Artificial intelligence-assisted ultrasound-guided focused ultrasound therapy: a feasibility study.

OBJECTIVES: Focused ultrasound (FUS) therapy has emerged as a promising noninvasive solution for tumor ablation. Accurate monitoring and guidance of ultrasound energy is crucial for effective FUS treatment. Although ultrasound (US) imaging is a well-suited modality for FUS monitoring, US-guided FUS (USgFUS) faces challenges in achieving precise monitoring, leading to unpredictable ablation shapes and a lack of quantitative monitoring. The demand for precise FUS monitoring heightens when complete tumor ablation involves controlling multiple sonication procedures. METHODS: To address these challenges, we propose an artificial intelligence (AI)-assisted USgFUS framework, incorporating an AI segmentation model with B-mode ultrasound imaging. This method labels the ablated regions distinguished by the hyperechogenicity effect, potentially bolstering FUS guidance. We evaluated our proposed method using the Swin-Unet AI architecture, conducting experiments with a USgFUS setup on chicken breast tissue. RESULTS: Our results showed a 93% accuracy in identifying ablated areas marked by the hyperechogenicity effect in B-mode imaging. CONCLUSION: Our findings suggest that AI-assisted ultrasound monitoring can significantly improve the precision and control of FUS treatments, suggesting a crucial advancement toward the development of more effective FUS treatment strategies.

Image Generation and Recognition for Railway Surface Defect Detection.

Railway defects can result in substantial economic and human losses. Among all defects, surface defects are the most common and prominent type, and various optical-based non-destructive testing (NDT) methods have been employed to detect them. In NDT, reliable and accurate interpretation of test data is vital for effective defect detection. Among the many sources of errors, human errors are the most unpredictable and frequent. Artificial intelligence (AI) has the potential to address this challenge; however, the lack of sufficient railway images with diverse types of defects is the major obstacle to training the AI models through supervised learning. To overcome this obstacle, this research proposes the RailGAN model, which enhances the basic CycleGAN model by introducing a pre-sampling stage for railway tracks. Two pre-sampling techniques are tested for the RailGAN model: image-filtration, and U-Net. By applying both techniques to 20 real-time railway images, it is demonstrated that U-Net produces more consistent results in image segmentation across all images and is less affected by the pixel intensity values of the railway track. Comparison of the RailGAN model with U-Net and the original CycleGAN model on real-time railway images reveals that the original CycleGAN model generates defects in the irrelevant background, while the RailGAN model produces synthetic defect patterns exclusively on the railway surface. The artificial images generated by the RailGAN model closely resemble real cracks on railway tracks and are suitable for training neural-network-based defect identification algorithms. The effectiveness of the RailGAN model can be evaluated by training a defect identification algorithm with the generated dataset and applying it to real defect images. The proposed RailGAN model has the potential to improve the accuracy of NDT for railway defects, which can ultimately lead to increased safety and reduced economic losses. The method is currently performed offline, but further study is planned to achieve real-time defect detection in the future.

Magnetic nanoparticles-enhanced focused ultrasound heating: size effect, mechanism, and performance analysis.

High intensity focused ultrasound (HIFU) has attracted great interest as a new energy-based technique to treat disordered tissues, such as tumors, through a hyperthermal mechanism using ultrasonic waves. However, long treatment times and collateral damage to healthy tissues due to high acoustic powers are still challenges for the clinical application of HIFU. One possible strategy to enhance the deposition efficiency of HIFU at the tumor site is to employ magnetic nanoparticles (MNPs) as ultrasound absorption agents for the thermal therapy. The objectives of the current study are threefold: (i) to examine the effects of MNP features, including the size and volume concentration, on the thermal mechanism of HIFU (ii) to investigate the performance of MNPs as they were exposed to ultrasound fields at different ranges of power and frequency (iii) and to study the interaction mechanism between MNPs and ultrasonic waves during the MNPs-enhanced HIFU process. To this end, we developed an ultrasound-guided HIFU system to conduct an in vitro experimental study on tissue phantoms containing MNPs of different sizes and volume concentrations. A set of HIFU parameters such as temperature rise and the rate of absorbed energy was monitored to examine the role of MNPs during the NPs-enhanced HIFU thermal procedure. Results showed that the MNPs significantly improved the thermal effect of HIFU by enhancing the rate of energy converted to heat and the temperature rise at the focal region. Moreover, it was demonstrated that the increase of MNP size and volume concentration greatly enhanced the HIFU parameters; the effects of MNPs were further improved by increasing the power and frequency of acoustic field.

Compressed Sensing for Elastography in Portable Ultrasound.

Portable ultrasound is recently emerging as a new medical imaging modality featuring high portability, easy connectivity, and real-time on-site diagnostic ability. However, it does not yet provide ultrasound elastography function that enables the diagnosis of malignant lesions using elastic properties. This is mainly due to the limitations of hardware performance and wireless data transfer speed for processing the large amount of data for elastography. Therefore, data transfer reduction is one of the feasible solutions to overcome these limitations. Recently, compressive sensing (CS) theory has been rigorously studied as a means to break the conventional Nyquist sampling rate and thus can significantly decrease the amount of measurement signals without sacrificing signal quality. In this research, we implemented various CS reconstruction frameworks and comparatively evaluated their reconstruction performance for realizing ultrasound elastography function on portable ultrasound. Combinations of three most common model bases (Fourier transform [FT], discrete cosine transform [DCT], and wave atom [WA]) and two reconstruction algorithms (L1 minimization and block sparse Bayesian learning [BSBL]) were considered for CS frameworks. Echoic and elastography phantoms, were developed to evaluate the performance of CS on B-mode images and elastograms. To assess the reconstruction quality, mean absolute error (MAE), signal-to-noise ratio (SNRe), and contrast-to-noise ratio (CNRe) were measured on the B-mode images and elastograms from CS reconstructions. Results suggest that CS reconstruction adopting BSBL algorithm with DCT model basis can yield the best results for all the measures tested, and the maximum data reduction rate for producing readily discernable elastograms is around 60%.

Application of Eshelby's Solution to Elastography for Diagnosis of Breast Cancer.

Eshelby's solution is the analytical method that can derive the elastic field within and around an ellipsoidal inclusion embedded in a matrix. Since breast tumor can be regarded as an elastic inclusion with different elastic properties from those of surrounding matrix when the deformation is small, we applied Eshelby's solution to predict the stress and strain fields in the breast containing a suspicious lesion. The results were used to investigate the effectiveness of strain ratio (SR) from elastography in representing modulus ratio (MR) that may be the meaningful indicator of the malignancy of the lesion. This study showed that SR significantly underestimates MR and is varied with the shape and the modulus of the lesion. Based on the results from Eshelby's solution and finite element analysis (FEA), we proposed a surface regression model as a polynomial function that can predict the MR of the lesion to the matrix. The model has been applied to gelatin-based phantoms and clinical ultrasound images of human breasts containing different types of lesions. The results suggest the potential of the proposed method to improve the diagnostic performance of breast cancer using elastography.

Model based deep learning method for focused ultrasound pathway scanning.

The primary purpose of high-intensity focused ultrasound (HIFU), a non-invasive medical therapy, is to precisely target and ablate tumors by focusing high-frequency ultrasound from an external power source. A series of ablations must be performed in order to treat a big volume of tumors, as a single ablation can only remove a small amount of tissue. To maximize therapeutic efficacy while minimizing adverse side effects such as skin burns, preoperative treatment planning is essential in determining the focal site and sonication duration for each ablation. Here, we introduce a machine learning-based approach for designing HIFU treatment plans, which makes use of a map of the material characteristics unique to a patient alongside an accurate thermal simulation. A numerical model was employed to solve the governing equations of HIFU process and to simulate the HIFU absorption mechanism, including ensuing heat transfer process and the temperature rise during the sonication period. To validate the accuracy of this numerical model, a series of tests was conducted using ex vivo bovine liver. The findings indicate that the developed models properly represent the considerable variances observed in tumor geometrical shapes and proficiently generate well-defined closed treated regions based on imaging data. The proposed strategy facilitated the formulation of high-quality treatment plans, with an average tissue over- or under-treatment rate of less than 0.06%. The efficacy of the numerical model in accurately predicting the heating process of HIFU, when combined with machine learning techniques, was validated through quantitative comparison with experimental data. The proposed approach in cooperation with HIFU simulation holds the potential to enhance presurgical HIFU plan.

Real-time monitoring of focused ultrasound therapy using intelligence-based thermography: A feasibility study.

Focused ultrasound (FUS) therapy has been widely studied for breast cancer treatment due to its potential as a fully non-invasive method to improve cosmetic and oncologic results. However, real-time imaging and monitoring of the therapeutic ultrasound delivered to the target area remain challenges for precision breast cancer therapy. The main objective of this study is to propose and evaluate a novel intelligence-based thermography (IT) method that can monitor and control FUS treatment using thermal imaging with the fusion of artificial intelligence (AI) and advanced heat transfer modeling. In the proposed method, a thermal camera is integrated into FUS system for thermal imaging of the breast surface, and an AI model is employed for the inverse analysis of the surface thermal monitoring, thereby estimating the features of the focal region. This paper presents experimental and computational studies conducted to assess the feasibility and efficiency of IT-guided FUS (ITgFUS). Tissue phantoms, designed to mimic the properties of breast tissue, were used in the experiments to investigate detectability and the impact of temperature rise at the focal region on the tissue surface. Additionally, an AI computational analysis employing an artificial neural network (ANN) and FUS simulation was carried out to provide a quantitative estimation of the temperature rise at the focal region. This estimation was based on the observed temperature profile on the breast model's surface. The results proved that the effects of temperature rise at the focused area could be detected by the thermal images acquired with thermography. Moreover, it was demonstrated that the AI analysis of the surface temperature measurement could result in near real-time monitoring of FUS by quantitative estimation of the temporal and spatial temperature rise profiles at the focal region.

Analytical and Numerical Model of High Intensity Focused Ultrasound Enhanced With Nanoparticles.

OBJECTIVE: High intensity focused ultrasound (HIFU) is a new noninvasive therapeutics that allows local treatment of solid tumors through a hyperthermal mechanism using ultrasonic energy. One promising strategy to increase the thermal efficiency of HIFU is to employ nanoparticles (NPs) as ultrasound agents for the hyperthermia procedure. However, the interaction mechanism between NPs and ultrasonic waves has not been well understood. METHODS: In an effort to investigate the heating process of NPs-enhanced HIFU, we derived a set of HIFU equations governing the temperature variation during the thermal ablation based on the principle of conservation of energy for heat transfer mechanism. A numerical model was developed to solve the HIFU equations to simulate the absorption mechanism of HIFU in the presence of NPs, the consequent heat transfer process, and the temperature rise profile during the sonication period. The accuracy of numerical model was verified by performing a series of experiments on tissue-mimicking phantoms embedded with magnetic NPs (MNPs). RESULTS: The transport processes taking place at the boundaries between NPs and surrounding medium played the major role in the temperature rise during HIFU sonication. Besides, the effects of MNPs on rising temperature were improved by amplifying the ultrasonic power and frequency as well as by increasing the MNP concentration. CONCLUSION: A quantitative comparison with experimental results demonstrated the potential of the numerical model to accurately predict the heating mechanism of HIFU mediated by NPs. SIGNIFICANCE: The proposed method can help with simulation of HIFU when NPs are employed as ultrasound agents.

Design and analysis of an ultrasonic tactile sensor using electro-mechanical analogy.

This paper proposed a hybrid design approach of a vibro-concentrator, a vital component of an ultrasonic tactile sensor, by using electro-mechanical analogy. Lab experiments on soft materials with elastic modulus from 14 kPa to 150 kPa were conducted using the tactile sensor installed with the vibro-concentrator to verify the performance of the design. Various mechanical and electrical parameters, such as resonance frequency shift and equivalent conductance, were discussed, focusing on their feasibility as new stiffness indicators. As a variant of tactile sensors, ultrasonic tactile sensors have the advantage of high sensitivity and minimal contact with the object over traditional tactile sensors based on force-displacement principle. They detect the changes in mechanical vibration characteristics, mostly resonance frequency shift of the sensor, as an indicator of the mechanical properties of the object. A vibro-concentrator has been frequently adopted to improve the performance an ultrasonic tactile sensor, but its design has yet been systematically considered. We propose a hybrid design approach based on electro-mechanical analogy for both mechanical and electrical analyses. Mechanically, impedance analogy was adopted to design an ultrasonic vibration concentrator for the sensor to localize the contact and reinforce the vibration behavior at ~40 kHz. Electrically, we used mobility analogy to derive electrical parameters from the tactile sensing tests in lab environment. The competence of the design was demonstrated by mechanical and electrical characteristic tests. By investigating various electrical parameters from tactile sensing tests, the equivalent conductance determined by the electro-mechanical analysis was found to have almost perfectly linear relationship (R2 = 0.9998) with the samples' elastic modulus ranging from 10 kPa to 70 kPa, and showed its potential as a new stiffness indicator for soft materials. Further analyses suggested that the electrically determined series resonance frequency shift, parallel resonance frequency shift, and maximum phase angle frequency shift also had excellent linearities (R2 = 0.9947, 0.9842, and 0.9935, respectively) with sample's modulus and can be considered as indicator candidates.

Enhancing Thermal Effect of Focused Ultrasound Therapy Using Gold Nanoparticles.

High intensity focused ultrasound (HIFU) has gained increasing attention as a noninvasive therapeutic method for wide range of biomedical applications from drug delivery to cancer treatment. However, high level of ultrasonic power required for efficient HIFU treatment can cause adverse effects such as damage to surrounding healthy tissues and skin burns. One of the strategies to improve the therapeutic mechanism of HIFU is to use ultrasound absorption agents during the treatment. The objectives of current study are to investigate the feasibility of adopting gold nanoparticles (AuNPs) as ultrasound absorption agents to enhance the HIFU thermal ablation when the NPs were injected locally to the focal region; and to examine the dose effects of AuNPs on both heating and cooling mechanisms of HIFU. To this end, we conducted an experimental study on tissue-mimicking phantoms where AuNPs were injected to the focal region under the guidance of ultrasound imaging. A set of thermal parameters including temperature, specific absorption rate of acoustic energy, and cooling rate were measured to monitor the mechanism of AuNPs-mediated HIFU. The results suggest that both heating and cooling rates of HIFU procedure could be greatly improved by injecting AuNPs, which demonstrates the feasibility of using AuNPs to reduce the level of ultrasonic power from extracorporeal source for HIFU treatment.

Elastography for portable ultrasound.

Portable wireless ultrasound has been emerging as a new ultrasound device due to its unique advantages including small size, lightweight, wireless connectivity and affordability. Modern portable ultrasound devices can offer high quality sonogram images and even multiple ultrasound modes such as color Doppler, echocardiography, and endovaginal examination. However, none of them can provide elastography function yet due to the limitations in computational performance and data transfer speed of wireless communication. Also phase-based strain estimator (PSE) that is commonly used for conventional elastography cannot be adopted for portable ultrasound, because ultrasound parameters such as data dumping interval are varied significantly in the practice of portable ultrasound. Therefore, this research aims to propose a new elastography method suitable for portable ultrasound, called the robust phase-based strain estimator (RPSE), which is not only robust to the variation of ultrasound parameters but also computationally effective. Performance and suitability of RPSE were compared with other strain estimators including time-delay, displacement-gradient and phase-based strain estimators (TSE, DSE and PSE, respectively). Three types of raw RF data sets were used for validation tests: two numerical phantom data sets modeled by an open ultrasonic simulation code (Field II) and a commercial FEA (Abaqus), and the one experimentally acquired with a portable ultrasound device from a gelatin phantom. To assess image quality of elastograms, signal-to-noise (SNRe) and contrast-to-noise (CNRe) ratios were measured on the elastograms produced by each strain estimator. The computational efficiency was also estimated and compared. Results from the numerical phantom experiment showed that RPSE could achieve highest values of SNRe and CNRe (around 5.22 and 47.62 dB) among all strain estimators tested, and almost 10 times higher computational efficiency than TSE and DSE (around 0.06 vs. 5.76 s per frame for RPSE and TSE, respectively).

An experimental and numerical study on tactile neuroimaging: A novel minimally invasive technique for intraoperative brain imaging.

BACKGROUND: The success of tumour neurosurgery is highly dependent on the ability to accurately localize the operative target, which may shift during the operation. Performing intraoperative brain imaging is crucial in minimally invasive neurosurgery to detect the effect of brain shift on the tumour's location, and to maximize the efficiency of tumour resection. METHOD: The major objective of this research is to introduce tactile neuroimaging as a novel minimally invasive technique for intraoperative brain imaging. To investigate the feasibility of the proposed method, an experimental and numerical study was first performed on silicone phantoms mimicking the brain tissue with a tumour. Then the study was extended to a clinical model with the meningioma tumour. RESULTS: The stress distribution on the brain surface has high potential to intraoperatively localize the tumour. CONCLUSION: Results suggest that tactile neuroimaging can be used to provide non-invasive and real-time intraoperative data on a tumour's features.

The rheological injectability of N-succinyl-chitosan solutions.

The viscosity of a set of N-succinyl-chitosan (NSC) solutions was characterized near the 0.2Pas rheological injectability limit. This is believed to be the first such report in the open literature. Viscosity was characterized at physiological pH and ionic strength as a function of NSC degree of substitution, NSC concentration, temperature, and shear rate. NSC was synthesized via Yamaguci's method and characterized using H-NMR, membrane osmometry, TGA and isothermal vacuum drying. NSC synthesis results were shown to fit a reproducible log-linear correlation and both optimum drying temperature and thermal decomposition temperature were found to be a function of NSC degree of substitution. Viscosity results were explained using Katchalsky's full model for polyampholyte ionization combined with a charge induced excluded volume model proposed by Higgs. The model predicted a polyelectrolyte/polyampholyte transition which agreed well with experimental data. For minimally injectable formulations a maximum in primary amine concentration is expected at 32sub% amine NSC.

Human corneal epithelial cell response to substrate stiffness.

It has been reported that mechanical stimulus can affect cellular behavior. While induced differentiation in stem cells and proliferation and directional migration in fibroblasts are reported as responses to mechanical stimuli, little is known about the response of cells from the cornea. In the present study, we investigated whether changes in substrate stiffness (measured by elastic modulus) affected the behavior of human corneal epithelial cells (HCECs). Polyacrylamide substrates with different elastic moduli (compliant, medium and stiff) were prepared and HCECs were cultured on them. HCECs responses, including cell viability, apoptosis, intercellular adhesion molecule-1 (ICAM-1) expression, integrin-α3ß1 expression and changes in cytoskeleton structure (actin fibers) and migratory behavior, were studied. No statistically significant cell activation, as measured by ICAM-1 expression, was observed. However, on compliant substrates, a higher number of cells were found to be apoptotic and disrupted actin fibers were observed. Furthermore, cells displayed a statistically significant lower migration speed on compliant substrates when compared with the stiffer substrates. Thus, corneal epithelial cells respond to changes in substrate stiffness, which may have implications in the understanding and perhaps treatment of corneal diseases, such as keratoconus.

Low-cost quasi-real-time elastography using B-mode ultrasound images.

A low cost, quasi real-time elastography system, displacement-gradient elastography (DGE), was developed by applying digital image correlation (DIC) method and smoothing algorithm to B-mode ultrasound images. In order to achieve quasi real-time elastogram display, a new fast pattern matching algorithm, decoupled cross-correlation (DCC), was proposed and validated. By applying the DGE to various phantoms, elastograms were generated to identify the lesion with wide variations of stiffness ratio and applied strain. The performance of DGE was qualitatively compared with those from a high-end ultrasound scanner using the elastograms of a commercial elastography breast phantom. DGE was also applied to the ultrasound images of human breast lesions in various BI-RADS categories. This study suggests that DGE may have comparable performance to conventional elastography in detecting breast cancer, while it can be easily implemented onto conventional ultrasound scanners.

Investigation of microstructure, mechanical properties and cellular viability of poly(L-lactic acid) tissue engineering scaffolds prepared by different thermally induced phase separation protocols.

Two thermally induced phase separation (TIPS) methods have been used to fabricate biodegradable poly(L-lactic acid) (PLLA) tissue engineering scaffolds each with fibrous (F-TIPS) and porous (P-TIPS) microstructures. Three levels of PLLA concentration (3, 5 and 7 wt%) were employed in each fabrication method and both wet and dry specimens were studied. Simple compression testing revealed that an elastic-plastic representation of the mechanical behavior was possible for all specimens. Both elastic and plastic moduli were higher for the P-TIPS, for higher polymer concentration, and might be somewhat higher for dry as opposed to wet specimens. For F-TIPS specimens, permanent deformation occurred successively during cyclic deformation but a "memory effect" simplified the behavior. Although F-TIPS microstructure better resembled the natural extracellular matrix, human osteosarcoma fibroblast cells showed more consistent viability in the P-TIPS scaffolds under our unloaded test protocols. Biodegradation in cell culture medium resulted in a decreased elastic moduli for F-TIPS specimens. Information presented regarding the microstructure, mechanical properties and cell viability of these PLLA scaffolds that should help reduce the number of iterations involved in developing tissue engineering products.

Application of digital image cross-correlation and smoothing function to the diagnosis of breast cancer.

Digital image correlation (DIC) algorithm was applied to 2D and 3D B-mode ultrasound (US) images to create 2D and 3D elastograms based on displacement-gradient. The roughness of elastograms caused by signal noises and sub-pixel errors could be greatly improved by employing the smoothing function based on the penalized least square regression method. Using the smoothed elastogram, the size and the relative modulus of the inclusion could be estimated with a reasonable accuracy. The study suggests that the 2D and 3D displacement-gradient elastograms acquired by the combination of DIC and smoothing function have the potential to diagnose pathological tissues in-vivo, and to provide new information that is related to tissue structure and/or pathology.