RESUMEN

Ultrasound (US) imaging is a widely used medical imaging modality for the diagnosis, monitoring, and surgical planning for kidney conditions. Thus, accurate segmentation of the kidney and internal structures in US images is essential for the assessment of kidney function and the detection of pathological conditions, such as cysts, tumors, and kidney stones. Therefore, there is a need for automated methods that can accurately segment the kidney and internal structures in US images. Over the years, automatic strategies were proposed for such purpose, with deep learning methods achieving the current state-of-the-art results. However, these strategies typically ignore the segmentation of the internal structures of the kidney. Moreover, they were evaluated in different private datasets, hampering the direct comparison of results, and making it difficult to determination the optimal strategy for this task. In this study, we perform a comparative analysis of 7 deep learning networks for the segmentation of the kidney and internal structures (Capsule, Central Echogenic Complex (CEC), Cortex and Medulla) in 2D US images in an open access multi-class kidney US dataset. The dataset includes 514 images, acquired in multiple clinical centers using different US machines and protocols. The dataset contains the annotation of two experts, but 321 images with complete segmentation of all 4 classes were used. Overall, the results demonstrate that the DeepLabV3+ network outperformed the inter-rater variability with a Dice score of 78.0% compared to 75.6% for inter-rater variability. Specifically, DeepLabV3Plus achieved mean Dice scores of 94.2% for the Capsule, 85.8% for the CEC, 62.4% for the Cortex, and 69.6% for the Medulla. These findings suggest the potential of deep learning-based methods in improving the accuracy of kidney segmentation in US images.Clinical Relevance- This study shows the potential of DL for improving accuracy of kidney segmentation in US, leading to increased diagnostic efficiency, and enabling new applications such as computer-aided diagnosis and treatment, ultimately resulting in improved patient outcomes and reduced healthcare costs.1.

Asunto(s)

RESUMEN

Training in surgery is essential for surgeons to develop skill and dexterity. Physical training phantoms provide excellent haptic feedback and tissue properties for stitching and operating with authentic instruments and are easily available. However, they lack realistic traits and fail to reflect the complex environment of a surgical scene. Generative Adversarial Networks can be used for image-to-image translation, addressing the lack of realism in physical phantoms, by mapping patterns from the intraoperative domain onto the video stream captured during training with these surgical simulators. This work aims to achieve a successful I2I translation, from intra-operatory mitral valve surgery images onto a surgical simulator, using the CycleGAN model. Different experiments are performed - comparing the Mean Square Error Loss with the Binary Cross Entropy Loss; validating the Fréchet Inception Distance as a training and image quality metric; and studying the impact of input variability on the model performance. Differences between MSE and BCE are modest, with MSE being marginally more robust. The FID score proves to be very useful in identifying the best training epochs for the CycleGAN I2I translation architecture. Carefully selecting the input images can have a great impact in the end results. Using less style variability and input images with good feature details and clearly defined characteristics enables the network to achieve better results.Clinical Relevance- This work further contributes for the domain of realistic surgical training, successfully generating fake intra operatory images from a surgical simulator of the cardiac mitral valve.

Asunto(s)

RESUMEN

Semantic segmentation of anatomical structures in laparoscopic videos is a crucial task to enable the development of new computer-assisted systems that can assist surgeons during surgery. However, this is a difficult task due to artifacts and similar visual characteristics of anatomical structures on the laparoscopic videos. Recently, deep learning algorithms have been showed promising results on the segmentation of laparoscopic instruments. However, due to the lack of large public datasets for semantic segmentation of anatomical structures, there are only a few studies on this task. In this work, we evaluate the performance of five networks, namely U-Net, U-Net++, DynUNet, UNETR and DeepLabV3+, for segmentation of laparoscopic cholecystectomy images from the recently released CholecSeg8k dataset. To the best of our knowledge, this is the first benchmark performed on this dataset. Training was performed with dice loss. The networks were evaluated on segmentation of 8 anatomical structures and instruments, performance was quantified through the dice coefficient, intersection over union, recall, and precision. Apart from the U-Net, all networks obtained scores similar to each other, with the U-Net++ being the network with the best overall score with a mean Dice value of 0.62. Overall, the results show that there is still room for improvement in the segmentation of anatomical structures from laparoscopic videos. Clinical Relevance- The results of this study show the potential of deep neural networks for the segmentation of anatomical structures in laparoscopic images which can later be incorporated into computer-aided systems for surgeons.

Asunto(s)

RESUMEN

Ultrasound (US) is a medical imaging modality widely used for diagnosis, monitoring, and guidance of surgical procedures. However, the accurate interpretation of US images is a challenging task. Recently, portable 2D US devices enhanced with Artificial intelligence (AI) methods to identify, in real-time, specific organs are widely spreading worldwide. Nevertheless, the number of available methods that effectively work in such devices is still limited. In this work, we evaluate the performance of the U-NET architecture to segment the kidney in 2D US images. To accomplish this task, we studied the possibility of using multiple sliced images extracted from 3D US volumes to achieve a large, variable, and multi-view dataset of 2D images. The proposed methodology was tested with a dataset of 66 3D US volumes, divided in 51 for training, 5 for validation, and 10 for testing. From the volumes, 3792 2D sliced images were extracted. Two experiments were conducted, namely: (i) using the entire database (WWKD); and (ii) using images where the kidney area is > 500 mm2 (500KD). As a proof-of-concept, the potential of our strategy was tested in real 2D images (acquired with 2D probes). An average error of 2.88 ± 2.63 mm in the testing dataset was registered. Moreover, satisfactory results were obtained in our initial proof-of-concept using pure 2D images. In short, the proposed method proved, in this preliminary study, its potential interest for clinical practice. Further studies are required to evaluate the real performance of the proposed methodology. Clinical Relevance- In this work a deep learning methodology to segment the kidney in 2D US images is presented. It may be a relevant feature to be included in the recent portable US ecosystems easing the interpretation of image and consequently the clinical analysis.

Asunto(s)

RESUMEN

In quantitative susceptibility mapping, the tissue susceptibility is determined from the magnitude and phase of the gradient echo signal, which is influenced by the interplay of complex susceptibility and diffusion effect. Herein, we analytically analyze the influence of diffusion on magnitude and phase images generated by randomly arranged magnetic spheres as a model of intracerebral iron depositions. We demonstrate that both gradient and spin echo relaxation rate constants have a strong and nonlinear dependence on diffusion strength and give empirical formulas for magnitude and phase. This may be used in the future to improve QSM processing methods. In addition, we show that, in theory, combined acquisitions of gradient and spin echo can be used to determine the dimension of the magnetic spheres and the diffusion strength.

Asunto(s)

RESUMEN

The spin echo signal from vessels in Krogh's capillary model as well in the random distribution vessel model are studied by numerically solving the Bloch-Torrey equation. A comparison is made with the Gaussian local phase approximation, the Gaussian phase approximation and the strong-collision approximation. Differences between the Gaussian local phase approximation and the Gaussian phase approximation are explained. In the intermediate diffusion regime, the full numerical solution shows oscillations which are absent in any of the approximate solutions. In the limit of large diffusion coefficients, where the approximations become exact, the signal shows a linear-exponential decay governed by a single parameter. The features of the exact numerical solution can be explained by an analytically solvable discrete two-level model. There is a one-to-one correspondence between the different diffusion regimes and the three cases of the damped harmonic oscillator.

RESUMEN

Magnetic resonance imaging of lung tissue is strongly influenced by susceptibility effects between spin-bearing water molecules and air-filled alveoli. The measured lineshape, however, also depends on the interplay between susceptibility effects and blood-flow around alveoli that can be approximated as pseudo-diffusion. Both effects are quantitatively described by the Bloch-Torrey-equation, which was so far only solved for dephasing on the alveolar surface. In this work, we extend this model to the whole range of physiological relevant air volume fractions. The results agree very well with in vivo measurements in human lung tissue.

Asunto(s)

RESUMEN

We analyze the gradient echo signal in the presence of blood vessel networks. Both, diffusion and susceptibility effects are analytically emphasized within the Bloch-Torrey equation. Solving this equation, we present the first exact description of the local magnetization around a single vessel. This allows us to deduce the gradient echo signal of parallel vessels randomly distributed in a plane, which is valid for arbitrary mean vessel diameters in the range of physiological relevant blood volume fractions. Thus, the results are potentially relevant for gradient echo measurements of blood vessel networks with arbitrary vessel size.

Asunto(s)

RESUMEN

In this work the frequency distribution around a vessel inside a cubic voxel is investigated. Therefore, the frequency distribution is calculated in dependence on the orientation of the voxel according to the external magnetic field. The frequency distribution exhibits an interesting peak structure that cannot be explained by the established Krogh's vessel model. The results were validated with phantom measurements and in vivo measurements that agree very well with the developed theory.

Asunto(s)

RESUMEN

Myelin sheath microstructure and composition produce MR signal decay characteristics that can be used to evaluate status and outcome of demyelinating disease. We extend a recently proposed model of neuronal magnetic susceptibility, that accounts for both the structural and inherent anisotropy of the myelin sheath, by including the whole dynamic range of diffusion effects. The respective Bloch-Torrey equation for local spin dephasing is solved with a uniformly convergent perturbation expansion method, and the resulting magnetization decay is validated with a numerical solution based on a finite difference method. We show that a variation of diffusion strengths can lead to substantially different MR signal decay curves. Our results may be used to adjust or control simulations for water diffusion in neuronal structures.

Asunto(s)

RESUMEN

The Brownian motion of spins diffusing in an inhomogeneous magnetic field created by susceptibility effects is considered. Applying the correct form of the Gaussian approximation, the method allows calculating the local magnetization as well as the free induction decay for all diffusion regimes. The phase accumulated during the diffusional motion is treated by an averaging over all possible trajectories in terms of the Gaussian local phase approximation. Predictions of the Gaussian local phase approximation are compared with the Gaussian phase approximation for diffusion in a constant gradient in a slab, a cylinder, and a sphere. The Gaussian local phase approximation, thereby, shows several advantages compared to the Gaussian phase approximation: it is also valid in the static dephasing regime, predicts correctly both transverse components of the magnetization, and yields information about the local magnetization.

RESUMEN

Tissue-inherent relaxation parameters offer valuable information about the arrangement of capillaries: in an external field, capillaries act as magnetic perturbers to generate local inhomogeneous fields due to the susceptibility difference of deoxygenated blood and the surrounding tissue. These field inhomogeneities influence the free induction decay in a characteristic way, and, conversely, the above tissue parameters can be recovered by multi-parametric fits of adequate theoretical models to experimentally sampled free induction decays. In this work we study the influence of different spatial patterns of capillary positions on the free induction decay. Starting from the standard single capillary approximation (Krogh cylinder) for a symmetric array of capillaries, the free induction decay is analyzed for increasingly random capillary positions, using a previously described Gibbs point field model. The effects of diffusion are implemented with a flexible and fast random walk simulation. We find that the asymmetric form of the obtained frequency distribution is more robust against variations of capillary radii than against shifts of capillary positions, and further that, for an inclusion of diffusion effects, the single capillary approximation models the uniform alignment of capillaries in the hexagonal lattice to great accuracy. An increase in randomization of capillary positions then leads to a significant change in relaxation times. This effect, however, is found less pronounced than that of changes in the off-resonance field strengths which are controlled by the oxygen extraction fraction, thus indicating that observed changes in BOLD imaging are more likely to be attributed to changes in oxygenation than to capillary alignment.

Asunto(s)

RESUMEN

We propose a surface model of spin dephasing in lung tissue that includes both susceptibility and diffusion effects to provide a closed-form solution of the Bloch-Torrey equation on the alveolar surface. The nonlocal susceptibility effects of the model are validated against numerical simulations of spin dephasing in a realistic lung tissue geometry acquired from synchotron-based µCT data sets of mouse lung tissue, and against simulations in the well-known Wigner-Seitz model geometry. The free induction decay is obtained in dependence on microscopic tissue parameters and agrees very well with in vivo lung measurements at 1.5 Tesla to allow a quantification of the local mean alveolar radius. Our results are therefore potentially relevant for the clinical diagnosis and therapy of pulmonary diseases.

Asunto(s)

RESUMEN

We present an analytical solution of the Bloch-Torrey equation for local spin dephasing in the magnetic dipole field around a capillary and for ensembles of capillaries, and adapt this solution for the study of spin dephasing around large capillaries. In addition, we provide a rigorous mathematical derivation of the slow diffusion approximation for the spin-bearing particles that is used in this regime. We further show that, in analogy to the local magnetization, the transverse magnetization of one MR imaging voxel in the regime of static dephasing (where diffusion effects are not considered) is merely the first term of a series expansion that constitutes the signal in the slow diffusion approximation. Theoretical results are in agreement with experimental data for capillaries in rat muscle at 7T.

Asunto(s)

RESUMEN

Transverse relaxation rates for Carr-Purcell-Meiboom-Gill (CPMG) sequences increase with inter-echo time in presence of microscopic magnetic field inhomogeneities due to nuclear spin diffusion. For a weak field approximation that includes diffusion effects, the CPMG relaxation rate shift for proton diffusion around capillaries in muscle tissue can be expressed in terms of a frequency correlation function and the inter-echo time. The present work provides an analytical expression for the local relaxation rate shift that is dependent on local blood volume fraction, diffusion coefficient, capillary radius, susceptibility difference and inter-echo time. Asymptotic regions of the model are in agreement with previous modeling results of Brooks et al., Luz et al. and Ziener et al. In comparison with simulation data, the model shows an equal or better accuracy than established approximations. Also, model behavior coincides with experimental data for rat heart and skeletal muscle. The present work provides analytical tools to extract sub-voxel information about uniform capillary networks that can be used to study capillary organization or micro-circulatory remodeling.

Asunto(s)