Realistic Breast Phantoms Research Articles

Automated assessment of task-based performance of digital mammography and tomosynthesis systems using an anthropomorphic breast phantom and deep learning-based scoring.

Conventional metrics used for assessing digital mammography (DM) and digital breast tomosynthesis (DBT) image quality, including noise, spatial resolution, and detective quantum efficiency, do not necessarily predict how well the system will perform in a clinical task. A number of existing phantom-based methods have their own limitations, such as unrealistic uniform backgrounds, subjective scoring using humans, and regular signal patterns unrepresentative of common clinical findings. We attempted to address this problem with a realistic breast phantom with random hydroxyapatite microcalcifications and semi-automated deep learning-based image scoring. Our goal was to develop a methodology for objective task-based assessment of image quality for tomosynthesis and DM systems, which includes an anthropomorphic phantom, a detection task (microcalcification clusters), and automated performance evaluation using a convolutional neural network. Experimental 2D and pseudo-3D mammograms of an anthropomorphic inkjet-printed breast phantom with inserted microcalcification clusters were collected on clinical mammography systems to train a signal-present/signal-absent image classifier based on Resnet-18 architecture. In a separate validation study using simulations, this Resnet-18 classifier was shown to approach the performance of an ideal observer. Microcalcification detection performance was evaluated as a function of four dose levels using receiver operating characteristic (ROC) analysis [i.e., area under the ROC curve (AUC)]. To demonstrate the use of this evaluation approach for assessing different technologies, the method was applied to two different mammography systems, as well as to mammograms with re-binned pixels emulating a lower-resolution X-ray detector. Microcalcification detectability, as assessed by the deep learning classifier, was observed to vary with the exposure incident on the breast phantom for both DM and tomosynthesis. At full dose, experimental AUC was 0.96 (for DM) and 0.95 (for DBT), whereas at half dose, it dropped to 0.85 and 0.71, respectively. AUC performance on DM was significantly decreased with an effective larger pixel size obtained with re-binning. The task-based assessment approach also showed the superiority of a newer mammography system compared with an older system. An objective task-based methodology for assessing the image quality of mammography and tomosynthesis systems is proposed. Possible uses for this tool could be quality control, acceptance, and constancy testing, assessing the safety and effectiveness of new technology for regulatory submissions, and system optimization. The results from this study showed that the proposed evaluation method using a deep learning model observer can track differences in microcalcification signal detectability with varied exposure conditions.

A comprehensive numerical procedure for high-intensity focused ultrasound ablation of breast tumour on an anatomically realistic breast phantom.

High-Intensity Focused Ultrasound (HIFU) as a promising and impactful modality for breast tumor ablation, entails the precise focalization of high-intensity ultrasonic waves onto the tumor site, culminating in the generation of extreme heat, thus ablation of malignant tissues. In this paper, a comprehensive three-dimensional (3D) Finite Element Method (FEM)-based numerical procedure is introduced, which provides exceptional capacity for simulating the intricate multiphysics phenomena associated with HIFU. Furthermore, the application of numerical procedures to an anatomically realistic breast phantom (ARBP) has not been explored before. The integrity of the present numerical procedure has been established through rigorous validation, incorporating comparative assessments with previous two-dimensional (2D) simulations and empirical data. For ARBP ablation, the administration of a 0.1 MPa pressure input pulse at a frequency of 1.5 MHz, sustained at the focal point for 10 seconds, manifests an ensuing temperature elevation to 80°C. It is noteworthy that, in contrast, the prior 2D simulation using a 2D phantom geometry reached just 72°C temperature under the identical treatment regimen, underscoring the insufficiency of 2D models, ascribed to their inherent limitations in spatially representing acoustic energy, which compromises their overall effectiveness. To underscore the versatility of this numerical platform, a simulation of a more clinically relevant HIFU therapy procedure has been conducted. This scenario involves the repositioning of the ultrasound focal point to three separate lesions, each spaced at 3 mm intervals, with ultrasound exposure durations of 6 seconds each and a 5-second interval for movement between focal points. This approach resulted in a more uniform high-temperature distribution at different areas of the tumour, leading to the ablation of almost all parts of the tumour, including its verges. In the end, the effects of different abnormal tissue shapes are investigated briefly as well. For solid mass tumors, 67.67% was successfully ablated with one lesion, while rim-enhancing tumors showed only 34.48% ablation and non-mass enhancement tumors exhibited 20.32% ablation, underscoring the need for multiple lesions and tailored treatment plans for more complex cases.

A Breast Tumor Monitoring Vest with Flexible UWB Antennas-A Proof-of-Concept Study Using Realistic Breast Phantoms.

Breast cancers can appear and progress rapidly, necessitating more frequent monitoring outside of hospital settings to significantly reduce mortality rates. Recently, there has been considerable interest in developing techniques for portable, user-friendly, and low-cost breast tumor monitoring applications, enabling frequent and cost-efficient examinations. Microwave technique-based breast cancer detection, which is based on differential dielectric properties of malignant and healthy tissues, is regarded as a promising solution for cost-effective breast tumor monitoring. This paper presents the development process of the first proof-of-concept of a breast tumor monitoring vest which is based on the microwave technique. Two unique vests are designed and evaluated on realistic 3D human tissue phantoms having different breast density types. Additionally, the measured results are verified using simulations carried out on anatomically realistic voxel models of the electromagnetic simulations. The radio channel characteristics are evaluated and analyzed between the antennas embedded in the vest in tumor cases and reference cases. Both measurements and simulation results show that the proposed vest can detect tumors even if only 1 cm in diameter. Additionally, simulation results show detectability with 0.5 cm tumors. It is observed that the detectability of breast tumors depends on the frequency, antenna selection, size of the tumors, and breast types, causing differences of 0.5-30 dB in channel responses between the tumorous and reference cases. Due to simplicity and cost-efficiency, the proposed channel analysis-based breast monitoring vests can be used for breast health checks in smaller healthcare centers and for user-friendly home monitoring which can prove beneficial in rural areas and developing countries.

Investigating the Use of Traveltime and Reflection Tomography for Deep Learning-Based Sound-Speed Estimation in Ultrasound Computed Tomography.

Ultrasound computed tomography (USCT) quantifies acoustic tissue properties such as the speed-of-sound (SOS). Although full-waveform inversion (FWI) is an effective method for accurate SOS reconstruction, it can be computationally challenging for large-scale problems. Deep learning-based image-to-image learned reconstruction (IILR) methods can offer computationally efficient alternatives. This study investigates the impact of the chosen input modalities on IILR methods for high-resolution SOS reconstruction in USCT. The selected modalities are traveltime tomography (TT) and reflection tomography (RT), which produce a low-resolution SOS map and a reflectivity map, respectively. These modalities have been chosen for their lower computational cost relative to FWI and their capacity to provide complementary information: TT offers a direct SOS measure, while RT reveals tissue boundary information. Systematic analyses were facilitated by employing a virtual USCT imaging system with anatomically realistic numerical breast phantoms. Within this testbed, a supervised convolutional neural network (CNN) was trained to map dual-channel (TT and RT images) to a high-resolution SOS map. Single-input CNNs were trained separately using inputs from each modality alone (TT or RT) for comparison. The accuracy of the methods was systematically assessed using normalized root mean squared error (NRMSE), structural similarity index measure (SSIM), and peak signal-to-noise ratio (PSNR). For tumor detection performance, receiver operating characteristic analysis was employed. The dual-channel IILR method was also tested on clinical human breast data. Ensemble average of the NRMSE, SSIM, and PSNR evaluated on this clinical dataset were 0.2355, 0.8845, and 28.33 dB, respectively.

Improvement of Phased Antenna Array Applied in Focused Microwave Breast Hyperthermia.

Focused microwave breast hyperthermia (FMBH) employs a phased antenna array to perform beamforming that can focus microwave energy at targeted breast tumors. Selective heating of the tumor endows the hyperthermia treatment with high accuracy and low side effects. The effect of FMBH is highly dependent on the applied phased antenna array. This work investigates the effect of polarizations of antenna elements on the microwave-focusing results by simulations. We explore two kinds of antenna arrays with the same number of elements using different digital realistic human breast phantoms. The first array has all the elements' polarization in the vertical plane of the breast, while the second array has half of the elements' polarization in the vertical plane and the other half in the transverse plane, i.e., cross polarization. In total, 96 sets of different simulations are performed, and the results show that the second array leads to a better focusing effect in dense breasts than the first array. This work is very meaningful for the potential improvement of the antenna array for FMBH, which is of great significance for the future clinical applications of FMBH. The antenna array with cross polarization can also be applied in microwave imaging and sensing for biomedical applications.

Learned Full Waveform Inversion Incorporating Task Information for Ultrasound Computed Tomography.

Ultrasound computed tomography (USCT) is an emerging imaging modality that holds great promise for breast imaging. Full-waveform inversion (FWI)-based image reconstruction methods incorporate accurate wave physics to produce high spatial resolution quantitative images of speed of sound or other acoustic properties of the breast tissues from USCT measurement data. However, the high computational cost of FWI reconstruction represents a significant burden for its widespread application in a clinical setting. The research reported here investigates the use of a convolutional neural network (CNN) to learn a mapping from USCT waveform data to speed of sound estimates. The CNN was trained using a supervised approach with a task-informed loss function aiming at preserving features of the image that are relevant to the detection of lesions. A large set of anatomically and physiologically realistic numerical breast phantoms (NBPs) and corresponding simulated USCT measurements was employed during training. Once trained, the CNN can perform real-time FWI image reconstruction from USCT waveform data. The performance of the proposed method was assessed and compared against FWI using a hold-out sample of 41 NBPs and corresponding USCT data. Accuracy was measured using relative mean square error (RMSE), structural self-similarity index measure (SSIM), and lesion detection performance (DICE score). This numerical experiment demonstrates that a supervised learning model can achieve accuracy comparable to FWI in terms of RMSE and SSIM, and better performance in terms of task performance, while significantly reducing computational time.

The Effect of Respiratory Motion in Breast Intensity-modulated Radiation Therapy: 3D-Printed Dynamic Phantom Study.

This study evaluated the dosimetric effect of respiratory motion in intensity-modulated radiation therapy (IMRT) for breast cancer using a three-dimensional (3D)-printed dynamic phantom. Computed tomography (CT) data from breast cancer patients were used to create a 3D-printed breast phantom. Various types of treatment plans were generated using CT images acquired at the exhalation phase. Different infill densities in the 3D-printed phantom were tested to validate their effect on simulating the average human breast tissue density. Plans were delivered to the 3D-printed dynamic phantom in the exhalation position and free-breathing motion. Dosimetric verification was performed using Gafchromic EBT3 films. After changing the infill density to obtain Hounsfield Unit values similar to those of human breast tissue, a realistic patient-specific breast phantom was fabricated using a 3D printer at 80% infill density. The gamma passing rates of the dose distribution delivered in the exhalation phase and free-breathing motion were 92% or more. In addition, the dynamic phantom doses with free-breathing motion were directly compared with the static phantom dose in terms of sagittal dose profiles. Gamma passing rates of >93% and 90% were achieved at 3%/3 mm and 3%/2 mm, respectively. Despite a blurred dose distribution, the dose map of the film measurement with respiratory motion could be delivered without significantly increasing the dose heterogeneity of the tumor bed or loss of target coverage. Our findings demonstrate that the impact of respiration on breast IMRT for whole-breast irradiation was not significant, even in the tumor bed.

Effect of HIFU-Induced Thermal Ablation in Numerical Breast Phantom

Breast cancer is a leading cause of cancer-related deaths in women, and treatment involved invasive surgery such as lumpectomy. In the last decade, a non-invasive, non-contact high-intensity focused ultrasound (HIFU) therapy was developed for treatment with promising results. However, its success rate depends on patient selection, tissue heterogeneities, HIFU operational parameters, and even imaging techniques. In this emerging field, computer simulations can provide us with a much-needed platform to learn, test, and deduce results virtually before conducting experiments. In this study, we used three different classes of anatomically realistic numerical breast phantoms from clinical contrast-enhanced magnetic resonance imaging (MRI) data, including scattered-, heterogeneous-, and extremely dense-type breasts. Upon assigning the appropriate acoustic and optical parameters to the tissues within, we simulated HIFU propagation by using the k-Wave toolbox in MATLAB and compared the changes introduced in the three types of breasts. It was found that scattered-type breast was best-suited for HIFU therapy. Furthermore, we simulated light-beam propagation with the ValoMC toolbox in MATLAB after introducing the lesion to compare the distribution of the initial pressure generated via the photoacoustic effect. This simulation study will be of significant clinical impact, especially in the study and management of HIFU-based treatments, which are individual/tissue-selective in nature.

Autonomous Robotic System for Breast Biopsy With Deformation Compensation

Image-guided biopsy is the most common technique for breast cancer diagnosis. Although magnetic resonance imaging (MRI) has the highest sensitivity in breast lesion detection, ultrasound (US) biopsy guidance is generally preferred due to its non-invasiveness and real-time image feedback during the insertion. In this work, we propose an autonomous robotic system for US-guided biopsy of breast lesions identified on pre-operative MRI. After initial MRI to breast registration, the US probe attached to the robotic manipulator compresses the breast tissues until a pre-determined force level is reached. This technique, known as preloading, will allow to minimize lesion displacement during the needle insertion. Our workflow integrates a deformation compensation strategy based on patient-specific biomechanical model to update the US probe orientation keeping the target lesion on the US image plane during compression. By relying on a deformation model, the proposed system does not require lesion visibility on US. Experimental evaluation is performed to assess the performance of the system on a realistic breast phantom with 15 internal lesions, considering different preloading forces. The deformation compensation strategy allows to improve localization accuracy, and as a consequence final probe positioning, for all considered lesions. Median lesion localization error is 3.3 mm, which is lower than the median lesion radius, when applying a preloading of 2 N, which guarantees both minimal needle insertion error and tissue stress.

Computationally efficient deep learning-based image reconstruction for sound speed estimation in ultrasound computed tomography

Ultrasound computed tomography (USCT) has great potential for breast cancer imaging due to its ability to quantify acoustic tissue properties such as the speed of sound (SOS). Full-waveform inversion (FWI)-based image reconstruction methods can produce high spatial resolution SOS images, but these methods are computationally demanding, especially when formulated in the time-domain. Deep learning (DL)-based methods can be computationally efficient and provide high-quality object estimates as studied in various medical image reconstruction problems. However, there remains a further need to develop and assess the DL-based image reconstruction methods for USCT for use in medical imaging applications. Here, we present a supervised DL approach for USCT image reconstruction that is formulated as an image-to-image translation problem. Specifically, a low-resolution SOS image obtained via traveltime tomography and a high-resolution reflection tomography image are used as inputs to a U-Net, which is trained to estimate a high-resolution SOS map. These input images are complementary, providing both quantitative information about SOS values and information about tissue boundaries. Numerical studies that employ realistic numerical breast phantoms demonstrated that the proposed method produced high-quality SOS estimates based on image quality metrics such as NMSE, SSIM, and PSNR.

Virtual imaging trials to investigate impact of skin color on three-dimensional optoacoustic tomography of the breast

Optoacoustic tomography (OAT), also known as photoacoustic computed tomography, is being actively developed for breast imaging applications. The endogenous optical contrast in OAT images is associated with oxygen saturation and concentrations of chromophores, such as hemoglobin, melanin, fat, and water, within the tissue. In OAT breast imaging, near-infrared light propagates through the skin, where the optical energy is absorbed primarily by melanin. The photoacoustic effect results in the generation of a pressure wavefield, and the propagated pressure wavefield is measured by ultrasonic transducers located on a measurement aperture surrounding the breast. Thus, the melanin concentration influences lesion contrast in OAT images. However, the extent to which skin color affects lesion detectability in OAT breast imaging remains unexplored. To address this, we generate realistic optoacoustic 3D numerical breast phantoms containing a lesion at different locations (three depths and two polar angles) with five skin colors and virtually acquire optoacoustic data employing them. To assess the skin color impact on lesion detectability, we quantify numerical observer performance for a signal-known-exactly and background-known-exactly detection task. The results confirm that the signal-to-noise ratio of the test statistic is degraded in darker skin, and the extent depends on lesion locations and the light delivery system design.

Three-dimensional time-domain full-waveform inversion for ring-array-based ultrasound computed tomography

Ultrasound computed tomography (USCT) is a promising imaging technique for breast cancer diagnosis. Full-waveform inversion (FWI)-based image reconstruction methods based on the acoustic wave equation can produce quantitatively accurate images of acoustic properties with higher spatial resolution compared to ray-based methods. One common USCT design employs a circular ring-array of elevation-focused ultrasonic transducers that can be translated orthogonally to the imaging plane to achieve volumetric scanning. Currently, slice-by-slice (SBS) reconstruction methods are often used in this setting. Such methods use a 2D wave physic model to reconstruct a cross-sectional image for each position of the ring-array and vertically stack them together to render the 3D volume. However, this approach neglects the 3D wave propagation physics and transducer’s focusing properties, leading to out-of-plane scattering artifacts and degraded spatial resolution. To address this, a 3D time-domain FWI method that utilizes measurement data acquired at multiple adjacent ring-array locations is proposed. The reconstruction method was investigated using virtual imaging studies of ring-array-based USCT that employed realistic 3D numerical breast phantoms. The results demonstrated improved spatial resolution (both in-plane and off-plane) and reduced image artifacts compared to the SBS approach. Additionally, the impact of the number of ring-array locations on image quality is assessed.

A BIM-Based Algorithm for Quantitative Monitoring of Temperature Distribution During Breast Hyperthermia Treatments

Microwave hyperthermia (MH) treatment for breast cancer is a research interest due to its capability to initiate cell necrosis in malignant tumor or to enhance the effect of other treatment modalities such as chemotherapy. The goal of MH treatment is to increase temperature of malignant tumor up to 45°C based on the treatment plan; however, microwave energy focusing is a challenging problem and may cause unwanted hotspots on healthy tissues; therefore, there is a need to monitor the temperature. In this paper, an iterative differential microwave imaging algorithm for temperature monitoring is presented. The algorithm is based on Born iterative method (BIM) and Tikhonov regularization. Feasibility of the algorithm is shown by a large computational study using realistic digital breast phantoms via TM <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">z</sub> polarized 2-D scattered fields. Also, some results are given for calibrated scattering parameters, which are obtained from both a 3-D electromagnetic simulation program and a simple measurement setup. An approach for selection of matching medium in hyperthermia monitoring applications is also presented. The reconstructions are performed with scattered field data collected at 11 discrete frequency points uniformly taken from the 0.5-1.5 GHz range. For a specific heating scenario in 2-D problem, reconstruction error is lower than 0.3% with ±10% noise on reference dielectric property distribution and 40 dB signal to noise ratio (SNR). The results show that the proposed approach provides up to 3°C and 0.1°C resolution in temperature estimation with ±10% noise on reference dielectric property distribution for 30 dB and 60 dB SNR values, respectively.

Continuous Monitoring of Hyperthermia Treatment of Breast Tumors With Singular Sources Method

Qualitative microwave imaging methods provide a fast and accurate reconstruction of the shapes of the targets from scattered electric field measurements. In this paper, a qualitative imaging method, the singular sources method (SSM), is analyzed for two-dimensional transverse magnetic electromagnetic (2D-TM EM) inverse scattering scenarios. The contribution of the paper can be briefly summarized as: (i) The SSM was previously proposed for the far-field scenario, here we extend the SSM in the near field - inhomogeneous background setup. While making the extension each step (which includes an integral equation) is explained by exploiting the linearity and reciprocity principles to give physical insights. (ii) The relation between the electrical parameters of the scatterers and the indicator of SSM is derived. (iii) To analyze the performance of SSM in real-world problems, the proposed method is tested for monitoring hyperthermia treatment problems with a realistic breast model. Obtained results show that the SSM can handle realistic breast phantoms for monitoring hyperthermia problems.

Contrast Source Inversion Enhanced Confocal Imaging for Highly Heterogeneous Breast Media in Microwave Mammography

This study introduces accuracy-enhanced confocal imaging (CI) algorithm for highly heterogeneous breast media that can be used in microwave mammography. In the presence of heterogeneous backgrounds, such as breast media, CI accuracy highly depends on several background assumptions, such as homogeneity, multilayered structure, and adipose-dominant tissue. In the case of a highly dense breast dominated by fibroglandular tissue, the traditional CI suffers from inaccuracy due to a large discrepancy between the actual and assumed backgrounds. In this study, we used contrast source inversion (CSI) known as a promising inverse scattering approach, which could provide an accurate estimation of total fields in the region of interest. This feature enables us to generate accurate propagation models as the Green's functions. The numerical tests using realistic breast phantoms show that our method significantly enhances reconstruction accuracy without any prior knowledge of the background media, thus highly contributing to cancer tissue detection.

Detection and Localization of Breast Cancer Using UWB Microwave Technology and CNN-LSTM Framework

Ultrawideband (UWB) microwave detection technology, which is low cost and harmless, has been intensively studied and developed for breast cancer detection. In this study, a composite end-to-end framework that consists of convolutional neural network (CNN) and long-short-term memory (LSTM) is proposed, which can realize the tasks of detecting and quadrant locating the breast tumor simultaneously without any complicated microwave imaging processing. In order to verify the proposed network, three datasets are constructed. First, the microwave signals are solved by the auxiliary differential equation-finite-difference time-domain (ADE-FDTD) solver and the magnetic resonance imaging (MRI)-derived breast models with varied breast densities. Then, Simulation datasets 1 and 2 with varied dielectric properties are constructed. Furthermore, an experimental dataset is constructed through an experimental system consisting of the UWB pulse generation module, UWB antennas, the realistic breast phantom, and sampling oscilloscope. With the proposed network, the overall prediction accuracies of the three datasets reach 99.56%, 98.94%, and 89.50%. These promising results demonstrate the effectiveness and accuracy of the proposed deep learning framework for microwave breast cancer detection.

Numerical study of magnetic hyperthermia ablation of breast tumor on an anatomically realistic breast phantom.

Magnetic fluid hyperthermia (MFH) is a novel reliable technique with excellent potential for thermal therapies and treating breast tumours. This method involves injecting a magnetic nanofluid into the tumour and applying an external AC magnetic field to induce heat in the magnetic nanoparticles (MNPs) and raise the tumour temperature to ablation temperature ranges. Because of the complexity of considering and coupling all different physics involves in this phenomenon and also due to the intricacy of a thorough FEM numerical study, few FEM-based studies address the entire MFH process as similar to reality as possible. The current study investigates a FEM-based three-dimensional numerical simulation of MFH of breast tumours as a multi-physics problem. An anatomically realistic breast phantom (ARBP) is considered, some magnetic nanofluid is injected inside the tumour, and the diffusion phenomenon is simulated. Then, the amount of heat generated in the MNP-saturated tumour area due to an external AC magnetic field is simulated. In the end, the fraction of tumour tissue necrotized by this temperature rise is evaluated. The study’s results demonstrate that by injecting nanofluid and utilizing seven circular copper windings with each coil carrying 400 A current with a frequency of 400 kHz for generating the external AC magnetic field, the temperature in tumour tissue can be raised to a maximum of about 51.4°C, which leads to necrosis of entire tumour tissue after 30 minutes of electromagnetic field (EMF) exposure. This numerical platform can depict all four various physics involved in the MFH of breast tumours by numerically solving all different equation sets coupled together with high precision. Thus, the proposed model can be utilized by clinicians as a reliable tool for predicting and identifying the approximate amount of temperature rise and the necrotic fraction of breast tumour, which can be very useful to opt for the best MFH therapeutic procedure and conditions based on various patients. In future works, this numerical platform’s results should be compared with experimental in-vivo results to improve and modify this platform in order to be ready for clinical applications.

A compact ultrawideband slotted patch antenna for early stage breast tumor detection applications

AbstractThis paper presents a compact ultrawideband slotted patch antenna for early stage tumor detection applications using microwave breast imaging systems. In order to reduce any reflections between the antenna and breast, the antenna is designed and optimized inside a coupling medium with electrical properties similar to that of normal breast tissue as a coupling medium. The wideband performance of the antenna is obtained by adding a rectangular slit and L-shaped sleeves on the ground plane. While the added rectangular slit mainly improves the matching, the L-shaped sleeves improve both the matching and the bandwidth. The antenna operates from 0.9 to 9.6 GHz with a fractional bandwidth of ~165$\%$, providing a good penetration and resolution. In addition, the proposed antenna has a very compact size of 16 × 16 mm2. Furthermore, an antenna array consisting of 31 antennas is formed around a realistic breast phantom in the simulation environment to demonstrate the imaging of a spherical tumor with a diameter of 5 mm inside the breast. The results confirm that the proposed antenna can be used for early stage breast tumor detection purposes.

Enhanced Machine Learning Approach for Accurate and Fast Resolution of Inverse Scattering Problem in Breast Cancer Detection

An improved machine learning approach is presented in this paper to guarantee the fast convergence of the Born Iterative Method, even in the presence of strong scatterers, by assuming a single operating frequency and a reduced number of antennas in the scattering setup. The initial estimation of the dielectric profile, provided by the Born Iterative Method, was processed by a specific convolutional neural network to improve the reconstruction using a fast machine learning approach. To ensure rapid convergence, a proper choice of the initial guess in terms of the minimum permittivity value over the entire domain was also made. Numerical validations on realistic breast phantoms were illustrated, demonstrating an average error of 2.4% and an accuracy greater than 96% for all considered tests, even when considering a single operating frequency and a reduced amount of training data.

An Adaptive Finite Element/Finite Difference Domain Decomposition Method for Applications in Microwave Imaging

A new domain decomposition method for Maxwell’s equations in conductive media is presented. Using this method, reconstruction algorithms are developed for the determination of the dielectric permittivity function using time-dependent scattered data of an electric field. All reconstruction algorithms are based on an optimization approach to find the stationary point of the Lagrangian. Adaptive reconstruction algorithms and space-mesh refinement indicators are also presented. Our computational tests show the qualitative reconstruction of the dielectric permittivity function using an anatomically realistic breast phantom.