Spatiotemporal Deep Learning-Based Cine Loop Quality Filter for Handheld Point-of-Care Echocardiography.

The reliability of automated image interpretation of point-of-care (POC) echocardiography scans depends on the quality of the acquired ultrasound data. This work reports on the development and validation of spatiotemporal deep learning models to assess the suitability of input ultrasound cine loops collected using a handheld echocardiography device for processing by an automated quantification algorithm (e.g. ejection fraction estimation). POC echocardiograms (n=885 DICOM cine loops from 175 patients) from two sites were collected using a handheld ultrasound device and annotated for image quality at the frame-level. Attributes of high-quality frames for left ventricular (LV) quantification included a temporally-stable LV, reasonable coverage of LV borders, and good contrast between the borders and chamber. Attributes of low-quality frames included temporal instability of the LV and/or imaging artifacts (e.g., lack of contrast, haze, reverberation, acoustic shadowing). Three different neural network architectures were investigated - (a) frame-level convolutional neural network (CNN) which operates on individual echo frames (VectorCNN), (b) single-stream sequence-level CNN which operates on a sequence of echo frames (VectorCNN+LSTM) and (c) two-stream sequence-level CNNs which operate on a sequence of echo and optical flow frames (VectorCNN+LSTM+Average, VectorCNN+LSTM+MinMax, and VectorCNN+LSTM+ConvPool). Evaluation on a sequestered test dataset containing 76 DICOM cine loops with 16,914 frames showed that VectorCNN+LSTM can effectively utilize both spatial and temporal information to regress the quality of an input frame (accuracy: 0.925, sensitivity = 0.860, specificity = 0.952), compared to the frame-level VectorCNN that only utilizes spatial information in that frame (accuracy: 0.903, sensitivity = 0.791, specificity = 0.949). Furthermore, an independent sample t-test indicated that the cine loops classified to be of adequate quality by the VectorCNN+LSTM model had a statistically significant lower bias in the automatically estimated EF (mean bias = - 3.73 ± 7.46 %, versus a clinically obtained reference EF) compared to the loops classified as inadequate (mean bias = -15.92 ± 12.17 %) (p = 0.007). Thus, cine loop stratification using the proposed spatiotemporal CNN model improves the reliability of automated point-of-care echocardiography image interpretation.

Comparison of Variations Between Spectral Doppler and Gaussian Surface Integration Methods for Umbilical Vein Blood Volume Flow.

OBJECTIVES: We are studying a new method for estimating blood volume flow that uses 3-dimensional ultrasound to measure the total integrated flux through an ultrasound-generated Gaussian surface that intersects the umbilical cord. This method makes none of the assumptions typically required with standard 1-dimensional spectral Doppler volume flow estimates. We compared the variations in volume flow estimates between techniques in the umbilical vein. METHODS: The study was Institutional Review Board approved, and all 12 patients gave informed consent. Because we had no reference standard for the true umbilical vein volume flow, we compared the variations of the measurements for the flow measurement techniques. At least 3 separate spectral Doppler and 3 separate Gaussian surface measurements were made along the umbilical vein. Means, standard deviations, and coefficients of variation (standard deviation/mean) for the flow estimation techniques were calculated for each patient. P < .05 was considered significant. RESULTS: The ranges of the mean volume flow estimates were 174 to 577 mL/min for the spectral Doppler method and 100 to 341 mL/min for the Gaussian surface integration (GSI) method. The mean standard deviations (mean ± SD) were 161 ± 95 and 45 ± 48 mL/min for the spectral Doppler and GSI methods, respectively (P < .003). The mean coefficients of variation were 0.46 ± 0.17 and 0.18 ± 0.14 for the spectral Doppler and GSI methods respectively (P < 0.002). CONCLUSIONS: The new volume flow estimation method using 3-dimensional ultrasound appears to have significantly less variation in estimates than the standard 1-dimensional spectral Doppler method.

A method for localized computation of Pulse Wave Velocity in carotid structure.

Pulse Wave Velocity (PWV) promises to be a useful clinical marker for noninvasive diagnosis of atherosclerosis. This work demonstrates the ability to perform localized carotid PWV measurements from the distention waveform derived from the Radio Frequency (RF) ultrasound signal using a carotid phantom setup. The proposed system consists of low cost custom-built ultrasound probe and algorithms for envelope detection, arterial wall identification, echo tracking, distension waveform computation and PWV estimation. The method is proposed on a phantom data acquired using custom-built prototype non-imaging probe. The proposed approach is non-image based and can be seamlessly integrated into existing clinical ultrasound scanners.

Gaussian mixture model based identification of arterial wall movement for computation of distension waveform.

This work proposes a novel Gaussian Mixture Model (GMM) based approach for accurate tracking of the arterial wall and subsequent computation of the distension waveform using Radio Frequency (RF) ultrasound signal. The approach was evaluated on ultrasound RF data acquired using a prototype ultrasound system from an artery mimicking flow phantom. The effectiveness of the proposed algorithm is demonstrated by comparing with existing wall tracking algorithms. The experimental results show that the proposed method provides 20% reduction in the error margin compared to the existing approaches in tracking the arterial wall movement. This approach coupled with ultrasound system can be used to estimate the arterial compliance parameters required for screening of cardiovascular related disorders.

Enhanced gene expression of systemically administered plasmid DNA in the liver with therapeutic ultrasound and microbubbles.

Ultrasound-mediated delivery (USMD) of novel therapeutic agents in the presence of microbubbles is a potentially safe and effective method for gene therapy offering many desired characteristics, such as low toxicity, potential for repeated treatment, and organ specificity. In this study, we tested the capability of USMD to improve gene expression in mice livers using glycogen storage disease Type Ia as a model disease under systemic administration of naked plasmid DNA. Image-guided therapeutic ultrasound was used in two studies to provide therapeutic ultrasound to mice livers. In the first study, involving wild-type mice, control animals received naked plasmid DNA (pG6Pase 150 µg) via the tail vein, followed by an infusion of microbubbles; the treated animals additionally received therapeutic ultrasound (1 MHz). Following the procedure, the animals were left to recover and were subsequently euthanized after 2 d and liver samples were extracted. Reverse transcription polymerase chain reaction (RT-PCR) assays were performed on the samples to quantify mRNA expression. In addition, Western blot assays of FLAG-tagged glucose-6-phosphatase (G6Pase) were performed to evaluate protein expression. Ultrasound-exposed animals showed a 4-fold increase in G6Pase RNA in the liver, in comparison with control animals. Furthermore, results from Western blot analysis demonstrated a 2-fold increased protein expression in ultrasound-exposed animals after two days ( p < 0.05). A second pilot study was performed with G6Pase knockout mice, and the animals were monitored for correction of hypoglycemia over a period of 3 weeks before tissue analysis. The RT-PCR assays of samples from these animals demonstrated increased G6Pase RNA in the liver following ultrasound treatment. These results demonstrate that USMD can increase gene expression of systemically injected naked pDNA in the liver and also provide insight into the development of realistic approaches that can be translated into clinical practice.

Plasmonic intravascular photoacoustic imaging for detection of macrophages in atherosclerotic plaques.

To detect macrophages in atherosclerotic plaques, plasmonic gold nanoparticles are introduced as a contrast agent for intravascular photoacoustic imaging. The phantom and ex vivo tissue studies show that the individual spherical nanoparticles, resonant at 530 nm wavelength, produce a weak photoacoustic signal at 680 nm wavelength while photoacoustic signal from nanoparticles internalized by macrophages is very strong due to the plasmon resonance coupling effect. These results suggest that intravascular photoacoustic imaging can assess the macrophage-mediated aggregation of nanoparticles and therefore identify the presence and the location of nanoparticles associated with macrophage-rich atherosclerotic plaques.

Spectroscopic intravascular photoacoustic imaging to differentiate atherosclerotic plaques.

The potential of intravascular photoacoustic (IVPA) imaging to detect atherosclerosis was previously demonstrated using a 532 nm nanosecond pulsed laser and an intravascular ultrasound (IVUS) imaging catheter. However, to differentiate vulnerable plaques, the composition of plaques needs to be imaged. Therefore, we introduce a multi-wavelength photoacoustic imaging method to distinguish various types of plaques. Multi-spectral IVPA imaging of ex vivo samples of normal and atherosclerotic rabbit aorta was performed at several wavelengths within 680-900 nm range. The spectral variation of photoacoustic response was extracted and a spectroscopic analysis was performed. The results of our preliminary study suggest that the spectroscopic intravascular photoacoustic imaging technique can be used to differentiate fibrous and lipid components of the atherosclerotic plaques.

Remote temperature estimation in intravascular photoacoustic imaging.

Intravascular photoacoustic (IVPA) imaging is based on the detection of laser-induced acoustic waves generated within the arterial tissue under pulsed laser irradiation. In general, laser radiant energy levels are kept low (20 mJ/cm(2)) during photoacoustic imaging to conform to general standards for safe use of lasers on biological tissues. However, safety standards in intravascular photoacoustic imaging are not yet fully established. Consequently, monitoring spatio-temporal temperature changes associated with laser-tissue interaction is important to address thermal safety of IVPA imaging. In this study we utilize the IVUS-based strain measurements to estimate the laser-induced temperature increase. Temporal changes in temperature were estimated in a phantom modeling a vessel with an inclusion. A cross-correlation-based time delay estimator was used to assess temperature-induced strains produced by different laser radiant energies. The IVUS-based remote measurements revealed temperature increases of 0.7+/-0.3 degrees C, 2.9+/-0.2 degrees C and 5.0+/-0.2 degrees C, for the laser radiant energies of 30 mJ/cm(2), 60 mJ/cm(2) and 85 mJ/cm(2), respectively. The technique was then used in imaging of ex vivo samples of a normal rabbit aorta. For arterial tissues, a temperature elevation of 1.1 degrees C was observed for a laser fluence of 60 mJ/cm(2) and lesser than 1 degrees C for lower energy levels normally associated with IVPA imaging. Therefore, the developed ultrasound technique can be used to monitor temperature during IVPA imaging. Furthermore, the analysis based on the Arrhenius thermal damage model indicates no thermal injury in the arterial tissue, suggesting the safety of IVPA imaging.

Intravascular ultrasound and photoacoustic imaging.

There is a need for an imaging technique that can reliably identify and characterize the vulnerability of atherosclerotic plaques. Catheter-based intravascular ultrasound (IVUS) is one of the imaging tools of the clinical evaluation of atherosclerosis. However, histopathological information obtained with IVUS imaging is limited. We present and discuss the applicability of a combined intravascular photoacoustic (IVPA) and intravascular ultrasound (IVUS) imaging approach to assess both vessel structure and tissue composition thus identifying rupture-prone atherosclerotic plaques. Photoacoustic (or optoacoustic and, generally, thermoacoustic) imaging relies on the absorption of electromagnetic energy, such as light, and the subsequent emission of an acoustic wave. Therefore, the amplitude and temporal characteristics of the photoacoustic signal is primarily determined by optical absorption properties of different types of tissues and can be used to differentiate the lipid, fibrous and fibro-cellular components of an inflammatory lesion. Simultaneous IVUS and IVPA imaging studies were conducted using 40 MHz clinical IVUS imaging catheter interfaced with a pulsed laser system. The performance of the IVPA/IVUS imaging was assessed using phantoms with point targets and vessel-mimicking phantoms. To detect the lipids in the plaque, ex-vivo IVPA imaging studies of a normal and an atherosclerotic rabbit aorta were performed at a 532 nm wavelength. To assess plaque composition, multi-wavelength (680-950 nm) spectroscopic IVPA imaging studies were carried out. Finally, molecular and cellular IVPA imaging was demonstrated using plasmonic nanoparticles. Overall, our studies suggest that plaque detection and characterization can be improved using the combined IVPA/IVUS imaging.

Intravascular photoacoustic imaging using an IVUS imaging catheter.

Catheter-based imaging of atherosclerosis with high resolution, albeit invasive, is extremely important for screening and characterization of vulnerable plaques. Currently, there is a need for an imaging technique capable of providing comprehensive morphological and functional information of plaques. In this paper, we present an intravascular photoacoustic imaging technique to characterize vulnerable plaques by using optical absorption contrast between normal tissue and atherosclerotic lesions. Specifically, we investigate the feasibility of obtaining intravascular photoacoustic (IVPA) images using a high-frequency intravascular ultrasound (IVUS) imaging catheter. Indeed, the combination of IVPA imaging with clinically available IVUS imaging may provide desired functional and morphological assessment of the plaque. The imaging studies were performed with tissue-mimicking arterial vessel phantoms and excised samples of rabbit artery. The results of our study suggest that catheter-based intravascular photoacoustic imaging is possible, and the combination of IVPA with IVUS has the potential to detect and differentiate atherosclerosis based on both the structure and composition of the plaque.

Ex vivo Characterization of Atherosclerosis using Intravascular Photoacoustic Imaging.

The imaging of plaque composition represents one of the important steps in the interventional management of atherosclerosis. Intravascular photoacoustic (IVPA) imaging has the potential to play a major role in the detection and differentiation of an atherosclerotic lesion. The difference in the optical properties of the arterial wall and plaque constituents could be utilized to obtain high resolution photoacoustic images. In this work, through ex vivo imaging studies using a rabbit model of atherosclerosis, we evaluate the ability of IVPA imaging to detect and characterize the plaque. Specifically, the difference in the magnitude of the photoacoustic signals from the free lipids, macrophage foam cells, blood and the rest of the arterial wall were helpful in providing the contrast and detecting the fibro-cellular inflammatory plaque. The constituents identified in the IVPA images were confirmed by the results from histology.

Differential phase optical coherence probe for depth-resolved detection of photothermal response in tissue.

We describe a differential phase low-coherence interferometric probe for non-invasive, quantitative imaging of photothermal phenomena in biological materials. Our detection method utilizes principles of optical coherence tomography with differential phase measurement of interference fringe signals. A dual-channel optical low-coherence probe is used to analyse laser-induced thermoelastic and thermorefractive effects in tissue with micrometre axial resolution and nanometre sensitivity. We demonstrate an application of the technique using tissue phantoms and ex-vivo tissue specimens of rodent dorsal skin.