Classification of blood cells and tumor cells using label-free ultrasound and photoacoustics

Diagnosis of diabetes mellitus using high frequency ultrasound and convolutional neural network

The incidence of diabetes mellitus has been increasing, prompting the search for non-invasive diagnostic methods. Although current methods exist, these have certain limitations, such as low reliability and accuracy, difficulty in individual patient adjustment, and discomfort during use. This paper presents a novel approach for diagnosing diabetes using high-frequency ultrasound (HFU) and a convolutional neural network (CNN). This method is based on the observation that glucose in red blood cells (RBCs) forms glycated hemoglobin (HbA1c) and accumulates on its surface. The study incubated RBCs with different glucose concentrations, collected acoustic reflection signals from them using a custom-designed 90-MHz transducer, and analyzed the signals using a CNN. The CNN was applied to the frequency spectra and spectrograms of the signal to identify correlations between changes in RBC properties owing to glucose concentration and signal features. The results confirmed the efficacy of the CNN-based approach with a classification accuracy of 0.98. This non-invasive diagnostic technology using HFU and CNN holds promise for in vivo diagnosis without the need for blood collection.

Dual-wavelength photoacoustic atlas method to estimate fractional methylene blue and hemoglobin contents

SIGNIFICANCE

Methylene blue (MB) is an exogenous contrast agent that has the potential to assist with visualization and penetration challenges in photoacoustic imaging. However, monitoring the local concentration between MB and endogenous chromophores is critical for avoiding unnecessary MB accumulations that could lead to adverse effects such as hemolysis when exposed to increased dose and photodamage when exposed to high laser energies.

AIM

We developed a modified version of a previously proposed acoustic-based atlas method to estimate concentration levels from a mixture of two photoacoustic-sensitive materials after two laser wavelength emissions.

APPROACH

Photoacoustic data were acquired from mixtures of 100-μM MB and either human or porcine blood (Hb) injected in a plastisol phantom, using laser wavelengths of 710 and 870 nm. An algorithm to perform linear regression of the acoustic frequency response from an atlas composed of pure concentrations was designed to assess the concentration levels from photoacoustic samples obtained from 11 known MB/Hb volume mixtures. The mean absolute error (MAE), coefficient of determination (i.e., R2), and Spearman's correlation coefficient (i.e., ρ) between the estimated results and ground-truth labels were calculated to assess the algorithm performance, linearity, and monotonicity, respectively.

RESULTS

The overall MAE, R2, and ρ were 12.68%, 0.80, and 0.89, respectively, for the human Hb dataset and 9.92%, 0.86, and 0.93, respectively, for the porcine Hb dataset. In addition, a similarly linear relationship was observed between the acoustic frequency response at 2.3 MHz and 870-nm laser wavelength and the ground-truth concentrations, with R2 and | ρ | values of 0.76 and 0.88, respectively.

CONCLUSIONS

Contrast agent concentration monitoring is feasible with the proposed approach. The potential for minimal data acquisition times with only two wavelength emissions is advantageous toward real-time implementation in the operating room.

Time-domain ultrasound as prior information for frequency-domain compressive ultrasound for intravascular cell detection: A 2-cell numerical model

This study proposes a new method for the detection of a weak scatterer among strong scatterers using prior-information ultrasound (US) imaging. A perfect application of this approach is in vivo cell detection in the bloodstream, where red blood cells (RBCs) serve as identifiable strong scatterers. In vivo cell detection can help diagnose cancer at its earliest stages, increasing the chances of survival for patients. This work combines time-domain US with frequency-domain compressive US imaging to detect a 20-μ MCF-7 circulating tumor cell (CTC) among a number of RBCs within a simulated venule inside the mouth. The 2D image reconstructed from the time-domain US is employed to simulate the reflected and scattered pressure field from the RBCs, which is then measured at the location of the receivers. The RBCs are tagged one time by a human operator and another time, automatically, by template-based computer vision. Next, the resulting signal from the RBCs is subtracted from the measured total signal in frequency domain to generate the scattered-field data, coming from the CTC alone. Feeding that signal and the background pressure field into a norm-one-based compressive sensing code enables detecting the CTC at various locations. As errors could arise in determining the location of the RBCs and their acoustic properties in the real world, small errors (up to 10% in the former and 5% in the latter) are purposefully introduced to the model, to which the proposed method is shown to be resilient. Localization errors are smaller than 12 μ when a human tags the RBCs and smaller than 25 μ when computer vision is applied. Despite its limitations, this study, for the first time, reports the results of combining two US modalities aimed at cell detection and introduces a unique and useful application for ultrahigh-frequency US imaging. It should be noted that this method can be used in detecting weak scatterers with ultrasound waves in other applications as well.

Broadband surface plasmon resonance sensor for fast spectroscopic photoacoustic microscopy

High-speed optical-resolution photoacoustic microscopy (OR-PAM), integrating the merits of high spatial resolution and fast imaging acquisition, can observe dynamic processes of the optical absorption-based molecular specificities. However, it remains challenging for the evaluation to morphological and physiological parameters that are closely associated with photoacoustic spectrum due to the inadequate ultrasonic frequency response of the routinely-employed piezoelectric transducer. By utilizing the galvanometer for fast optical scanning and our previously-developed surface plasmon resonance sensor as an unfocused broadband ultrasonic detector, high-speed spectroscopic photoacoustic imaging was accessed in the OR-PAM system, achieving an acoustic bandwidth of ∼125 MHz and B-scan rate at ∼200 Hz over a scanning range of ∼0.5 mm. Our system demonstrated the dynamic imaging of the moving phantoms' structures and the simultaneous characterization of their photoacoustic spectra over time. Further, fast volumetric imaging and spectroscopic analysis of microanatomic features of a zebrafish eye ex vivo was obtained label-freely.

Recent Technical Progression in Photoacoustic Imaging—Towards Using Contrast Agents and Multimodal Techniques

For combining optical and ultrasonic imaging methodologies, photoacoustic imaging (PAI) is the most important and successful hybrid technique, which has greatly contributed to biomedical research and applications. Its theoretical background is based on the photoacoustic effect, whereby a modulated or pulsed light is emitted into tissue, which selectively absorbs the optical energy of the light at optical wavelengths. This energy produces a fast thermal expansion in the illuminated tissue, generating pressure waves (or photoacoustic waves) that can be detected by ultrasonic transducers. Research has shown that optical absorption spectroscopy offers high optical sensitivity and contrast for ingredient determination, for example, while ultrasound has demonstrated good spatial resolution in biomedical imaging. Photoacoustic imaging combines these advantages, i.e., high contrast through optical absorption and high spatial resolution due to the low scattering of ultrasound in tissue. In this review, we focus on advances made in PAI in the last five years and present categories and key devices used in PAI techniques. In particular, we highlight the continuously increasing imaging depth achieved by PAI, particularly when using exogenous reagents. Finally, we discuss the potential of combining PAI with other imaging techniques.

Real-time red blood cell counting and osmolarity analysis using a photoacoustic-based microfluidic system

Counting the number of red blood cells (RBCs) in blood samples is a common clinical diagnostic procedure, but conventional methods are unable to provide the size and other physical properties of RBCs at the same time. In this work, we explore photoacoustic (PA) detection as a rapid label-free and noninvasive analysis technique that can potentially be used for single RBC characterization based on their photoabsorption properties. We have demonstrated an on-chip PA flow cytometry system using a simple microfluidic chip combined with a PA imaging system to count and characterize up to ∼60 RBCs per second. Compared with existing microfluidic-based RBC analysis methods, which typically use camera-captured image sequences to characterize cell morphology and deformation, the PA method discussed here requires only the processing of one-dimensional time-series data instead of two- or three-dimensional time-series data acquired by computer vision methods. Therefore, the PA method will have significantly lower computational requirements when large numbers of RBCs are to be analyzed. Moreover, we have demonstrated that the PA signals of RBCs flowing in a microfluidic device could be directly used to acquire the osmolarity conditions (in the range of 124 to 497 mOsm L-1) of the medium surrounding the RBCs. This finding suggests a potential extension of applicability to blood tests via PA-based biomedical detection.

Deep tissue volumetric optoacoustic tracking of individual circulating tumor cells in an intracardially perfused mouse model

Widespread metastasis is the major cause of death from melanoma and other types of cancer. At present, the dynamic aspects of the metastatic cascade remain enigmatic. The feasibility to track circulating melanoma cells deep within living intact organisms can greatly impact our knowledge on tumor metastasis, but existing imaging approaches lack the sensitivity, spatio-temporal resolution or penetration depth to capture flowing tumor cells over large fields of view within optically-opaque biological tissues. Vast progress with the development of optoacoustic tomography technologies has recently enabled two- and three-dimensional imaging at unprecedented frame rates in the order of hundreds of Hertz, effectively mapping up to a million image voxels within a single volumetric snapshot. Herein, we employ volumetric optoacoustic tomography for real-time visualization of passage and trapping of individual B16 melanoma cells in the whole mouse brain. Detection of individual circulating melanoma cells was facilitated by substituting blood with an artificial cerebrospinal fluid that removes the strong absorption background in the optoacoustic images. The approach can provide new opportunities for studying trafficking and accumulation of metastatic melanoma cells in different organs.

Sizing biological cells using a microfluidic acoustic flow cytometer

We describe a new technique that combines ultrasound and microfluidics to rapidly size and count cells in a high-throughput and label-free fashion. Using 3D hydrodynamic flow focusing, cells are streamed single file through an ultrasound beam where ultrasound scattering events from each individual cell are acquired. The ultrasound operates at a center frequency of 375 MHz with a wavelength of 4 μm; when the ultrasound wavelength is similar to the size of a scatterer, the power spectra of the backscattered ultrasound waves have distinct features at specific frequencies that are directly related to the cell size. Our approach determines cell sizes through a comparison of these distinct spectral features with established theoretical models. We perform an analysis of two types of cells: acute myeloid leukemia cells, where 2,390 measurements resulted in a mean size of 10.0 ± 1.7 μm, and HT29 colorectal cancer cells, where 1,955 measurements resulted in a mean size of 15.0 ± 2.3 μm. These results and histogram distributions agree very well with those measured from a Coulter Counter Multisizer 4. Our technique is the first to combine ultrasound and microfluidics to determine the cell size with the potential for multi-parameter cellular characterization using fluorescence, light scattering and quantitative photoacoustic techniques.

Simultaneous acoustic and photoacoustic microfluidic flow cytometry for label-free analysis

We developed a label-free microfluidic acoustic flow cytometer (AFC) based on interleaved detection of ultrasound backscatter and photoacoustic waves from individual cells and particles flowing through a microfluidic channel. The AFC uses ultra-high frequency ultrasound, which has a center frequency of 375 MHz, corresponding to a wavelength of 4 μm, and a nanosecondpulsed laser, to detect individual cells. We validate the AFC by using it to count different color polystyrene microparticles and comparing the results to data from fluorescence-activated cell sorting (FACS). We also identify and count red and white blood cells in a blood sample using the AFC, and observe an excellent agreement with results obtained from FACS. This new label-free, non-destructive technique enables rapid and multi-parametric studies of individual cells of a large heterogeneous population using parameters such as ultrasound backscatter, optical absorption, and physical properties, for cell counting and sizing in biomedical and diagnostics applications.

New advances in microfluidic flow cytometry

In recent years, researchers are paying the increasing attention to the development of portable microfluidic diagnostic devices including microfluidic flow cytometry for the point-of-care testing. Microfluidic flow cytometry, where microfluidics and flow cytometry work together to realize novel functionalities on the microchip, provides a powerful tool for measuring the multiple characteristics of biological samples. The development of a portable, low-cost, and compact flow cytometer can benefit the health care in underserved areas such as Africa or Asia. In this article, we review recent advancements of microfluidics including sample pumping, focusing and sorting, novel detection approaches, and data analysis in the field of flow cytometry. The challenge of microfluidic flow cytometry is also examined briefly.

Extended photoacoustic transport model for characterization of red blood cell morphology in microchannel flow

The dynamic response behavior of red blood cells holds the key to understanding red blood cell related diseases. In this regard, an understanding of the physiological functions of erythrocytes is significant before focusing on red blood cell aggregation in the microcirculatory system. In this work, we present a theoretical model for a photoacoustic signal that occurs when deformed red blood cells pass through a microfluidic channel. Using a Green's function approach, the photoacoustic pressure wave is obtained analytically by solving a combined Navier-Stokes and photoacoustic equation system. The photoacoustic wave expression includes determinant parameters for the cell deformability such as plasma viscosity, density, and red blood cell aggregation, as well as involving laser parameters such as beamwidth, pulse duration, and repetition rate. The effects of aggregation on blood rheology are also investigated. The results presented by this study show good agreements with the experimental ones in the literature. The comprehensive analytical solution of the extended photoacoustic transport model including a modified Morse type potential function sheds light on the dynamics of aggregate formation and demonstrates that the profile of a photoacoustic pressure wave has the potential for detecting and characterizing red blood cell aggregation.

High frequency ultrasound imaging and simulations of sea urchin oocytes

High frequency ultrasound backscatter signals from sea urchin oocytes were measured using a 40 MHz transducer and compared to numerical simulations. The Faran scattering model was used to calculate the ultrasound scattered from single oocytes in suspension. The urchin oocytes are non-nucleated with uniform size and biomechanical properties; the backscatter from each cell is similar and easy to simulate, unlike typical nucleated mammalian cells. The time domain signal measured from single oocytes in suspension showed two distinct peaks, and the power spectrum was periodic with minima spaced approximately 10 MHz apart. Good agreement to the Faran scattering model was observed. Measurements from tightly packed oocyte cell pellets showed similar periodic features in the power spectra, which was a result of the uniform size and consistent biomechanical properties of the cells. Numerical simulations that calculated the ultrasound scattered from individual oocytes within a three dimensional volume showed good agreement to the measured signals and B-scan images. A cepstral analysis of the signal was used to calculate the size of the cells, which was 78.7 μm (measured) and 81.4 μm (simulated). This work supports the single scattering approximation, where ultrasound is discretely scattered from single cells within a bulk homogeneous sample, and that multiple scattering has a negligible effect. This technique can be applied towards understanding the complex scattering behaviour from heterogeneous tissues.

Biophysical isolation and identification of circulating tumor cells

Isolation and enumeration of circulating tumor cells (CTCs) from blood is important for determining patient prognosis and monitoring treatment. Methods based on affinity to cell surface markers have been applied to both purify (via immunoseparation) and identify (via immunofluorescence) CTCs. However, variability of cell biomarker expression associated with tumor heterogeneity and evolution and cross-reactivity of antibody probes have long complicated CTC enrichment and immunostaining. Here, we report a truly label-free high-throughput microfluidic approach to isolate, enumerate, and characterize the biophysical properties of CTCs using an integrated microfluidic device. Vortex-mediated deformability cytometry (VDC) consists of an initial vortex region which enriches large CTCs, followed by release into a downstream hydrodynamic stretching region which deforms the cells. Visualization and quantification of cell deformation with a high-speed camera revealed populations of large (>15 μm diameter) and deformable (aspect ratio >1.2) CTCs from 16 stage IV lung cancer samples, that are clearly distinguished by increased deformability compared to contaminating blood cells and rare large cells isolated from healthy patients. The VDC technology demonstrated a comparable positive detection rate of putative CTCs above healthy baseline (93.8%) with respect to standard immunofluorescence (71.4%). Automation allows full enumeration of CTCs from a 10 mL vial of blood within <1 h after sample acquisition, compared with 4+ hours with standard approaches. Moreover, cells are released into any collection vessel for further downstream analysis. VDC shows potential for accurate CTC enumeration without labels and confirms the unique highly deformable biophysical properties of large CTCs circulating in blood.

Properties of cells through life and death - an acoustic microscopy investigation

Current methods to evaluate the status of a cell are largely focused on fluorescent identification of molecular biomarkers. The invasive nature of these methods - requiring either fixation, chemical dyes, genetic alteration, or a combination of these - prevents subsequent analysis of samples. In light of this limitation, studies have considered the use of physical markers to differentiate cell stages. Acoustic microscopy is an ultrahigh frequency (>100 MHz) ultrasound technology that can be used to calculate the mechanical and physical properties of biological cells in real-time, thereby evaluating cell stage in live cells without invasive biomarker evaluation. Using acoustic microscopy, MCF-7 human breast adenocarcinoma cells within the G1, G2, and metaphase phases of the proliferative cell cycle, in addition to early and late programmed cell death, were examined. Physical properties calculated include the cell height, sound speed, acoustic impedance, cell density, adiabatic bulk modulus, and the ultrasonic attenuation. A total of 290 cells were measured, 58 from each cell phase, assessed using fluorescent and phase contrast microscopy. Cells actively progressing from G1 to metaphase were marked by a 28% decrease in attenuation, in contrast to the induction of apoptosis from G1, which was marked by a significant 81% increase in attenuation. Furthermore late apoptotic cells separated into 2 distinct groups based on ultrasound attenuation, suggesting that presently-unidentified sub-stages may exist within late apoptosis. A methodology has been implemented for the identification of cell stages without the use of chemical dyes, fixation, or genetic manipulation.

High resolution ultrasound and photoacoustic imaging of single cells

High resolution ultrasound and photoacoustic images of stained neutrophils, lymphocytes and monocytes from a blood smear were acquired using a combined acoustic/photoacoustic microscope. Photoacoustic images were created using a pulsed 532 nm laser that was coupled to a single mode fiber to produce output wavelengths from 532 nm to 620 nm via stimulated Raman scattering. The excitation wavelength was selected using optical filters and focused onto the sample using a 20× objective. A 1000 MHz transducer was co-aligned with the laser spot and used for ultrasound and photoacoustic images, enabling micrometer resolution with both modalities. The different cell types could be easily identified due to variations in contrast within the acoustic and photoacoustic images. This technique provides a new way of probing leukocyte structure with potential applications towards detecting cellular abnormalities and diseased cells at the single cell level.