Doga Gürsoy

Reconstruction artefacts in magnetic induction tomography due to patient's movement during data acquisition

Magnetic induction tomography (MIT) attempts to obtain the distribution of passive electrical properties inside the body. Eddy currents are induced in the body using an array of transmitter coils and the magnetic fields of these currents are measured by receiver coils. In clinical usage, the relative position of the coils to the body can change during data acquisition because of the expected/unexpected movements of the patient. Especially in respiration monitoring these movements will inevitably cause artefacts in the reconstructed images. In this paper, this effect was investigated for both state and frequency differential variants of MIT. It was found that a slight shift of the body in the transverse plane causes spurious perturbations on the surface. In reconstructions, this artefact on the surface propagates towards the centre in an oscillatory manner. It was observed that the movement can corrupt all the valuable information in state differential MIT, while frequency differential MIT seems more robust against movement effects. A filtering strategy is offered in order to decrease the movement artefacts in the images. To this end, monitoring of the patients movement during data acquisition is required.

An experimental clinical evaluation of EIT imaging with ℓ1 data and image norms

Electrical impedance tomography (EIT) produces an image of internal conductivity distributions in a body from current injection and electrical measurements at surface electrodes. Typically, image reconstruction is formulated using regularized schemes in which ℓ2-norms are used for both data misfit and image prior terms. Such a formulation is computationally convenient, but favours smooth conductivity solutions and is sensitive to outliers. Recent studies highlighted the potential of ℓ1-norm and provided the mathematical basis to improve image quality and robustness of the images to data outliers. In this paper, we (i) extended a primal-dual interior point method (PDIPM) algorithm to 2.5D EIT image reconstruction to solve ℓ1 and mixed ℓ1/ℓ2 formulations efficiently, (ii) evaluated the formulation on clinical and experimental data, and (iii) developed a practical strategy to select hyperparameters using the L-curve which requires minimum user-dependence. The PDIPM algorithm was evaluated using clinical and experimental scenarios on human lung and dog breathing with known electrode errors, which requires a rigorous regularization and causes the failure of reconstruction with an ℓ2-norm solution. The results showed that an ℓ1 solution is not only more robust to unavoidable measurement errors in a clinical setting, but it also provides high contrast resolution on organ boundaries.

Optimum Receiver Array Design for Magnetic Induction Tomography

Magnetic induction tomography (MIT) is an imaging modality that aims at mapping the distribution of the electrical conductivity inside the body. Eddy currents are induced in the body by magnetic induction and the resulting fields are measured by an array of receiver coils. In MIT, the location of the receivers affects the quality of the image reconstruction. In this paper, a fast deterministic algorithm was applied to obtain optimum receiver array designs for a given specific excitation. The design strategy is based on the iterative exclusion of receiver locations, which yield poor conductivity information, from the space spanning all possible locations until a feasible design is reached. The applicability of ldquoregionally focusedrdquo MIT designs that increase the image resolution at a particular region was demonstrated. Currently used design geometries and the corresponding reconstructed images were compared to the images obtained by optimized designs. The eigenvalue analysis of the Hessian matrix showed that the algorithm tends to maintain identical conductivity information content sensed by the receivers. Although the method does not guarantee finding the optimum design globally, the results demonstrate the practical usability of this algorithm in MIT experimental designs.

Imaging artifacts in magnetic induction tomography caused by the structural incorrectness of the sensor model

Magnetic induction tomography (MIT) is a noninvasive imaging modality that aims to reconstruct the interior electrical conductivity distribution of the human body. It uses magnetic induction to excite eddy currents in the body and an array of sensor coils to detect the perturbations in the magnetic field. Image reconstruction in MIT is usually carried out by minimizing the residuals between the estimated and measured quantities assuming a structurally correct model. Thus, any mismatch between the simulated and the true experimental coil setup alters the data and may cause artifacts in the images. In this paper, a simulation study was performed to investigate the effect of modeling mismatches on measurements and corresponding reconstructed images. It was found that slight distortions of the receivers may cause up to 20% deviations in the data considering a local and small perturbation in conductivity. Unless the geometry is modeled correctly, these artifacts may spoil the images particularly for the case of flexible systems that have many degrees of freedom and systems that require different adjustments for different imaging sessions. If the system does not need calibration, for instance as in the case of head applications, then a rigid mechanical support appears to be an important design issue to achieve a better image quality.

The effect of receiver coil orientations on the imaging performance of magnetic induction tomography

Magnetic induction tomography is an imaging modality which aims to reconstruct the conductivity distribution of the human body. It uses magnetic induction to excite the body and an array of sensor coils to detect the perturbations in the magnetic field. Up to now, much effort has been expended with the aim of finding an efficient coil configuration to extend the dynamic range of the measured signal. However, the merits of different sensor orientations on the imaging performance have not been studied in great detail so far. Therefore, the aim of the study is to fill the void of a systematic investigation of coil orientations on the reconstruction quality of the designs. To this end, a number of alternative receiver array designs with different coil orientations were suggested and the evaluations of the designs were performed based on the singular value decomposition. A generalized class of quality measures, the subclasses of which are linked to both the spatial resolution and uncertainty measures, was used to assess the performance on the radial and axial axes of a cylindrical phantom. The detectability of local conductivity perturbations in the phantom was explored using the reconstructed images. It is possible to draw the conclusion that the proper choice of the coil orientations significantly influences the number of usable singular vectors and accordingly the stability of image reconstruction, although the effect of increased stability on the quality of the reconstructed images was not of paramount importance due to the reduced independent information content of the associated singular vectors.

Anisotropic conductivity tensor imaging using magnetic induction tomography

Magnetic induction tomography aims to reconstruct the electrical conductivity distribution of the human body using non-contact measurements. The potential of the method has been demonstrated by various simulation studies and a number of phantom experiments. These studies have all relied on models having isotropic distributions of conductivity, although the human body has a highly heterogeneous structure with partially anisotropic properties. Therefore, whether the conventional modeling approaches used so far are appropriate for clinical applications or not is still an open question. To investigate the problem, we performed a simulation study to investigate the feasibility of (1) imaging anisotropic perturbations within an isotropic medium and (2) imaging isotropic perturbations inside a partially anisotropic background. The first is the case for the imaging of anomalies that have anisotropic characteristics and the latter is the case e.g. in lung imaging where an anisotropic skeletal muscle tissue surrounds the lungs and the rib cage. An anisotropic solver based on the singular value decomposition was used to attain conductivity tensor images to be compared with the ones obtained from isotropic solvers. The results indicate the importance of anisotropic modeling in order to obtain satisfactory reconstructions, especially for the imaging of the anisotropic anomalies, and address the resolvability of the conductivity tensor components.

Enhancing Impedance Imaging Through Multimodal Tomography

Several noninvasive modalities including electrical impedance tomography (EIT), magnetic induction tomography (MIT), and induced-current EIT (ICEIT) have been developed for imaging the electrical conductivity distribution within a human body. Although these modalities differ in how the excitation and detection circuitry (electrodes or coils) are implemented, they share a number of common principles not only within the image reconstruction approaches but also with respect to the basic principle of generating a current density distribution inside a body and recording the resultant electric fields. In this paper, we are interested in comparing differences between these modalities and in theoretically understanding the compromises involved, despite the increased hardware cost and complexity that such a multimodal system brings along. To systematically assess the merits of combining data, we performed 3-D simulations for each modality and for the multimodal system by combining all available data. The normalized sensitivity matrices were computed for each modality based on the finite element method, and singular value decomposition was performed on the resultant matrices. We used both global and regional quality measures to evaluate and compare different modalities. This study has shown that the condition number of the sensitivity matrix obtained from the multimodal tomography with 16-electrode and 16-coil is much lower than the condition number produced in the conventional 16-channel EIT and MIT systems, and thus, produced promising results in terms of image stability. An improvement of about 20% in image resolution can be achieved considering feasible signal-to-noise ratio levels.

Magnetic induction pneumography: a planar coil system for continuous monitoring of lung function via contactless measurements

Abstract Continuous monitoring of lung function is of particular interest to the mechanically ventilated patients during critical care. Recent studies have shown that magnetic induction measurements with single coils provide signals which are correlated with the lung dynamics and this idea is extended here by using a 5 by 5 planar coil matrix for data acquisition in order to image the regional thoracic conductivity changes. The coil matrix can easily be mounted onto the patient bed, and thus, the problems faced in methods that use contacting sensors can readily be eliminated and the patient comfort can be improved. In the proposed technique, the data are acquired by successively exciting each coil in order to induce an eddy-current density within the dorsal tissues and measuring the corresponding response magnetic field strength by the remaining coils. The recorded set of data is then used to reconstruct the internal conductivity distribution by means of algorithms that minimize the residual norm between the estimated and measured data. To investigate the feasibility of the technique, the sensitivity maps and the point spread functions at different locations and depths were studied. To simulate a realistic scenario, a chest model was generated by segmenting the tissue boundaries from NMR images. The reconstructions of the ventilation distribution and the development of an edematous lung injury were presented. The imaging artifacts caused by either the incorrect positioning of the patient or the expansion of the chest wall due to breathing were illustrated by simulations.

Single-step phase contrast x-ray imaging using photon counting detectors

Using solutions of spectral transport-of-intensity equations, we have demonstrated a single step method to retrieve absorption and phase changes for a wide range of x-ray imaging energies and material composition. We simulated a Cadmium-Zinc-Telluride based spectral detection system using a cascade model for investigations of breast mass and microcalcification detectability when using both absorption and phase images simultaneously.

Experimental/clinical evaluation of EIT image reconstruction with ℓ1 data and image norms

Electrical impedance tomography (EIT) image reconstruction is ill-posed, and the spatial resolution of reconstructed images is low due to the diffuse propagation of current and limited number of independent measurements. Generally, image reconstruction is formulated using a regularized scheme in which l2 norms are preferred for both the data misfit and image prior terms due to computational convenience which result in smooth solutions. However, recent work on a Primal Dual-Interior Point Method (PDIPM) framework showed its effectiveness in dealing with the minimization problem. l1 norms on data and regularization terms in EIT image reconstruction address both problems of reconstruction with sharp edges and dealing with measurement errors. We aim for a clinical and experimental evaluation of the PDIPM method by selecting scenarios (human lung and dog breathing) with known electrode errors, which require a rigorous regularization and cause the failure of reconstructions with l2 norm. Results demonstrate the applicability of PDIPM algorithms, especially l1 data and regularization norms for clinical applications of EIT showing that l1 solution is not only more robust to measurement errors in clinical setting, but also provides high contrast resolution on organ boundaries.