Explore the vast range of titles available.

Scientific progress depends on the ability of independent researchers to scrutinize the results of a research study, to reproduce the study's main results using its materials, and to build on them in future studies (https://www.nature.com/nature-research/editorial-policies/reporting-standards). Nuances in the computer code may have marked effects on the training and evaluation of results4, potentially leading to unintended consequences5. [...]transparency in the form of the actual computer code used to train a model and arrive at its final set of parameters is essential for research reproducibility. The many software dependencies of large-scale machine learning applications require appropriate control of the software environment, which can be achieved through package managers including Conda, as well as container and virtualization systems, including Code Ocean, Gigantum, Colaboratory and Docker. Sharing the fitted model (architecture along with learned parameters) should be simple aside from privacy concerns that the model may reveal sensitive information about the set of patients used to train it.

Globally, there is a huge unmet need for effective treatments for neurodegenerative diseases. The complexity of the molecular mechanisms underlying neuronal degeneration and the heterogeneity of the patient population present massive challenges to the development of early diagnostic tools and effective treatments for these diseases. Machine learning, a subfield of artificial intelligence, is enabling scientists, clinicians and patients to address some of these challenges. In this Review, we discuss how machine learning can aid early diagnosis and interpretation of medical images as well as the discovery and development of new therapies. A unifying theme of the different applications of machine learning is the integration of multiple high-dimensional sources of data, which all provide a different view on disease, and the automated derivation of actionable insights.In this Review, the authors describe the latest developments in the use of machine learning to interrogate neurodegenerative disease-related datasets. They discuss applications of machine learning to diagnosis, prognosis and therapeutic development, and the challenges involved in analysing health-care data.

Inflammatory depression is a treatment-resistant subtype of depression. A causal role of the gut microbiota as a source of low-grade inflammation remains unclear. Here, as part of an observational trial, we first analyze the gut microbiota composition in the stool, inflammatory factors and short-chain fatty acids (SCFAs) in plasma, and inflammatory and permeability markers in the intestinal mucosa of patients with inflammatory depression (ChiCTR1900025175). Gut microbiota of patients with inflammatory depression exhibits higher Bacteroides and lower Clostridium , with an increase in SCFA-producing species with abnormal butanoate metabolism. We then perform fecal microbiota transplantation (FMT) and probiotic supplementation in animal experiments to determine the causal role of the gut microbiota in inflammatory depression. After FMT, the gut microbiota of the inflammatory depression group shows increased peripheral and central inflammatory factors and intestinal mucosal permeability in recipient mice with depressive and anxiety-like behaviors. Clostridium butyricum administration normalizes the gut microbiota, decreases inflammatory factors, and displays antidepressant-like effects in a mouse model of inflammatory depression. These findings suggest that inflammatory processes derived from the gut microbiota can be involved in neuroinflammation of inflammatory depression. Inflammatory depression is a treatment-resistant subtype of depression. Here, the authors show that patients with inflammatory depression exhibit a disrupted microbiota, which upon FMT in mice leads to increased peripheral and central inflammatory factors, intestinal mucosal permeability, and depressive and anxiety-like behaviors. Probiotic administration ameliorates the disease phenotype.

Artificial intelligence (AI) systems have been shown to help dermatologists diagnose melanoma more accurately, however they lack transparency, hindering user acceptance. Explainable AI (XAI) methods can help to increase transparency, yet often lack precise, domain-specific explanations. Moreover, the impact of XAI methods on dermatologists’ decisions has not yet been evaluated. Building upon previous research, we introduce an XAI system that provides precise and domain-specific explanations alongside its differential diagnoses of melanomas and nevi. Through a three-phase study, we assess its impact on dermatologists’ diagnostic accuracy, diagnostic confidence, and trust in the XAI-support. Our results show strong alignment between XAI and dermatologist explanations. We also show that dermatologists’ confidence in their diagnoses, and their trust in the support system significantly increase with XAI compared to conventional AI. This study highlights dermatologists’ willingness to adopt such XAI systems, promoting future use in the clinic. Artificial intelligence has become popular as a cancer classification tool, but there is distrust of such systems due to their lack of transparency. Here, the authors develop an explainable AI system which produces text- and region-based explanations alongside its classifications which was assessed using clinicians’ diagnostic accuracy, diagnostic confidence, and their trust in the system.

Diffuse myocardial fibrosis resulting from the excessive deposition of collagen fibres through the entire myocardium is encountered in a number of chronic cardiac diseases. This lesion results from alterations in the regulation of fibrillary collagen turnover by fibroblasts, facilitating the excessive deposition of type I and type III collagen fibres within the myocardial interstitium and around intramyocardial vessels. The available evidence suggests that, beyond the extent of fibrous deposits, collagen composition and the physicochemical properties of the fibres are also relevant in the detrimental effects of diffuse myocardial fibrosis on cardiac function and clinical outcomes in patients with heart failure. In this regard, findings from the past 20 years suggest that various clinicopathological phenotypes of diffuse myocardial fibrosis exist in patients with heart failure. In this Review, we summarize the current knowledge on the mechanisms and detrimental consequences of diffuse myocardial fibrosis in heart failure. Furthermore, we discuss the validity and usefulness of available imaging techniques and circulating biomarkers to assess the clinicopathological variation in this lesion and to track its clinical evolution. Finally, we highlight the currently available and potential future therapeutic strategies aimed at personalizing the prevention and reversal of diffuse myocardial fibrosis in patients with heart failure.In this Review, Díez and colleagues summarize the mechanisms of diffuse myocardial fibrosis in heart failure, discuss imaging techniques and circulating biomarkers to characterize the variability of this lesion in patients, and highlight the available and potential future therapeutic strategies for personalizing the prevention and reversal of diffuse myocardial fibrosis.

Labeled medical imaging data is scarce and expensive to generate. To achieve generalizable deep learning models large amounts of data are needed. Standard data augmentation is a method to increase generalizability and is routinely performed. Generative adversarial networks offer a novel method for data augmentation. We evaluate the use of CycleGAN for data augmentation in CT segmentation tasks. Using a large image database we trained a CycleGAN to transform contrast CT images into non-contrast images. We then used the trained CycleGAN to augment our training using these synthetic non-contrast images. We compared the segmentation performance of a U-Net trained on the original dataset compared to a U-Net trained on the combined dataset of original data and synthetic non-contrast images. We further evaluated the U-Net segmentation performance on two separate datasets: The original contrast CT dataset on which segmentations were created and a second dataset from a different hospital containing only non-contrast CTs. We refer to these 2 separate datasets as the in-distribution and out-of-distribution datasets, respectively. We show that in several CT segmentation tasks performance is improved significantly, especially in out-of-distribution (noncontrast CT) data. For example, when training the model with standard augmentation techniques, performance of segmentation of the kidneys on out-of-distribution non-contrast images was dramatically lower than for in-distribution data (Dice score of 0.09 vs. 0.94 for out-of-distribution vs. in-distribution data, respectively, p < 0.001). When the kidney model was trained with CycleGAN augmentation techniques, the out-of-distribution (non-contrast) performance increased dramatically (from a Dice score of 0.09 to 0.66, p < 0.001). Improvements for the liver and spleen were smaller, from 0.86 to 0.89 and 0.65 to 0.69, respectively. We believe this method will be valuable to medical imaging researchers to reduce manual segmentation effort and cost in CT imaging.

There is a limited access to liver transplantation, however, many organs are discarded based on subjective assessment only. Here we report the VITTAL clinical trial (ClinicalTrials.gov number NCT02740608) outcomes, using normothermic machine perfusion (NMP) to objectively assess livers discarded by all UK centres meeting specific high-risk criteria. Thirty-one livers were enroled and assessed by viability criteria based on the lactate clearance to levels ≤2.5 mmol/L within 4 h. The viability was achieved by 22 (71%) organs, that were transplanted after a median preservation time of 18 h, with 100% 90-day survival. During the median follow up of 542 days, 4 (18%) patients developed biliary strictures requiring re-transplantation. This trial demonstrates that viability testing with NMP is feasible and in this study enabled successful transplantation of 71% of discarded livers, with 100% 90-day patient and graft survival; it does not seem to prevent non-anastomotic biliary strictures in livers donated after circulatory death with prolonged warm ischaemia. The shortage of viable donated livers limits patient access to liver transplantation. Here the authors report the use of normothermic machine perfusion to help identify viable organs from livers discarded based on current clinical criteria, which are then transplanted to recipients in a single-arm clinical trial.

The Cancer Genome Atlas (TCGA) is one of the largest biorepositories of digital histology. Deep learning (DL) models have been trained on TCGA to predict numerous features directly from histology, including survival, gene expression patterns, and driver mutations. However, we demonstrate that these features vary substantially across tissue submitting sites in TCGA for over 3,000 patients with six cancer subtypes. Additionally, we show that histologic image differences between submitting sites can easily be identified with DL. Site detection remains possible despite commonly used color normalization and augmentation methods, and we quantify the image characteristics constituting this site-specific digital histology signature. We demonstrate that these site-specific signatures lead to biased accuracy for prediction of features including survival, genomic mutations, and tumor stage. Furthermore, ethnicity can also be inferred from site-specific signatures, which must be accounted for to ensure equitable application of DL. These site-specific signatures can lead to overoptimistic estimates of model performance, and we propose a quadratic programming method that abrogates this bias by ensuring models are not trained and validated on samples from the same site. Deep learning models have been trained on The Cancer Genome Atlas to predict numerous features directly from histology, including survival, gene expression patterns, and driver mutations. Here, the authors demonstrate that site-specific histologic signatures can lead to biased estimates of accuracy for such models, and propose a method to minimize such bias.

This paper reports a highly sensitive, non-invasive sensor for real-time glucose monitoring from interstitial fluid. The structure is comprised of a chip-less tag sensor which may be taped over the patient’s skin and a reader, that can be embedded in a smartwatch. The tag sensor is energized through the established electromagnetic coupling between the tag and the reader and its frequency response is reflected on the spectrum of the reader in the same manner. The tag sensor consumes zero power as there is no requirement for any active readout or communication circuitry on the tag side. When measuring changes in glucose concentrations within saline replicating interstitial fluid, the sensor was able to detect glucose with an accuracy of ~ 1 mM/l over a physiological range of glucose concentrations with 38 kHz of the resonance frequency shift. This high sensitivity is attained as a result of the proposed new design and extended field concentration on the tag. The impact of some of the possible interferences on the response of the sensor’s performance was also investigated. Variations in electrolyte concentrations within the test samples have a negligible effect on the response of the sensor unless these variations are supra-physiologically large.

Animal studies implicate meningeal lymphatic dysfunction in the pathogenesis of neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease (PD). However, there is no direct evidence in humans to support this role 1 – 5 . In this study, we used dynamic contrast-enhanced magnetic resonance imaging to assess meningeal lymphatic flow in cognitively normal controls and patients with idiopathic PD (iPD) or atypical Parkinsonian (AP) disorders. We found that patients with iPD exhibited significantly reduced flow through the meningeal lymphatic vessels (mLVs) along the superior sagittal sinus and sigmoid sinus, as well as a notable delay in deep cervical lymph node perfusion, compared to patients with AP. There was no significant difference in the size (cross-sectional area) of mLVs in patients with iPD or AP versus controls. In mice injected with α-synuclein (α-syn) preformed fibrils, we showed that the emergence of α-syn pathology was followed by delayed meningeal lymphatic drainage, loss of tight junctions among meningeal lymphatic endothelial cells and increased inflammation of the meninges. Finally, blocking flow through the mLVs in mice treated with α-syn preformed fibrils increased α-syn pathology and exacerbated motor and memory deficits. These results suggest that meningeal lymphatic drainage dysfunction aggravates α-syn pathology and contributes to the progression of PD. Reduced meningeal lymphatic flow detected in patients with idiopathic Parkinson’s disease compared to patients with atypical Parkinsonian disorders and cognitively normal controls.