacoustic imaging with limited sampling: a review of machine learning approaches.

Photoacoustic imaging combines high optical absorption contrast and deep acoustic penetration, and can reveal structural, molecular, and functional information about biological tissue non-invasively. Due to practical restrictions, photoacoustic imaging systems often face various challenges, such as complex system configuration, long imaging time, and/or less-than-ideal image quality, which collectively hinder their clinical application. Machine learning has been applied to improve photoacoustic imaging and mitigate the otherwise strict requirements in system setup and data acquisition. In contrast to the previous reviews of learned methods in photoacoustic computed tomography (PACT), this review focuses on the application of machine learning approaches to address the limited spatial sampling problems in photoacoustic imaging, specifically the limited view and undersampling issues. We summarize the relevant PACT works based on their training data, workflow, and model architecture. Notably, we also introduce the recent limited sampling works on the other major implementation of photoacoustic imaging, i.e., photoacoustic microscopy (PAM). With machine learning-based processing, photoacoustic imaging can achieve improved image quality with modest spatial sampling, presenting great potential for low-cost and user-friendly clinical applications.

Absolute Grüneisen parameter measurement in deep tissue based on X-ray-induced acoustic computed tomography.

The Grüneisen parameter is a primary parameter of the initial sound pressure signal in the photoacoustic effect, which can provide unique biological information and is related to the temperature change information of an object. The accurate measurement of this parameter is of great significance in biomedical research. Combining X-ray-induced acoustic tomography and conventional X-ray computed tomography, we proposed a method to obtain the absolute Grüneisen parameter. The theory development, numerical simulation, and biomedical application scenarios are discussed. The results reveal that our method not only can determine the Grüneisen parameter but can also obtain the body internal temperature distribution, presenting its potential in the diagnosis of a broad range of diseases.

Recent Advances in Flexible Ultrasonic Transducers: From Materials Optimization to Imaging Applications.

Ultrasonic (US) transducers have been widely used in the field of ultrasonic and photoacoustic imaging system in recent years, to convert acoustic and electrical signals into each other. As the core part of imaging systems, US transducers have been extensively studied and achieved remarkable progress recently. Imaging systems employing conventional rigid US transducers impose certain constraints, such as not being able to conform to complex surfaces and comfortably come into contact with skin and the sample, and meet the applications of continuous monitoring and diagnosis. To overcome these drawbacks, significant effort has been made in transforming the rigid US transducers to become flexible and wearable. Flexible US transducers ensure self-alignment to complex surfaces and maximize the transferred US energy, resulting in high quality detection performance. The advancement in flexible US transducers has further extended the application range of imaging systems. This review is intended to summarize the most recent advances in flexible US transducers, including advanced functional materials optimization, representative US transducers designs and practical applications in imaging systems. Additionally, the potential challenges and future directions of the development of flexible US transducers are also discussed.

Cantilever-enhanced photoacoustic spectroscopy for gas sensing: A comparison of different displacement detection methods.

Photoacoustic spectroscopy (PAS) combines the advantages of high sensitivity, high specificity and zero background, which is very suitable for trace gas detection. Cantilever-enhanced photoacoustic spectroscopy (CEPAS) utilizes highly sensitive mechanical cantilevers to further enhance the photoacoustic signal and shows a gas concentration detection limit of parts per trillion. This review is intended to summarize the recent advancements in CEPAS based on different displacement detection methods, such as Michelson interference, Fabry-Perot interference, light intensity detection, capacitive, piezoelectric and piezoresistive detection. Fundamental mechanisms and technical requirements of CEPAS are also provided in the literature. Finally, potential challenges and further opportunities are also discussed.

Non-Invasive Monitoring of Human Health by Photoacoustic Spectroscopy.

For certain diseases, the continuous long-term monitoring of the physiological condition is crucial. Therefore, non-invasive monitoring methods have attracted widespread attention in health care. This review aims to discuss the non-invasive monitoring technologies for human health based on photoacoustic spectroscopy. First, the theoretical basis of photoacoustic spectroscopy and related devices are reported. Furthermore, this article introduces the monitoring methods for blood glucose, blood oxygen, lipid, and tumors, including differential continuous-wave photoacoustic spectroscopy, microscopic photoacoustic spectroscopy, mid-infrared photoacoustic detection, wavelength-modulated differential photoacoustic spectroscopy, and others. Finally, we present the limitations and prospects of photoacoustic spectroscopy.

A Review of High-Frequency Ultrasonic Transducers for Photoacoustic Imaging Applications.

Photoacoustic imaging (PAI) is a new and rapidly growing hybrid biomedical imaging modality that combines the virtues of both optical and ultrasonic (US) imaging. The nature of the interaction between light and ultrasound waves allows PAI to make good use of the rich contrast produced by optics while retaining the imaging depths in US imaging. High-frequency US transducers are an important part of the PAI systems, used to detect the high-frequency and broad-bandwidth photoacoustic signals excited by the target tissues irradiated by short laser pulses. Advancement in high-frequency US transducer technology has influenced the boost of PAI to broad applications. Here, we present a review on high-frequency US transducer technologies for PAI applications, including advanced piezoelectric materials and representative transducers. In addition, we discuss the new challenges and directions facing the development of high-frequency US transducers for PAI applications.

Recent advances in high-speed photoacoustic microscopy.

Photoacoustic (PA) microscopy (PAM) has achieved remarkable progress in biomedicine in the past decade. It is a fast-rising imaging modality with diverse applications, such as hemodynamics, oncology, metabolism, and neuroimaging. Combining optical excitation and acoustic detection, the hybrid nature of PAM provides advantages of rich contrast and deep penetration. In recent years, high-speed PAM has flourished and enabled high-speed wide-field imaging of functional activity. In this review, we summarize the most recent advances in high-speed PAM technologies, including high-repetition-rate multi-wavelength laser development, fast scanning techniques, and novel PA signal acquisition strategies.

In vivo characterization of connective tissue remodeling using infrared photoacoustic spectra.

Premature cervical remodeling is a critical precursor of spontaneous preterm birth, and the remodeling process is characterized by an increase in tissue hydration. Nevertheless, current clinical measurements of cervical remodeling are subjective and detect only late events, such as cervical effacement and dilation. Here, we present a photoacoustic endoscope that can quantify tissue hydration by measuring near-infrared cervical spectra. We quantify the water contents of tissue-mimicking hydrogel phantoms as an analog of cervical connective tissue. Applying this method to pregnant women in vivo, we observed an increase in the water content of the cervix throughout pregnancy. The application of this technique in maternal healthcare may advance our understanding of cervical remodeling and provide a sensitive method for predicting preterm birth.

Transvaginal fast-scanning optical-resolution photoacoustic endoscopy.

Photoacoustic endoscopy offers in vivo examination of the visceral tissue using endogenous contrast, but its typical B-scan rate is ∼10 Hz, restricted by the speed of the scanning unit and the laser pulse repetition rate. Here, we present a transvaginal fast-scanning optical-resolution photoacoustic endoscope with a 250-Hz B-scan rate over a 3-mm scanning range. Using this modality, we not only illustrated the morphological differences of vasculatures among the human ectocervix, uterine body, and sublingual mucosa but also showed the longitudinal and cross-sectional differences of cervical vasculatures in pregnant women. This technology is promising for screening the visceral pathological changes associated with angiogenesis.

Dry coupling for whole-body small-animal photoacoustic computed tomography.

We have enhanced photoacoustic computed tomography with dry acoustic coupling that eliminates water immersion anxiety and wrinkling of the animal and facilitates incorporating complementary modalities and procedures. The dry acoustic coupler is made of a tubular elastic membrane enclosed by a closed transparent water tank. The tubular membrane ensures water-free contact with the animal, and the closed water tank allows pressurization for animal stabilization. The dry coupler was tested using a whole-body small-animal ring-shaped photoacoustic computed tomography system. Dry coupling was found to provide image quality comparable to that of conventional water coupling.

Photoacoustic elastography.

Elastography can noninvasively map the elasticity distribution in biological tissue, which can potentially be used to reveal disease conditions. In this Letter, we have demonstrated photoacoustic elastography by using a linear-array photoacoustic computed tomography system. The feasibility of photoacoustic elastography was first demonstrated by imaging the strains of single-layer and bilayer gelatin phantoms with various stiffness values. The measured strains agreed well with theoretical values, with an average error of less than 5.2%. Next, in vivo photoacoustic elastography was demonstrated on a mouse leg, where the fat and muscle distribution was mapped based on the elasticity contrast. We confirmed the photoacoustic elastography results by ultrasound elastography performed simultaneously.

Encrypted Three-dimensional Dynamic Imaging using Snapshot Time-of-flight Compressed Ultrafast Photography.

Compressed ultrafast photography (CUP), a computational imaging technique, is synchronized with short-pulsed laser illumination to enable dynamic three-dimensional (3D) imaging. By leveraging the time-of-flight (ToF) information of pulsed light backscattered by the object, ToF-CUP can reconstruct a volumetric image from a single camera snapshot. In addition, the approach unites the encryption of depth data with the compressed acquisition of 3D data in a single snapshot measurement, thereby allowing efficient and secure data storage and transmission. We demonstrated high-speed 3D videography of moving objects at up to 75 volumes per second. The ToF-CUP camera was applied to track the 3D position of a live comet goldfish. We have also imaged a moving object obscured by a scattering medium.

Optical-resolution photoacoustic endomicroscopy in vivo.

Optical-resolution photoacoustic microscopy (OR-PAM) has become a major experimental tool of photoacoustic tomography, with unique imaging capabilities for various biological applications. However, conventional imaging systems are all table-top embodiments, which preclude their use in internal organs. In this study, by applying the OR-PAM concept to our recently developed endoscopic technique, called photoacoustic endoscopy (PAE), we created an optical-resolution photoacoustic endomicroscopy (OR-PAEM) system, which enables internal organ imaging with a much finer resolution than conventional acoustic-resolution PAE systems. OR-PAEM has potential preclinical and clinical applications using either endogenous or exogenous contrast agents.

Handheld photoacoustic probe to detect both melanoma depth and volume at high speed in vivo.

We applied a linear-array-based photoacoustic probe to detect melanin-containing melanoma tumor depth and volume in nude mice in vivo. This system can image melanomas at five frames per second (fps), which is much faster than our previous handheld single transducer system (0.1 fps). We first theoretically show that, in addition to the higher frame rate, almost the entire boundary of the melanoma can be detected by the linear-array-based probe, while only the horizontal boundary could be detected by the previous system. Then we demonstrate the ability of this linear-array-based system in measuring both the depth and volume of melanoma through phantom, ex vivo, and in vivo experiments. The volume detection ability also enables us to accurately calculate the rate of growth of the tumor, which is an important parameter in quantifying the tumor activity. Our results show that this system can be used for clinical melanoma diagnosis and treatment in humans at the bedside. Linear-array-based PA images of melanoma acquired in vivo on day 3 (a) and day 6 (b).

Optical drug monitoring: photoacoustic imaging of nanosensors to monitor therapeutic lithium in vivo.

Personalized medicine could revolutionize how primary care physicians treat chronic disease and how researchers study fundamental biological questions. To realize this goal, we need to develop more robust, modular tools and imaging approaches for in vivo monitoring of analytes. In this report, we demonstrate that synthetic nanosensors can measure physiologic parameters with photoacoustic contrast, and we apply that platform to continuously track lithium levels in vivo. Photoacoustic imaging achieves imaging depths that are unattainable with fluorescence or multiphoton microscopy. We validated the photoacoustic results that illustrate the superior imaging depth and quality of photoacoustic imaging with optical measurements. This powerful combination of techniques will unlock the ability to measure analyte changes in deep tissue and will open up photoacoustic imaging as a diagnostic tool for continuous physiological tracking of a wide range of analytes.

Photoacoustic tomography of vascular compliance in humans.

Characterization of blood vessel elastic properties can help in detecting thrombosis and preventing life-threatening conditions such as acute myocardial infarction or stroke. Vascular elastic photoacoustic tomography (VE-PAT) is proposed to measure blood vessel compliance in humans. Implemented on a linear-array-based photoacoustic computed tomography system, VE-PAT can quantify blood vessel compliance changes due to simulated thrombosis and occlusion. The feasibility of the VE-PAT system was first demonstrated by measuring the strains under uniaxial loading in perfused blood vessel phantoms and quantifying their compliance changes due to the simulated thrombosis. The VE-PAT system detected a decrease in the compliances of blood vessel phantoms with simulated thrombosis, which was validated by a standard compression test. The VE-PAT system was then applied to assess blood vessel compliance in a human subject. Experimental results showed a decrease in compliance when an occlusion occurred downstream from the measurement point in the blood vessels, demonstrating VE-PAT's potential for clinical thrombosis detection.

Single-shot compressed ultrafast photography at one hundred billion frames per second.

The capture of transient scenes at high imaging speed has been long sought by photographers, with early examples being the well known recording in 1878 of a horse in motion and the 1887 photograph of a supersonic bullet. However, not until the late twentieth century were breakthroughs achieved in demonstrating ultrahigh-speed imaging (more than 10(5) frames per second). In particular, the introduction of electronic imaging sensors based on the charge-coupled device (CCD) or complementary metal-oxide-semiconductor (CMOS) technology revolutionized high-speed photography, enabling acquisition rates of up to 10(7) frames per second. Despite these sensors' widespread impact, further increasing frame rates using CCD or CMOS technology is fundamentally limited by their on-chip storage and electronic readout speed. Here we demonstrate a two-dimensional dynamic imaging technique, compressed ultrafast photography (CUP), which can capture non-repetitive time-evolving events at up to 10(11) frames per second. Compared with existing ultrafast imaging techniques, CUP has the prominent advantage of measuring an x-y-t (x, y, spatial coordinates; t, time) scene with a single camera snapshot, thereby allowing observation of transient events with temporal resolution as tens of picoseconds. Furthermore, akin to traditional photography, CUP is receive-only, and so does not need the specialized active illumination required by other single-shot ultrafast imagers. As a result, CUP can image a variety of luminescent--such as fluorescent or bioluminescent--objects. Using CUP, we visualize four fundamental physical phenomena with single laser shots only: laser pulse reflection and refraction, photon racing in two media, and faster-than-light propagation of non-information (that is, motion that appears faster than the speed of light but cannot convey information). Given CUP's capability, we expect it to find widespread applications in both fundamental and applied sciences, including biomedical research.

Label-free photoacoustic nanoscopy.

Super-resolution microscopy techniques - capable of overcoming the diffraction limit of light - have opened new opportunities to explore subcellular structures and dynamics not resolvable in conventional far-field microscopy. However, relying on staining with exogenous fluorescent markers, these techniques can sometimes introduce undesired artifacts to the image, mainly due to large tagging agent sizes and insufficient or variable labeling densities. By contrast, the use of endogenous pigments allows imaging of the intrinsic structures of biological samples with unaltered molecular constituents. Here, we report label-free photoacoustic (PA) nanoscopy, which is exquisitely sensitive to optical absorption, with an 88 nm resolution. At each scanning position, multiple PA signals are successively excited with increasing laser pulse energy. Because of optical saturation or nonlinear thermal expansion, the PA amplitude depends on the nonlinear incident optical fluence. The high-order dependence, quantified by polynomial fitting, provides super-resolution imaging with optical sectioning. PA nanoscopy is capable of super-resolution imaging of either fluorescent or nonfluorescent molecules.

Reversibly switchable fluorescence microscopy with enhanced resolution and image contrast.

Confocal microscopy with optical sectioning has revolutionized biological studies by providing sharper images than conventional optical microscopy. Here, we introduce a fluorescence imaging method with enhanced resolution and imaging contrast, which can be implemented using a commercial confocal microscope setup. This approach, called the reversibly switchable photo-imprint microscopy (rsPIM), is based on the switching dynamics of reversibly switchable fluorophores. When the fluorophores are switched from the bright (ON) state to the dark (OFF) state, their switching rate carries the information about the local excitation light intensity. In rsPIM, a polynomial function is used to fit the fluorescence signal decay during the transition. The extracted high-order coefficient highlights the signal contribution from the center of the excitation volume, and thus sharpens the resolution in all dimensions. In particular, out-of-focus signals are greatly blocked for large targets, and thus the image contrast is considerably enhanced. Notably, since the fluorophores can be cycled between the ON and OFF states, the whole imaging process can be repeated. RsPIM imaging with enhanced image contrast was demonstrated in both fixed and live cells using a reversibly switchable synthetic dye and a genetically encoded red fluorescent protein. Since rsPIM does not require the modification of commercial microscope systems, it may provide a simple and cost-effective solution for subdiffraction imaging of live cells.