Elia A. Attardo

Bone dielectric property variation as a function of mineralization at microwave frequencies

A critical need exists for new imaging tools to more accurately characterize bone quality beyond the conventional modalities of dual energy X-ray absorptiometry (DXA), ultrasound speed of sound, and broadband attenuation measurements. In this paper we investigate the microwave dielectric properties of ex vivo trabecular bone with respect to bulk density measures. We exploit a variation in our tomographic imaging system in conjunction with a new soft prior regularization scheme that allows us to accurately recover the dielectric properties of small, regularly shaped and previously spatially defined volumes. We studied six excised porcine bone samples from which we extracted cylindrically shaped trabecular specimens from the femoral heads and carefully demarrowed each preparation. The samples were subsequently treated in an acid bath to incrementally remove volumes of hydroxyapatite, and we tested them with both the microwave measurement system and a micro-CT scanner. The measurements were performed at five density levels for each sample. The results show a strong correlation between both the permittivity and conductivity and bone volume fraction and suggest that microwave imaging may be a good candidate for evaluating overall bone health.

Whole-System Electromagnetic Modeling for Microwave Tomography

This letter presents a full-wave, whole-system modeling of microwave imaging tomography (MWT) systems to be used as the forward model in reconstruction (inverse) algorithms. The full geometry including antennas and their ports is simulated via a finite element method (FEM) approach. A new technique is used to compute the antenna operation in the system, which provides a general method to enforce the excitation as a specific modal distribution and to extract the voltage and current from the employed antenna. We report results for a complete microwave imaging (MWI) system with comparison between measured and simulated data.

GPU acceleration of Algebraic Multigrid for low-frequency finite element methods

This paper introduces a GPU acceleration of a Wavelet-based Algebraic Multigrid used as preconditioner for solving the Laplaces equation discretized by Finite Element Method. We conduct some tests using a CPU-based direct solver, a CPU-based Preconditined Conjugate Gradient (PCG), and a GPU-based PCG. Finally, we report the solution time and the speed-up achieved in solving the discretized problem.

Finite element modeling for microwave tomography

This paper presents a Finite Element Modelling for Microwave Imaging tomography based on edge basis functions. In particular a new technique is used to compute the antenna parameters. This novel procedure will provide a general method to enforce the excitation as a specific modal distribution, and to extract the voltage from the employed antenna. We show the model for the Microwave Imaging system including the antennas and we report a comparison between measured and simulated data.

A multi-GPU acceleration for 3D imaging of the prostate

Transrectal Electric Impedance Tomography (TREIT) has been proposed jointly with ultrasound (US) imaging of the prostate to enhance the standard clinical imaging. Reconstructing TREIT images involves a solution of an inverse problem. The reconstruction is based on two steps: solving and updating an estimate of the dielectric property distribution through solution of an inverse problem. In this paper we consider a multi-GPU acceleration, which will allow us to significantly speed up the solution of the inverse problem. By conducting numerical experiments we compare results in assembling the Jacobian matrix by a CPU-based multiple computational cores server, a single GPU acceleration, and eventually a multi-GPU.

GPU-Accelerated real time simulation of Radio Frequency Ablation thermal dose

In this paper we present results showing that Graphic Processing Units (GPUs) can be used to accelerate the simulation of the thermal field dose during Radio Frequency Ablation (RFA). Specifically we show that this simulation can be conducted in real-time, allowing the development of intraoperative guidance platforms that track and display the thermal lesion as it forms during the intervention. Real time simulation has not been reported before, and has been a critical factor preventing development of intraoperative guidance tools.

A finite element based hybrid source-type scheme for Microwave Imaging

We present a novel imaging approach to tackle inverse scattering problems in non-canonical configurations. The approach relies on a hybrid formulation of the problem in terms of source-type integral equations, in which the non-canonical nature of the scenario is handled by means of the underlying finite element method framework. In particular, the adopted hybrid model mixes the contrast source formulation and the contrast source extended Born one, and takes advantage of their different (complementary) features. As shown by means of a preliminary numerical example, the proposed hybrid method is capable to achieve meaningful reconstruction in cases in which the same inversion algorithm (cast with respect to one of the two models) fails.

Contrast Source Extended Born Inversion in Noncanonical Scenarios Via FEM Modeling

The contrast source extended Born formulation of the electromagnetic scattering is based on a simple rewriting of standard source type integral equations and allows enhancing the reliability of the inverse problems solution, especially when losses are present in the scenario. The method has been developed in the framework of integral equations, so that its application has been so far limited to canonical scenarios, where the Greens function is available. To overcome this drawback, in this work we consider a generalization of the contrast source extended Born scheme based on a finite element formalism, which opens the way to a much more flexible algorithm, capable of handling noncanonical configurations. Additionally, we also propose a new hybrid strategy that simultaneously exploits both the well-known contrast source formulation and the contrast source extended Born one and it is capable to achieve improved imaging performance with respect to both of them. Numerical examples are reported to illustrate the performance of the proposed inversion methods.

Magnetic field shimming in MRI with controlled polarization and SAR limitation

The homogeneity of RF field B1 used in Magnetic Resonance Imaging (MRI) is a crucial problem in such a kind of diagnostic tool. Because of the non homogeneity of the scenario, and of the intrinsic difficulty of such a constraint, the field usually obtained is far from the one required. This is especially the case for recent high-and ultra-high field (B0>7 T), with the consequent increase of the RF frequency of field. As a result, the issue of shimming the B1 field has received considerable attention in the recent literature [1]. The problem is rendered even more complex from the need, in partial and whole body imaging, to keep under control the Specific Absorption Rate (SAR) for partial and whole body imaging. Recent methods have therefore addressed the combined task of optimizing the field and limiting the SAR [2, 3]; these approaches, however, do not address control of the polarization of the RF field itself, which is also a relevant issue. As a matter of fact, even in the presence of a significant homogeneity of the field intensity, degradation of field polarization from the desired circular left-hand to elliptical up to linear polarization is the cause of higher power emission for the coils and a reduced signal-to-noise ratio (SNR) [4]. Herein, we show that controlling the ripple of the RF field is not enough to control the polarization, and we present a method to optimize the field intensity while keeping under control the homogeneity of intensity and polarization of the RF magnetic field B1, as well as SAR limitations for the case of partial body imaging. In particular, the proposed method is able to comply with constraints on SAR and B1 intensity as enforced by FDA [5] or recommended by ICNIRP [6].

Full-wave assessment of feasibility guidelines for 3-d microwave imaging of brain strokes

Microwave imaging (MWI) can provide truly non-invasive and affordable tools to perform biomedical monitoring for diagnostic purposes, as it exploits non-ionizing - therefore harmless - radiations and relatively cheap and portable devices. In this respect, besides breast cancer imaging, which is still the mostly addressed topic by researchers worldwide, there are several emerging applications relevant to the so-called aging society scenario, such as early diagnosis of bones degradation [P. M. Meaney, T. Zhou, D. Goodwin, A. Golnabi, E. A. Attardo and K. D. Paulsen, “Bone Dielectric Property Variation as a Function of Mineralization at Microwave Frequencies,” Int. Journal of Biom. Imag. 2012] and monitoring of brain injuries due to strokes. In these applications, the capabilities of a microwave technology can be helpful to cooperate with existing diagnostic technologies, in order to improve the overall reliability and timeliness of the diagnosis.