Search

Background Obstructive hydrocephalus in adulthood can be caused by stenosis in the aqueductal area. Chronic changes lead to a dilatation of the lateral ventricles and ballooning of infratentorial recesses. In rare cases a rupture of the floor of the third ventricle (so-called spontaneous ventriculostomy) has been described in the literature. Case presentation: We present two cases of chronic obstructive hydrocephalus due to aqueductal stenosis in adult patients. Magnetic resonance imaging included phase-contrast-imaging and revealed significant flow through the floor of the third ventricle in keeping with spontaneous ventriculostomy. In addition to other typical changes associated with chronic hydrocephalus, a distinct flattening of the tectal plate could be identified in one case. Conclusion We present two cases of spontaneous ventriculostomy in patients with chronic hydrocephalus. To our knowledge, flattening of the tectal plate has not yet been described in the literature and may be caused by continuous cerebrospinal fluid-pulsation.

SummaryThe point centromere of budding yeast specifies assembly of the large multi-subunit kinetochore complex. By direct attachment to the mitotic spindle, kinetochores couple the forces of microtubule dynamics to power chromatid segregation at mitosis. Kinetochores share a conserved architecture comprising the centromere-associated inner kinetochore CCAN (constitutive centromere-associated network) complex and the microtubule-binding outer kinetochore KMN network. The budding yeast inner kinetochore additionally includes the centromere-binding CBF1 and CBF3 complexes. Here, we reconstituted the complete yeast inner kinetochore complex assembled onto the centromere-specific CENP-A nucleosome (CENP-ANuc) and determined its structure using cryo-EM. This revealed a central CENP-ANuc, wrapped by only one turn of DNA, and harboring extensively unwrapped DNA ends. These free DNA duplexes function as binding sites for two CCAN protomers, one of which entraps DNA topologically and is positioned precisely on the centromere by the sequence-specific DNA-binding complex CBF1. The CCAN protomers are connected through CBF3 to form an arch-like configuration, binding 150 bp of DNA. We also define a structural model for a CENP-ANuc-pathway to the outer kinetochore involving only CENP-QU. This study presents a framework for understanding the basis of complete inner kinetochore assembly onto a point centromere, and how it organizes the outer kinetochore for robust chromosome attachment to the mitotic spindle.

To assess the impact of prosthesis choice and aortic valve calcifications on the occurrence of conduction disturbances after transcatheter aortic valve implantation (TAVI).We retrospectively analyzed the preoperative clinical characteristics, electrocardiograms, contrast-enhanced multidetector computed tomography scans and procedural strategies of patients who underwent TAVI in our center between January 2012 and June 2017. Quantification of calcium volume was performed for each aortic cusp above (aortic valve) and below (left ventricular outflow tract, LVOT) the basal plane. Multivariate analysis was performed to evaluate risk factors for the onset of new bundle branch block (BBB), transient and permanent atrioventricular block (tAVB, pAVB).A total of 569 patients were included in the study. Six different prostheses were implanted (Edwards Sapien XT,LVOT calcification beneath the non-coronary cusp, baseline left anterior hemiblock, right BBB, balloon post-dilatation, prior valvuloplasty and oversizing are independently associated with postprocedural conduction disturbances after TAVI. Use of a self-expandable prosthesis may show a lower incidence of AVB, if applied in lower calcified aortic valves.

Aims Peripartum cardiomyopathy (PPCM) establishes late in pregnancy or in the first postpartum months. Many patients recover well within the first year, but long-term outcome studies on morbidity and mortality are rare. Here, we present 5-year follow-up data of a German PPCM cohort. Methods and results Five-year follow-up data were available for 66 PPCM patients (mean age 34 ± 5 years) with a mean left ventricular ejection fraction (LVEF) of 26 ± 9% at diagnosis. Ninety-eight percent initially received standard heart failure therapy (beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, and/or mineralocorticoid receptor antagonists), and 86% were additionally treated with dopamine D2 receptor agonists (mainly bromocriptine) and anticoagulation. After 1 year, mean LVEF had improved to 50 ± 11% (n = 48) and further increased to 54 ± 7% at 5-year follow-up with 72% of patients having achieved full cardiac recovery (LVEF >50%). At 5-year follow-up, only three patients (5%) displayed no recovery, of whom one had died. However, 20% had arterial hypertension and 17% arrhythmias, including paroxysmal supraventricular tachycardia, ventricular tachycardia, or ventricular fibrillation. Moreover, 70% were still on at least one heart failure drug. Subsequent pregnancy occurred in 16 patients with two abortions and 14 uneventful pregnancies. Mean LVEF was 55 ± 7% at 5-year follow-up in these patients. Conclusion Our PPCM collective treated with standard therapy for heart failure, dopamine D2 receptor agonists, and anticoagulation displays a high and stable long-term recovery rate with low mortality at 5-year follow-up. However, long-term use of cardiovascular medication, persisting or de novo hypertension and arrhythmias were frequent.

A demonstrative Fischer-Tropsch fuel without any cost intensive post-processing treatment has been investigated for its application potential as a synthetic blending component with conventional petroleum-derived aviation fuels. As a first step, the focus of the analysis was purely on combustion related properties. The Fischer-Tropsch fuel was obtained via a specific Power-to-Liquid Fischer-Tropsch process, developed by Ineratec. Whereas the already approved Fischer-Tropsch–SPK process (ASTM D7566 Annex A1) utilizes hydrotreatment and is applied in large-scale plants, the herein presented plant features a unique and compact container-scale set-up, with no further downstream hydrotreatment, which allows for a significant reduction of production time and costs. Main objective of this paper is to provide the fuel producer with fast feedback to find the minimum complexity of fuel processing technology to achieve a synthetic blending component for aviation fuels directly from a container plant. As a first step in the ongoing process, the combustion properties of the non-hydroprocessed Fischer-Tropsch fuels are assessed regarding their suitability for aviation purposes. Fuel characterization was carried out regarding the physiochemical properties of the fuels and their chemical composition to monitor selected “fit-for-purpose” properties for aviation with regard to combustion properties. Additional combustion experiments were conducted in a high-temperature flow-reactor with coupled molecular beam mass spectrometer (MBMS) for two stoichiometries to map lean and rich combustion (Φ = 0.8 and 1.2), allowing quantitative access to the chemical reaction species formed within the combustion. The general combustion chemistry and reaction temperature regime was found similar to Jet A-1 and pure n-alkane decane. This indicates the dominant species for the observed combustion process are aliphatic hydrocarbons. The detailed evaluation of relevant intermediates allows for an observation on typical soot precursors (e.g. benzene, naphthalene) in the combustion process and enables the estimation on the pollutant reduction potential of the Fischer-Tropsch fuel when used as blending component to Jet A-1. Blending analysis has been performed utilizing the data from the CRC world fuel survey to evaluate the range of blending ratios of the Fischer-Tropsch fuel with conventional jet fuels determined by identified limiting factors. The presented evaluations demonstrate the potential of the Fischer-Tropsch fuel as a blending component with conventional jet fuels considering the combustion behavior only.

Background: The vascular access in transcatheter aortic valve replacement (TAVR) was initially dominated by a surgical approach. Meanwhile, percutaneous closure systems became a well-established alternative. The aim of this study was to compare the clinical outcome between the two approaches. Methods: In this retrospective study, we observed 787 patients undergoing a TAVR-Procedure between 2013 and 2019. Of those, 338 patients were treated with surgical access and 449 with the Perclose ProGlide™-System (Abbott, Chicago, IL, USA). According to the Bleeding Academic Research Consortium (BARC) and Valve Academic Research Consortium (VARC) criteria, the primary combined endpoints were defined. Results: Overall hospital mortality was 2.8% with no significant difference between surgical (3.8%) and percutaneous (2.2%) access (p = 0.182). Major vascular complications or bleeding defined as the primary combined endpoint was not significantly different in either group (Surgical group 5.3%, ProGlide group 5.1%, p = 0.899). In the ProGlide group, women with pre-existing peripheral artery disease (PAD) were significantly more often affected by a vascular complication (p = 0.001 for female sex and p = 0.03 for PAD). Conclusions: We were able to show that the use of both accesses is safe. However, the surgical access route should also be considered in case of peripheral artery disease.

Background Acute respiratory distress syndrome (ARDS) is an important disease with a high incidence among patients admitted to intensive care units. Over the last decades, the survival of critically ill patients has improved; however, cognitive deficits are among the long-term sequelae. We hypothesize that acute lung injury leads to upregulation of cerebral cytokine synthesis. Methods After approval of the institutional and animal care committee, 20 male pigs were randomized to one of three groups: (1) Lung injury by oleic acid injection (OAI), (2) ventilation only (CTR) or (3) untreated. We compared neuronal numbers, proportion of neurons with markers for apoptosis, activation state of Iba-1 stained microglia cells and cerebral mRNA levels of different cytokines between the groups 18 hours after onset of lung injury. Results We found an increase in hippocampal TNFalpha (p p p p p Conclusion Hippocampal cytokine transcription increases within 18 hours after the induction of acute lung injury with histological evidence of neuronal damage. It remains to be elucidated if increased cytokine mRNA synthesis plays a role in the cognitive decline observed in survivors of ARDS.

Background Calcification of native aortic valve is a well known variable causing paravalvular leakage (PVL) following transcatheter aortic valve implantation (TAVI). However, there is currently no evidence to support the fact that patients with high calcium load could be better treated with surgical aortic valve replacement (SAVR). Purpose To assess the utility of preoperative assessment of valvular calcium load through computed tomography in patients affected by severe aortic valve stenosis undergoing SAVR or TAVI. Methods Between June 2016 and June 2018, 109 candidates for isolated SAVR through minimal invasive access underwent preoperative contrast enhanced multidetector computed tomography (MDCT) for the assessment of valve and aortic calcifications. Calcium load was quantitatively measured using a dedicated software in three regions on interest (aortic valve [AV], left ventricular outflow tract [LVOT] and device landing zone [DLZ], which is the sum of the earlier 2). Clinical, echocardiographic, and MDCT variables were collected and compared to a sample population of 107 patients that underwent TAVI (87 transfemoral, 20 transapical) for native aortic valve stenosis in the same period of time, in the same institution. A univariate and multivariate logistic regression analysis were performed on the whole study population to assess risk factors for the onset of postoperative PVL (any grade, defined as ≥ mild) at discharge. Results The two study groups were significantly different in terms of age (71.9±5 in SAVR; 81.5±6 in TAVI), gender (36% female in SAVR vs 51% in TAVI), Euroscore II (1.9%±0.8 in SAVR; 4.8%±2.7 in TAVI), annulus perimeter (79.5mm±8.2 in SAVR; 61.8mm±30.5 in TAVI), baseline ejection fraction (57%±8 in SAVR; 51%±12 in TAVI) and severe pulmonary hypertension (2%±13 in SAVR; 26%±44 in TAVI). Calcium load was not different between groups (DLZ 1066 mm3±716 vs 955mm3±639; total calcium in AV 987 mm3±678 vs 879 mm3±601; total calcium in LVOT 78 mm3±130 vs 77 mm3±100). 30-days-mortality was 1.8% in SAVR and 5.6% in TAVI group (p=0.17). At discharge, incidence of all grades PVL was 5.5% in SAVR (0.9% trace, 3.6% mild, 0.9% moderate) and 41% in TAVI group (12.1% trace, 25.2% mild, 3.7% moderate) (p Conclusions Aortic valve calcifications are a risk factor for the onset of PVL for both TAVI and SAVR. Nevertheless, the risk increases considerably with the use of TAVI. A deeper anatomical analysis of preoperative MDCT could improve the treatment selection and the outcome of patients affected by aortic valve stenosis. Funding Acknowledgement Type of funding source: None

We propose a sparse reconstruction framework (aNETT) for solving inverse problems. Opposed to existing sparse reconstruction techniques that are based on linear sparsifying transforms, we train an autoencoder network $D \circ E$ with $E$ acting as a nonlinear sparsifying transform and minimize a Tikhonov functional with learned regularizer formed by the $\ell^q$-norm of the encoder coefficients and a penalty for the distance to the data manifold. We propose a strategy for training an autoencoder based on a sample set of the underlying image class such that the autoencoder is independent of the forward operator and is subsequently adapted to the specific forward model. Numerical results are presented for sparse view CT, which clearly demonstrate the feasibility, robustness and the improved generalization capability and stability of aNETT over post-processing networks., Comment: The original proceeding is part of the ISBI 2020 and only contains 4 pages due to page restrictions

The development of fast and accurate image reconstruction algorithms is a central aspect of computed tomography. In this paper, we investigate this issue for the sparse data problem in photoacoustic tomography (PAT). We develop a direct and highly efficient reconstruction algorithm based on deep learning. In our approach, image reconstruction is performed with a deep convolutional neural network (CNN), whose weights are adjusted prior to the actual image reconstruction based on a set of training data. The proposed reconstruction approach can be interpreted as a network that uses the PAT filtered backprojection algorithm for the first layer, followed by the U-net architecture for the remaining layers. Actual image reconstruction with deep learning consists in one evaluation of the trained CNN, which does not require time-consuming solution of the forward and adjoint problems. At the same time, our numerical results demonstrate that the proposed deep learning approach reconstructs images with a quality comparable to state of the art iterative approaches for PAT from sparse data.

The development of fast and accurate image reconstruction algorithms is a central aspect of computed tomography. In this paper we address this issue for photoacoustic computed tomography in circular geometry. We investigate the Galerkin least squares method for that purpose. For approximating the function to be recovered we use subspaces of translation invariant spaces generated by a single Funktion. This includes many systems that have previously been employed in PAT such as generalized Kaiser-Bessel basis functions or the natural pixel basis. By exploiting an isometry property of the forward problem we are able to efficiently set up the Galerkin equation for a wide class of generating functions and Devise efficient algorithms for its solution. We establish a convergence analysis and present numerical simulations that demonstrate the efficiency and accuracy of the derived algorithm., 31 pages, 5 figures

Unter transsektoraler Versorgung von Patienten mit Herzinsuffizienz versteht man das koordinierte optimale Zusammenwirken verschiedener Leistungserbringer im stationaren und ambulanten Bereich. Hierdurch kann die Mortalitat und Rehospitalisierungsrate signifikant reduziert werden bzw. die Leistungsfahigkeit und Lebensqualitat nicht zuletzt durch eine Starkung der Selbstverantwortung und -bestimmung dieser Patienten gesteigert werden. Voraussetzung ist allerdings ein supervidiertes, spezialisiertes Pflegepersonal, das die Vernetzung mehrerer Versorgungsebenen mit dem heimischen Umfeld des Patienten sicherstellt. Entscheidend fur den Erfolg eines solchen multidisziplinaren Managements ist weiterhin, das Disease-Management-Programm den individuellen regionalen Anforderungen, Bedurfnissen und nicht zuletzt limitierten finanziellen Ressourcen anzupassen.

Background Left ventricular noncompaction cardiomyopathy (LVNC) is characterized by an increased number of LV trabeculations with deep intertrabecular recesses. This abnormality is associated with heart failure, arrhythmias and arterial embolic events (AE). At present, it is unknown if AE is mainly due to blood stasis within the intertrabecular recesses, reduced LV ejection fraction or concomitant atrial fibrillation. LVNC is usually diagnosed by echocardiography but cardiac magnetic resonance imaging (CMRI) has evolved as an alternative method. This study assessed the prognostic value of CMRI for arterial embolic events in patients (pts) with LVNC. Methods 34 consecutive pts (19m, 15f, age 53±16) with LVNC underwent cine and contrast-enhanced CMRI with a 1.5 T scanner. LV diameter, volume, ejection fraction, and ratio of noncompacted to compacted myocardium (NC/C) were determined, and in 32 pts presence and localization of late gadolinium enhancement (LGE) was assessed. Clinical and CMRI findings were compared in pts with and without LV thrombus and/or AE. Results Overall, 20 pts (59%) were in heart failure NYHA III or IV, 14 (41%) had left bundle branch block (LBBB), 7 (21%) paroxysmal atrial fibrillation and 6 (19%) ventricular tachycardia (VT). By CMRI, LV diameter in end-diastole (66±8 mm), end-systole (53±10 mm), end-diastolic (229±69 ml) and end-systolic volume (150±68 ml) were enlarged and ejection fraction (36±14%) was reduced. The NC/C ratio was 3.2±1.4 in end-diastole and 2.6±1.4 in end-systole. One pt had right ventricular involvement with a thrombus. LGE was seen in 9/32 pts (28%) in the compacted myocardial layer (n=6), in the noncompacted trabecular layer (n=6) and within the papillary muscles (n=3). LGE was present in 3 areas in 1 and in 2 areas in 4 pts. In 3 pts (9%) a thrombus was seen within the trabecular layer which resolved under anticoagulation, and 6 additional pts (18%) without detectable thrombus experienced AE (transient ischemic attack n=1, stroke n=5). Thrombus and/or AE were not associated with age, sex, NYHA class, larger left atrial or LV diameter, LV volume, LBBB or documented VT. Atrial fibrillation (2/9 vs 5/25 pts, p=ns), LV ejection fraction (33±13% vs 38±15%, p=ns) and the NC/C ratio in end-diastole (median 3.2 vs 3) or end-systole (both median 2.6, p=ns) were similar. Thrombus and/or AE occurred mainly in pts with LGE (6/9 vs 2/23 pts, p=0.002). Conclusion In LVNC, evaluation by CMRI and demonstration of LGE in the compacted or noncompacted myocardium identifies patients at high risk for thrombus formation and/or arterial embolic events, warranting anticoagulation.

Aortic valve calcification is supposed to be a possible cause of embolic stroke or subclinical valve thrombosis after transcatheter aortic valve replacement (TAVR). We aimed to assess the role of aortic valve calcification in the occurrence of in-hospital clinical complications and survival after TAVR.We retrospectively analyzed preoperative contrast-enhanced multidetector computed tomography scans of patients who underwent TAVR on the native aortic valve in our center. Calcium volume was calculated for each aortic cusp, above and below the aortic annulus. Outcomes were recorded according to VARC-2 criteria.Overall, 581 patients were included in the study (SapienXT = 192; Sapien3 = 228; CoreValve/EvolutR = 45; Engager = 5; Acurate = 111). Median survival was 4.98 years (interquartile range 4.41-5.54). Logistic regression identified calcium load beneath the right coronary cusp in left ventricular outflow tract (LVOT) as significantly associated with stroke (odds ratio [OR] 1.2; 95% confidence interval [CI] 1.03-1.3; p = 0.0019) and in-hospital mortality (OR 1.1; 95% CI 1.004-1.2; p = 0.04), whereas total calcium volume of the LVOT was associated with both in-hospital and 30 day-mortality (OR 1.2; 95% CI 1.01-1.4; p = 0.03, and OR 1.2; 95% CI 1.02-1.43; p = 0.029, respectively). Cox regression identified total calcium of LVOT (hazard ratio [HR] 1.18; 95% CI 1.02-1.38; p = 0.026), male sex (HR 1.88; 95% CI 1.06-3.32; p = 0.031), baseline creatinine clearance (HR 0.96; 95% CI 0.93-0.98; p 0.001), and baseline severe aortic regurgitation (HR 7.48; 95% CI 2.76-20.26; p 0.001) as risk factors associated with lower survival.LVOT calcification is associated with increased risk of peri-procedural stroke and mortality as well as shorter long-term survival.

Filtered backprojection (FBP) is an efficient and popular class of tomographic image reconstruction methods. In photoacoustic tomography, these algorithms are based on theoretically exact analytic inversion formulas which results in accurate reconstructions. However, photoacoustic measurement data are often incomplete (limited detection view and sparse sampling), which results in artefacts in the images reconstructed with FBP. In addition to that, properties such as directivity of the acoustic detectors are not accounted for in standard FBP, which affects the reconstruction quality, too. To account for these issues, in this papers we propose to improve FBP algorithms based on machine learning techniques. In the proposed method, we include additional weight factors in the FBP, that are optimized on a set of incomplete data and the corresponding ground truth photoacoustic source. Numerical tests show that the learned FBP improves the reconstruction quality compared to the standard FBP., Comment: This paper is a proceedings to our presentation (Paper 1087837) at the Photons Plus Ultrasound: Imaging and Sensing, (within the SPIE Photonics West), Poster Sunday, February 03, 2019

We discuss several methods for image reconstruction in compressed sensing photoacoustic tomography (CS-PAT). In particular, we apply the deep learning method of [H. Li, J. Schwab, S. Antholzer, and M. Haltmeier. NETT: Solving Inverse Problems with Deep Neural Networks (2018), arXiv:1803.00092], which is based on a learned regularizer, for the first time to the CS-PAT problem. We propose a network architecture and training strategy for the NETT that we expect to be useful for other inverse problems as well. All algorithms are compared and evaluated on simulated data, and validated using experimental data for two different types of phantoms. The results on the one the hand indicate great potential of deep learning methods, and on the other hand show that significant future work is required to improve their performance on real-word data.

We propose a sparse reconstruction framework for solving inverse problems. Opposed to existing sparse regularization techniques that are based on frame representations, we train an encoder-decoder network by including an $\ell^1$-penalty. We demonstrate that the trained decoder network allows sparse signal reconstruction using thresholded encoded coefficients without losing much quality of the original image. Using the sparse synthesis prior, we propose minimizing the $\ell^1$-Tikhonov functional, which is the sum of a data fitting term and the $\ell^1$-norm of the synthesis coefficients, and show that it provides a regularization method., Comment: Presented at the SAMPTA 2019 conference

Applying standard algorithms to sparse data problems in photoacoustic tomography (PAT) yields low-quality images containing severe under-sampling artifacts. To some extent, these artifacts can be reduced by iterative image reconstruction algorithms which allow to include prior knowledge such as smoothness, total variation (TV) or sparsity constraints. These algorithms tend to be time consuming as the forward and adjoint problems have to be solved repeatedly. Further, iterative algorithms have additional drawbacks. For example, the reconstruction quality strongly depends on a-priori model assumptions about the objects to be recovered, which are often not strictly satisfied in practical applications. To overcome these issues, in this paper, we develop direct and efficient reconstruction algorithms based on deep learning. As opposed to iterative algorithms, we apply a convolutional neural network, whose parameters are trained before the reconstruction process based on a set of training data. For actual image reconstruction, a single evaluation of the trained network yields the desired result. Our presented numerical results (using two different network architectures) demonstrate that the proposed deep learning approach reconstructs images with a quality comparable to state of the art iterative reconstruction methods., Comment: This paper is a proceeding to our presentation (10494-168) at the Photons Plus Ultrasound: Imaging and Sensing conference (within the SPIE Photonics West), Poster Monday, January 29, 2018

We develop a data-driven regularization method for the severely ill-posed problem of photoacoustic image reconstruction from limited view data. Our approach is based on the regularizing networks that have been recently introduced and analyzed in [J. Schwab, S. Antholzer, and M. Haltmeier. Big in Japan: Regularizing networks for solving inverse problems (2018), arXiv:1812.00965] and consists of two steps. In the first step, an intermediate reconstruction is performed by applying truncated singular value decomposition (SVD). In order to prevent noise amplification, only coefficients corresponding to sufficiently large singular values are used, whereas the remaining coefficients are set zero. In a second step, a trained deep neural network is applied to recover the truncated SVD coefficients. Numerical results are presented demonstrating that the proposed data driven estimation of the truncated singular values significantly improves the pure truncated SVD reconstruction. We point out that proposed reconstruction framework can straightforwardly be applied to other inverse problems, where the SVD is either known analytically or can be computed numerically., Comment: This paper is a proceedings to our presentation (Paper 10878-111) at the Photons Plus Ultrasound: Imaging and Sensing, (within the SPIE Photonics West), Poster Sunday, February 03, 2019

The treatment of severe symptomatic aortic valve stenosis by conventional aortic valve replacement (AVR) or by transcatheter aortic valve implantation (TAVI) has a good perinterventional prognosis even for patients of advanced age. Having a heart team select the best management strategies based on current guidelines for each individual patient is essential for success. Especially in elderly and increasingly multimorbid patients with sometimes severe preconditions, the detection of functional deficits is relevant not only for the mortality but also for perioperative and postoperative complications as well as the functional outcome. Various methods of geriatric assessment are important supplements to standard risk scores. The aim is to implement targeted interventions to minimize the risk factors and to improve the prognosis for elderly patients. The aim of this article is to provide an overview of the current therapy options for aortic valve replacement and to summarize current aspects of treatment options for elderly patients.

Recently, a large number of efficient deep learning methods for solving inverse problems have been developed and show outstanding numerical performance. For these deep learning methods, however, a solid theoretical foundation in the form of reconstruction guarantees is missing. In contrast, for classical reconstruction methods, such as convex variational and frame-based regularization, theoretical convergence and convergence rate results are well established. In this paper, we introduce deep synthesis networks for regularizing inverse problems (DESYRE) using neural networks as nonlinear synthesis operator bridging the gap between these two worlds. The proposed method allows to exploit the deep learning benefits of being well adjustable to available training data and on the other hand comes with a solid mathematical foundation. We present a complete convergence analysis with convergence rates for the proposed deep synthesis regularization. We present a strategy for constructing a synthesis network as part of an analysis–synthesis sequence together with an appropriate training strategy. Numerical results show the plausibility of our approach.

Aims Despite recent advances in the treatment of chronic heart failure (HF), mortality and hospitalizations still remain high. Additional therapies to improve mortality and morbidity are urgently needed. The efficacy of cardiac glycosides - although regularly used for HF treatment - remains unclear. DIGIT-HF was designed to demonstrate that digitoxin on top of standard of care treatment improves mortality and morbidity in patients with HF and a reduced ejection fraction (HFrEF). Methods Patients with chronic HF, New York Heart Association (NYHA) functional class III-IV and left ventricular ejection fraction (LVEF) ≤ 40%, or patients in NYHA functional class II and LVEF ≤ 30% are randomized 1:1 in a double-blind fashion to treatment with digitoxin (target serum concentration 8-18 ng/mL) or matching placebo. Randomization is stratified by centre, sex, NYHA functional class (II, III, or IV), atrial fibrillation, and treatment with cardiac glycosides at baseline. A total of 2190 eligible patients will be included in this clinical trial (1095 per group). All patients receive standard of care treatment recommended by expert guidelines upon discretion of the treating physician. The primary outcome is a composite of all-cause mortality or hospital admission for worsening HF (whatever occurs first). Key secondary endpoints are all-cause mortality, hospital admission for worsening HF, and recurrent hospital admission for worsening HF. Conclusion The DIGIT-HF trial will provide important evidence, whether the cardiac glycoside digitoxin reduces the risk for all-cause mortality and/or hospital admission for worsening HF in patients with advanced chronic HFrEF on top of standard of care treatment.

We investigate compressed sensing (CS) techniques for reducing the number of measurements in photoacoustic tomography (PAT). High resolution imaging from CS data requires particular image reconstruction algorithms. The most established reconstruction techniques for that purpose use sparsity and $\ell^{1}$ -minimization. Recently, deep learning appeared as a new paradigm for CS and other inverse problems. In this paper, we compare a recently invented joint $\ell^{1}$ -minimization algorithm with two deep learning methods, namely a residual network and an approximate nullspace network. We present numerical results showing that all developed techniques perform well for deterministic sparse measurements as well as for random Bernoulli measurements. For the deterministic sampling, deep learning shows more accurate results, whereas for Bernoulli measurements the $\ell^{1}$ -minimization algorithm performs best. Comparing the implemented deep learning approaches, we show that the nullspace network uniformly outperforms the residual network in terms of the mean squared error (MSE).

AIMS To assess the contribution of aortic valve calcification to the occurrence of transient or permanent atrioventricular block (AVB) and the need for permanent pacemaker implantation (PPI) after transcatheter aortic valve implantation (TAVI) in a large single-centre cohort. METHODS AND RESULTS We retrospectively analysed pre-operative contrast-enhanced multidetector computed tomography scans of patients who underwent TAVI in our centre between 2012 and 2016. Calcium volume was calculated for each aortic cusp above (aortic valve), and below [left ventricular outflow tract (LVOT)] the basal plane. Clinical and procedural data as well as pre-operative electrocardiograms were evaluated. Multivariate analysis was performed to evaluate risk factors for transient and permanent AVB. A total of 342 patients receiving a balloon-expandable prosthesis were included in the study. Overall incidence of transient and permanent AVB was 4% (n = 14) and 7.6% (n = 26), respectively. On logistic regression analysis, baseline right bundle branch block [odds ratio (OR) 7.36, 95% confidence interval (CI) 2.6-20.6; P

Background Right ventricular (RV) dysfunction predicts adverse outcome in peripartum cardiomyopathy (PPCM). We recently demonstrated beneficial effects associated with the prolactin release inhibitor bromocriptine at different doses when added to standard heart failure therapy in PPCM. Here, we evaluated for the first time the therapeutic potential of bromocriptine particularly in PPCM patients with RV involvement. Methods In this study, 40 patients with PPCM were included, of whom 24 patients had reduced RV ejection fraction (RVEF

Recovering a function or high-dimensional parameter vector from indirect measurements is a central task in various scientific areas. Several methods for solving such inverse problems are well developed and well understood. Recently, novel algorithms using deep learning and neural networks for inverse problems appeared. While still in their infancy, these techniques show astonishing performance for applications like low-dose CT or various sparse data problems. However, there are few theoretical results for deep learning in inverse problems. In this paper, we establish a complete convergence analysis for the proposed NETT (network Tikhonov) approach to inverse problems. NETT considers nearly data-consistent solutions having small value of a regularizer defined by a trained neural network. We derive well-posedness results and quantitative error estimates, and propose a possible strategy for training the regularizer. Our theoretical results and framework are different from any previous work using neural networks for solving inverse problems. A possible data driven regularizer is proposed. Numerical results are presented for a tomographic sparse data problem, which demonstrate good performance of NETT even for unknowns of different type from the training data. To derive the convergence and convergence rates results we introduce a new framework based on the absolute Bregman distance generalizing the standard Bregman distance from the convex to the non-convex case.