Joint reconstruction and segmentation in undersampled 3D knee MRI combining shape knowledge and deep learning.

Objective.Task-adapted image reconstruction methods using end-to-end trainable neural networks (NNs) have been proposed to optimize reconstruction for subsequent processing tasks, such as segmentation. However, their training typically requires considerable hardware resources and thus, only relatively simple building blocks, e.g. U-Nets, are typically used, which, albeit powerful, do not integrate model-specific knowledge.Approach.In this work, we extend an end-to-end trainable task-adapted image reconstruction method for a clinically realistic reconstruction and segmentation problem of bone and cartilage in 3D knee MRI by incorporating statistical shape models (SSMs). The SSMs model the prior information and help to regularize the segmentation maps as a final post-processing step. We compare the proposed method to a simultaneous multitask learning approach for image reconstruction and segmentation (MTL) and to a complex SSMs-informed segmentation pipeline (SIS).Main results.Our experiments show that the combination of joint end-to-end training and SSMs to further regularize the segmentation maps obtained by MTL highly improves the results, especially in terms of mean and maximal surface errors. In particular, we achieve the segmentation quality of SIS and, at the same time, a substantial model reduction that yields a five-fold decimation in model parameters and a computational speedup of an order of magnitude.Significance.Remarkably, even for undersampling factors of up toR= 8, the obtained segmentation maps are of comparable quality to those obtained by SIS from ground-truth images.

Neural networks-based regularization for large-scale medical image reconstruction.

In this paper we present a generalized Deep Learning-based approach for solving ill-posed large-scale inverse problems occuring in medical image reconstruction. Recently, Deep Learning methods using iterative neural networks (NNs) and cascaded NNs have been reported to achieve state-of-the-art results with respect to various quantitative quality measures as PSNR, NRMSE and SSIM across different imaging modalities. However, the fact that these approaches employ the application of the forward and adjoint operators repeatedly in the network architecture requires the network to process the whole images or volumes at once, which for some applications is computationally infeasible. In this work, we follow a different reconstruction strategy by strictly separating the application of the NN, the regularization of the solution and the consistency with the measured data. The regularization is given in the form of an image prior obtained by the output of a previously trained NN which is used in a Tikhonov regularization framework. By doing so, more complex and sophisticated network architectures can be used for the removal of the artefacts or noise than it is usually the case in iterative NNs. Due to the large scale of the considered problems and the resulting computational complexity of the employed networks, the priors are obtained by processing the images or volumes as patches or slices. We evaluated the method for the cases of 3D cone-beam low dose CT and undersampled 2D radial cine MRI and compared it to a total variation-minimization-based reconstruction algorithm as well as to a method with regularization based on learned overcomplete dictionaries. The proposed method outperformed all the reported methods with respect to all chosen quantitative measures and further accelerates the regularization step in the reconstruction by several orders of magnitude.

Correlation between urodynamic function and 3D cat scan anatomy in neobladders: does it exist?

INTRODUCTION: We compared the functional and anatomical differences among three different orthotopic neobladders, utilizing video urodynamics and 3D CT to determine what parameters, if any, correlate to function. MATERIALS AND METHODS: Thirty-four patients were able to participate in the evaluation of their neobladder by 3D CT and video urodynamics. Three different orthotopic neobladders were identified (12 ileal, 7 ileocecal, 15 sigmoid). Multiple measurements, observations and functional data have been obtained. Statistical analysis for this study employed a linear regression test and an odds ratio calculation (using StatSoft V. 5.1). RESULTS: In comparing three different neobladders, no significant differences were noted. Looking at the entire population, the following association was statistically significant in linear correlation: the maximal capacity and the neobladder volume; the pressure at the maximal capacity and the distance from the symphysis, the pressure at maximal flow and both the distance from the symphysis and the thickness of the neobladder. The distance from the left femoral head was directly correlated with the post void residual and inversely correlated with the maximal flow. The Odds ratio calculation revealed (with significant P < 0.05) that the further the center of the neobladder is from the right femoral head, the higher risk of incontinence. CONCLUSIONS: The study seems to show no significant anatomical or functional difference among the three different types of neobladders. A possible correlation between the position of the neobladder and urinary incontinence is suggested, recognizing further study in a larger population is required.

Improved survival after live donor adult liver transplantation (LDALT) using right lobe grafts: program experience and lessons learned.

We present our program experience with 85 live donor adult liver transplantation (LDALT) procedures using right lobe grafts with five simultaneous live donor kidney transplants using different donors performed over a 6-year period. After an "early" 2-year experience of 25 LDALT procedures, program improvements in donor and recipient selection, preoperative imaging, donor and recipient surgical technique and immunosuppressive management significantly reduced operative mortality (16% vs. 3.3%, p = 0.038) and improved patient and graft 1-year survival in recipients during our "later" experience with the next 60 cases (January 2001 and March 2005; patient survival: early 70.8% vs. later 92.7%, p = 0.028; graft survival: Early 64% vs. later 91.1%, p = 0.019, respectively). Overall patient and graft survival were 82% and 80%. There was a trend for less postoperative complications (major and minor) with program experience (early 88% vs. later 66.7%; p = 0.054) but overall morbidity remained at 73.8%. Biliary complications (cholangitis, disruption, leak or stricture) were not influenced by program experience (early 32% vs. later 38%). Liver volume adjusted to 100% of standard liver volume (SLV) within 1 month post-transplant. Despite a high rate of morbidity after LDALT, excellent patient and graft survival can be achieved with program experience.

Noise-induced transport of two coupled particles.

We study the motion of two harmonically coupled particles in a sawtooth potential. The particles are subject to temporally correlated multiplicative noise. The stationary current is calculated in an expansion about the limit of rigid coupling. For two coupled particles a driving mechanism occurs which is different from the one occurring in the case of a single particle. In particular this mechanism does not need diffusion. Depending on the equilibrium distance of the particles and the coupling constant, a current reversal occurs. Possible relevance as a model for motor proteins is discussed.

In vivo MRI thermometry using a phase-sensitive sequence: preliminary experience during MRI-guided laser-induced interstitial thermotherapy of brain tumors.

The purpose of this study was the application of the proton-resonance-frequency method to monitor laser-induced interstitial thermotherapy (LITT) in a patient with an astrocytoma WHO II. A phase-sensitive two-dimensional (2D) fast low-angle shot (FLASH) sequence was used to determine the temperature-related phase shifts during LITT. Temperature maps were displayed during therapy with a temporal resolution of 20 seconds. Irradiation was discontinued as soon as the 60 to 65 degrees C isotherm reached the margin of the tumor. A contrast-enhanced MRI study performed immediately after therapy showed a good correlation of the size of an enhancing rim around the lesion with the 60 to 65 degrees C isotherm. The preliminary results of our study indicate that MRI guidance of LITT may be improved by temperature quantification based on the proton-resonance-frequency method.

Age-dependent changes in magnetization transfer contrast of white matter in the pediatric brain.

BACKGROUND AND PURPOSE: It is unknown to what extent magnetization transfer contrast (MTC) in white matter of the brain changes during myelination. The goal of this study was to measure the age-dependent changes of MTC in different regions of the pediatric brain and to evaluate their relation to T2 relaxation times. METHODS: Seventy children aged 1 week to 80 months without evidence of organic brain disease underwent MR imaging of the brain. A double-echo spin-echo (SE) sequence and an SE sequence with and without an off-resonance pulse were performed in the axial orientation. Using paired images, we calculated MTC ratios in 13 predefined regions of the brain and compared them with the T2 relaxation times measured in the same areas. Regression analysis was performed for both parameters to evaluate age dependency. RESULTS: MTC in white matter increased during myelination from a range of 13% to 19% to a range of 34% to 37%. At the same time, T2 relaxation times decreased from a range of 115 to 160 milliseconds to a range of 60 to 70 milliseconds after myelination. For both MTC and T2 relaxation times, age dependency could be expressed by a monoexponential function. CONCLUSION: A strong positive correlation exists between MTC ratios and the degree of myelination in the pediatric brain, and an inverse correlation exists between MTC and T2 relaxation times. Fast proton relaxation within macromolecules in the myelinated white matter and subsequent MT may be the most important reason for the decreasing T2 relaxation time of white matter during brain myelination.

Influence of age on the rhythm in basal and forskolin-stimulated adenylate cyclase activity of the rat heart.

In male rats of three different ages (Group I: 7 weeks, Group II: 4 months, Group III: greater than 10 months) basal and forskolin-stimulated adenylate cyclase (AC) activity were investigated in rat heart ventricles of animals sacrificed at 8 time points within 24 hours of a day. Maximum (Emax) and half-maximum (EC50) stimulation by a water-soluble forskolin (FOR; 0.1-100 mumol/l) were determined from dose-response curves. In young rats (Group I) significant rhythmicity was found in basal and FOR-stimulated AC activity, which was successively reduced or abolished with age by a reduction in amplitude. Mean basal AC activity increased about 2-fold from Group I to Group II/III. Mean maximum in FOR stimulation was about 12-fold in Group I and about 7-fold in Group II and III. EC50-values displayed neither a significant rhythmicity in either group nor an age-dependency. It is concluded that the stimulatory potency of the catalytic subunit of the AC is reduced with age.