Search

Writing in professional contexts requires professional techniques. Have a closer look at the techniques in this toolbox. They help you organize your workflow and manage the text production process during its four main phases: goal setting, planning, controlling, and revising. The toolbox is the result of long-term research into professional writing – a best-of collection of the tools applied by hundreds of professional writers under investigation. More: https://benjamins.com/catalog/aals.11 or https://link.springer.com/book/10.1007/978-3-658-10890-8 – Organize your workflow and workspace to prepare for a smooth and effective writing process. – Set the goal, the topic, the crucial idea of your communication project and of your text. Reduce it to the max. – View the parts of your text as a whole and recognize the interrelations among them before writing. – Write your text parts in a single flow, without jumping back and for¬th. Keep the momentum going. – Evaluate the text you have written – without being misled by your initial vision of the text you wanted to write.

Objective: To examine the relationship between current and former smoking and the occurrence of delirium in surgical Intensive Care Unit (ICU) patients. Methods: We conducted a single center, case-control study involving 244 delirious and 251 non-delirious patients that were admitted to our ICU between 2018 and 2022. Using propensity score analysis, we obtained 115 pairs of delirious and non-delirious patients matched for age and Simplified Acute Physiology Score II (SAPS II). Both groups of patients were further stratified into non-smokers, active smokers and former smokers, and logistic regression was performed to further investigate potential confounders. Results: Our study revealed a significant association between former smoking and the incidence of delirium in ICU patients, both in unmatched (adjusted odds ratio (OR): 1.82, 95% confidence interval (CI): 1.17-2.83) and matched cohorts (OR: 3.0, CI: 1.53-5.89). Active smoking did not demonstrate a significant difference in delirium incidence compared to non-smokers (unmatched OR = 0.98, CI: 0.62-1.53, matched OR = 1.05, CI: 0.55-2.0). Logistic regression analysis of the matched group confirmed former smoking as an independent risk factor for delirium, irrespective of other variables like surgical history (p = 0.010). Notably, also respiratory and vascular surgeries were associated with increased odds of delirium (respiratory: OR: 4.13, CI: 1.73-9.83; vascular: OR: 2.18, CI: 1.03-4.59). Medication analysis showed that while Ketamine and Midazolam usage did not significantly correlate with delirium, Morphine use was linked to a decreased likelihood (OR: 0.27, 95% CI: 0.13-0.55). Discussion: Nicotine’s complex neuropharmacological impact on the brain is still not fully understood, especially its short-term and long-term implications for critically ill patients. Although our retrospective study cannot establish causality, our findings suggest that smoking may induce structural changes in the brain, potentially heigh

Method development for the analysis of cell-free DNA (cfDNA) sequencing data is impeded by limited data sharing due to the strict control of sensitive genomic data. An existing solution for facilitating data sharing removes nucleotide-level information from raw cfDNA sequencing data, keeping alignment coordinates only. This simplified format can be publicly shared and would, theoretically, suffice for common functional analyses of cfDNA data. However, current bioinformatics software requires nucleotide-level information and cannot process the simplified format. We present Fragmentstein, a command-line tool for converting non-sensitive cfDNA-fragmentation data into alignment mapping (BAM) files. Fragmentstein complements fragment coordinates with sequence information from a reference genome to reconstruct BAM files. We demonstrate the utility of Fragmentstein by showing the feasibility of copy number variant (CNV), nucleosome occupancy, and fragment length analyses from non-sensitive fragmentation data. Availability and implementatio: Implemented in bash, Fragmentstein is available at https://github.com/uzh-dqbm-cmi/fragmentstein, licensed under GNU GPLv3.

AI-driven precision oncology has the transformative potential to reshape cancer treatment by leveraging the power of AI models to analyze the interaction between complex patient characteristics and their corresponding treatment outcomes. New technological platforms have facilitated the timely acquisition of multimodal data on tumor biology at an unprecedented resolution, such as single-cell multi-omics data, making this quality and quantity of data available for data-driven improved clinical decision-making. In this work, we propose a modular machine learning framework designed for personalized counterfactual cancer treatment suggestions based on an ensemble of machine learning experts trained on diverse multi-omics technologies. These specialized counterfactual experts per technology are consistently aggregated into a more powerful expert with superior performance and can provide both confidence and an explanation of its decision. The framework is tailored to address critical challenges inherent in data-driven cancer research, including the high-dimensional nature of the data, and the presence of treatment assignment bias in the retrospective observational data. The framework is showcased through comprehensive demonstrations using data from in-vitro and in-vivo treatment responses from a cohort of patients with ovarian cancer. Our method aims to empower clinicians with a reality-centric decision-support tool including probabilistic treatment suggestions with calibrated confidence and personalized explanations for tailoring treatment strategies to multi-omics characteristics of individual cancer patients., Comment: arXiv admin comment: This version has been removed by arXiv administrators as the submitter did not have the rights to agree to the license at the time of submission

The identification of phenotypes within complex diseases or syndromes is a fundamental component of precision medicine, which aims to adapt healthcare to individual patient characteristics. Postoperative delirium (POD) is a complex neuropsychiatric condition with significant heterogeneity in its clinical manifestations and underlying pathophysiology. We hypothesize that POD comprises several distinct phenotypes, which cannot be directly observed in clinical practice. Identifying these phenotypes could enhance our understanding of POD pathogenesis and facilitate the development of targeted prevention and treatment strategies. In this paper, we propose an approach that combines supervised machine learning for personalized POD risk prediction with unsupervised clustering techniques to uncover potential POD phenotypes. We first demonstrate our approach using synthetic data, where we simulate patient cohorts with predefined phenotypes based on distinct sets of informative features. We aim to mimic any clinical disease with our synthetic data generation method. By training a predictive model and applying SHAP, we show that clustering patients in the SHAP feature importance space successfully recovers the true underlying phenotypes, outperforming clustering in the raw feature space. We then present a case study using real-world data from a cohort of elderly surgical patients. The results showcase the utility of our approach in uncovering clinically relevant subtypes of complex disorders like POD, paving the way for more precise and personalized treatment strategies.

Background: Federated learning methods offer the possibility of training machine learning models on privacy-sensitive data sets, which cannot be easily shared. Multiple regulations pose strict requirements on the storage and usage of healthcare data, leading to data being in silos (i.e. locked-in at healthcare facilities). The application of federated algorithms on these datasets could accelerate disease diagnostic, drug development, as well as improve patient care. Methods: We present an extensive evaluation of the impact of different federation and differential privacy techniques when training models on the open-source MIMIC-III dataset. We analyze a set of parameters influencing a federated model performance, namely data distribution (homogeneous and heterogeneous), communication strategies (communication rounds vs. local training epochs), federation strategies (FedAvg vs. FedProx). Furthermore, we assess and compare two differential privacy (DP) techniques during model training: a stochastic gradient descent-based differential privacy algorithm (DP-SGD), and a sparse vector differential privacy technique (DP-SVT). Results: Our experiments show that extreme data distributions across sites (imbalance either in the number of patients or the positive label ratios between sites) lead to a deterioration of model performance when trained using the FedAvg strategy. This issue is resolved when using FedProx with the use of appropriate hyperparameter tuning. Furthermore, the results show that both differential privacy techniques can reach model performances similar to those of models trained without DP, however at the expense of a large quantifiable privacy leakage. Conclusions: We evaluate empirically the benefits of two federation strategies and propose optimal strategies for the choice of parameters when using differential privacy techniques.

Recent transformer-based models have made significant strides in generating radiology reports from chest X-ray images. However, a prominent challenge remains: these models often lack prior knowledge, resulting in the generation of synthetic reports that mistakenly reference non-existent prior exams. This discrepancy can be attributed to a knowledge gap between radiologists and the generation models. While radiologists possess patient-specific prior information, the models solely receive X-ray images at a specific time point. To tackle this issue, we propose a novel approach that leverages a rule-based labeler to extract comparison prior information from radiology reports. This extracted comparison prior is then seamlessly integrated into state-of-the-art transformer-based models, enabling them to produce more realistic and comprehensive reports. Our method is evaluated on English report datasets, such as IU X-ray and MIMIC-CXR. The results demonstrate that our approach surpasses baseline models in terms of natural language generation metrics. Notably, our model generates reports that are free from false references to non-existent prior exams, setting it apart from previous models. By addressing this limitation, our approach represents a significant step towards bridging the gap between radiologists and generation models in the domain of medical report generation., Comment: Accepted at ACL 2023, BioNLP Workshop

Contrastive learning methods have shown an impressive ability to learn meaningful representations for image or time series classification. However, these methods are less effective for time series forecasting, as optimization of instance discrimination is not directly applicable to predicting the future state from the history context. Moreover, the construction of positive and negative pairs in current technologies strongly relies on specific time series characteristics, restricting their generalization across diverse types of time series data. To address these limitations, we propose SimTS, a simple representation learning approach for improving time series forecasting by learning to predict the future from the past in the latent space. SimTS does not rely on negative pairs or specific assumptions about the characteristics of the particular time series. Our extensive experiments on several benchmark time series forecasting datasets show that SimTS achieves competitive performance compared to existing contrastive learning methods. Furthermore, we show the shortcomings of the current contrastive learning framework used for time series forecasting through a detailed ablation study. Overall, our work suggests that SimTS is a promising alternative to other contrastive learning approaches for time series forecasting., Comment: 13 pages, 6 figures

Objectives: This survey aims to provide an overview of the current state of biomedical and clinical Natural Language Processing (NLP) research and practice in Languages other than English (LoE). We pay special attention to data resources, language models, and popular NLP downstream tasks. Methods: We explore the literature on clinical and biomedical NLP from the years 2020-2022, focusing on the challenges of multilinguality and LoE. We query online databases and manually select relevant publications. We also use recent NLP review papers to identify the possible information lacunae. Results: Our work confirms the recent trend towards the use of transformer-based language models for a variety of NLP tasks in medical domains. In addition, there has been an increase in the availability of annotated datasets for clinical NLP in LoE, particularly in European languages such as Spanish, German and French. Common NLP tasks addressed in medical NLP research in LoE include information extraction, named entity recognition, normalization, linking, and negation detection. However, there is still a need for the development of annotated datasets and models specifically tailored to the unique characteristics and challenges of medical text in some of these languages, especially low-resources ones. Lastly, this survey highlights the progress of medical NLP in LoE, and helps at identifying opportunities for future research and development in this field.

OBJECTIVES: The first objective of this study was to implement and assess the performance and reliability of a vision transformer (ViT)-based deep-learning model, an 'off-the-shelf' artificial intelligence solution, for identifying distinct signs of microangiopathy in nailfold capilloroscopy (NFC) images of patients with SSc. The second objective was to compare the ViT's analysis performance with that of practising rheumatologists. METHODS: NFC images of patients prospectively enrolled in our European Scleroderma Trials and Research group (EUSTAR) and Very Early Diagnosis of Systemic Sclerosis (VEDOSS) local registries were used. The primary outcome investigated was the ViT's classification performance for identifying disease-associated changes (enlarged capillaries, giant capillaries, capillary loss, microhaemorrhages) and the presence of the scleroderma pattern in these images using a cross-fold validation setting. The secondary outcome involved a comparison of the ViT's performance vs that of rheumatologists on a reliability set, consisting of a subset of 464 NFC images with majority vote-derived ground-truth labels. RESULTS: We analysed 17 126 NFC images derived from 234 EUSTAR and 55 VEDOSS patients. The ViT had good performance in identifying the various microangiopathic changes in capillaries by NFC [area under the curve (AUC) from 81.8% to 84.5%]. In the reliability set, the rheumatologists reached a higher average accuracy, as well as a better trade-off between sensitivity and specificity compared with the ViT. However, the annotators' performance was variable, and one out of four rheumatologists showed equal or lower classification measures compared with the ViT. CONCLUSIONS: The ViT is a modern, well-performing and readily available tool for assessing patterns of microangiopathy on NFC images, and it may assist rheumatologists in generating consistent and high-quality NFC reports; however, the final diagnosis of a scleroderma pattern in any individual cas

We propose a new automated evaluation metric for machine-generated radiology reports using the successful COMET architecture adapted for the radiology domain. We train and publish four medically-oriented model checkpoints, including one trained on RadGraph, a radiology knowledge graph. Our results show that our metric correlates moderately to high with established metrics such as BERTscore, BLEU, and CheXbert scores. Furthermore, we demonstrate that one of our checkpoints exhibits a high correlation with human judgment, as assessed using the publicly available annotations of six board-certified radiologists, using a set of 200 reports. We also performed our own analysis gathering annotations with two radiologists on a collection of 100 reports. The results indicate the potential effectiveness of our method as a radiology-specific evaluation metric. The code, data, and model checkpoints to reproduce our findings will be publicly available., Comment: 9 pages

Irregular multivariate time series data is characterized by varying time intervals between consecutive observations of measured variables/signals (i.e., features) and varying sampling rates (i.e., recordings/measurement) across these features. Modeling time series while taking into account these irregularities is still a challenging task for machine learning methods. Here, we introduce TADA, a Two-stageAggregation process with Dynamic local Attention to harmonize time-wise and feature-wise irregularities in multivariate time series. In the first stage, the irregular time series undergoes temporal embedding (TE) using all available features at each time step. This process preserves the contribution of each available feature and generates a fixed-dimensional representation per time step. The second stage introduces a dynamic local attention (DLA) mechanism with adaptive window sizes. DLA aggregates time recordings using feature-specific windows to harmonize irregular time intervals capturing feature-specific sampling rates. Then hierarchical MLP mixer layers process the output of DLA through multiscale patching to leverage information at various scales for the downstream tasks. TADA outperforms state-of-the-art methods on three real-world datasets, including the latest MIMIC IV dataset, and highlights its effectiveness in handling irregular multivariate time series and its potential for various real-world applications., Comment: A short version of this paper has been accepted for presentation at the Findings of Machine Learning for Health (ML4H) 2023 conference

In this paper, we propose a deep generative time series approach using latent temporal processes for modeling and holistically analyzing complex disease trajectories. We aim to find meaningful temporal latent representations of an underlying generative process that explain the observed disease trajectories in an interpretable and comprehensive way. To enhance the interpretability of these latent temporal processes, we develop a semi-supervised approach for disentangling the latent space using established medical concepts. By combining the generative approach with medical knowledge, we leverage the ability to discover novel aspects of the disease while integrating medical concepts into the model. We show that the learned temporal latent processes can be utilized for further data analysis and clinical hypothesis testing, including finding similar patients and clustering the disease into new sub-types. Moreover, our method enables personalized online monitoring and prediction of multivariate time series including uncertainty quantification. We demonstrate the effectiveness of our approach in modeling systemic sclerosis, showcasing the potential of our machine learning model to capture complex disease trajectories and acquire new medical knowledge., Comment: Extended Abstract presented at Machine Learning for Health (ML4H) symposium 2023, December 10th, 2023, New Orleans, United States, 23 pages

Human genetic diseases often arise from point mutations, emphasizing the critical need for precise genome editing techniques. Among these, base editing stands out as it allows targeted alterations at the single nucleotide level. However, its clinical application is hindered by low editing efficiency and unintended mutations, necessitating extensive trial-and-error experimentation in the laboratory. To speed up this process, we present an attention-based two-stage machine learning model that learns to predict the likelihood of all possible editing outcomes for a given genomic target sequence. We further propose a multi-task learning schema to jointly learn multiple base editors (i.e. variants) at once. Our model's predictions consistently demonstrated a strong correlation with the actual experimental results on multiple datasets and base editor variants. These results provide further validation for the models' capacity to enhance and accelerate the process of refining base editing designs., Comment: Extended Abstract presented at Machine Learning for Health (ML4H) symposium 2023, December 10th, 2023, New Orleans, United States, 15 pages. arXiv admin note: substantial text overlap with arXiv:2310.02919

Human genetic diseases often arise from point mutations, emphasizing the critical need for precise genome editing techniques. Among these, base editing stands out as it allows targeted alterations at the single nucleotide level. However, its clinical application is hindered by low editing efficiency and unintended mutations, necessitating extensive trial-and-error experimentation in the laboratory. To speed up this process, we present an attention-based two-stage machine learning model that learns to predict the likelihood of all possible editing outcomes for a given genomic target sequence. We further propose a multi-task learning schema to jointly learn multiple base editors (i.e. variants) at once. Our model's predictions consistently demonstrated a strong correlation with the actual experimental results on multiple datasets and base editor variants. These results provide further validation for the models' capacity to enhance and accelerate the process of refining base editing designs.

We propose a novel framework that combines deep generative time series models with decision theory for generating personalized treatment strategies. It leverages historical patient trajectory data to jointly learn the generation of realistic personalized treatment and future outcome trajectories through deep generative time series models. In particular, our framework enables the generation of novel multivariate treatment strategies tailored to the personalized patient history and trained for optimal expected future outcomes based on conditional expected utility maximization. We demonstrate our framework by generating personalized insulin treatment strategies and blood glucose predictions for hospitalized diabetes patients, showcasing the potential of our approach for generating improved personalized treatment strategies. Keywords: deep generative model, probabilistic decision support, personalized treatment generation, insulin and blood glucose prediction, Comment: Extended Abstract presented at Machine Learning for Health (ML4H) symposium 2023, December 10th, 2023, New Orleans, United States, 17 pages

Drug synergy arises when the combined impact of two drugs exceeds the sum of their individual effects. While single-drug effects on cell lines are well-documented, the scarcity of data on drug synergy, considering the vast array of potential drug combinations, prompts a growing interest in computational approaches for predicting synergies in untested drug pairs. We introduce a Graph Neural Network (\textit{GNN}) based model for drug synergy prediction, which utilizes drug chemical structures and cell line gene expression data. We extract data from the largest available drug combination database (DrugComb) and generate multiple synergy scores (commonly used in the literature) to create seven datasets that serve as a reliable benchmark with high confidence. In contrast to conventional models relying on pre-computed chemical features, our GNN-based approach learns task-specific drug representations directly from the graph structure of the drugs, providing superior performance in predicting drug synergies. Our work suggests that learning task-specific drug representations and leveraging a diverse dataset is a promising approach to advancing our understanding of drug-drug interaction and synergy.

Metabolic reprogramming is an emerging hallmark of resistance to cancer therapy but may generate vulnerabilities that can be targeted with small molecules. Multi-omics analysis revealed that NRAS-mutated melanoma cells with a mesenchymal transcriptional profile adopt a quiescent metabolic program to resist cellular stress response induced by MEK-inhibitor resistance. However, as a result of elevated baseline ROS levels, these cells become highly sensitive to ROS induction. In vivo xenograft experiments and single-cell RNA sequencing demonstrated that intra-tumor heterogeneity requires the combination of a ROS-inducer and a MEK-inhibitor to target both tumor growth and metastasis. By ex vivo pharmacoscopy of 62 human metastatic melanomas, we found that MEK-inhibitor resistant tumors significantly benefitted from the combination therapy. Finally, we profiled 486 cancer cell lines and revealed that oxidative stress responses and translational suppression are biomarkers of ROS-inducer sensitivity, independent of cancer indication. These findings link transcriptional plasticity to a metabolic phenotype that can be inhibited by ROS-inducers in melanoma and other cancers. Statement of Significance Targeted-therapy resistance in cancer arises from genetic selection and both transcriptional and metabolic adaptation. We show that metabolic reprogramming sensitizes resistant cells to ROS-induction in combination with pathway inhibitors. Predictive biomarkers of metabolic sensitivity to ROS-inducing agents were identified in many cancer entities, highlighting the generalizability of this treatment approach.

BACKGROUND: Drug synergy occurs when the combined effect of two drugs is greater than the sum of the individual drugs' effect. While cell line data measuring the effect of single drugs are readily available, there is relatively less comparable data on drug synergy given the vast amount of possible drug combinations. Thus, there is interest to use computational approaches to predict drug synergy for untested pairs of drugs. METHODS: We introduce a Graph Neural Network (GNN) based model for drug synergy prediction, which utilizes drug chemical structures and cell line gene expression data. We use information from the largest drug combination database available (DrugComb), combining drug synergy scores in order to construct high confidence benchmark datasets. RESULTS: Our proposed solution for drug synergy predictions offers a number of benefits: 1) It utilizes a combination of 34 distinct drug synergy datasets to learn on a wide variety of drugs and cell lines representations. 2) It is trained on constructed high confidence benchmark datasets. 3) It learns task-specific drug representations, instead of relying on generalized and pre-computed chemical drug features. 4) It achieves similar or better prediction performance (AUPR scores ranging from 0.777 to 0.964) compared to state-of-the-art baseline models when tested on various benchmark datasets. CONCLUSIONS: We demonstrate that a GNN based model can provide state-of-the-art drug synergy predictions by learning task-specific representations of drugs.

Biobanking of surplus human healthy and disease-derived tissues is essential for diagnostics and translational research. An enormous amount of formalin-fixed and paraffin-embedded (FFPE), Tissue-Tek OCT embedded or snap-frozen tissues are preserved in many biobanks worldwide and have been the basis of translational studies. However, their usage is limited to assays that do not require viable cells. The access to intact and viable human material is a prerequisite for translational validation of basic research, for novel therapeutic target discovery, and functional testing. Here we show that surplus tissues from multiple solid human cancers directly slow-frozen after resection can subsequently be used for different types of methods including the establishment of 2D, 3D, and ex vivo cultures as well as single-cell RNA sequencing with similar results when compared to freshly analyzed material.

In this NTCIR-16 Real-MedNLP shared task paper, we present the methods of the ZuKyo-JA subteam for solving the Japanese part of Subtask1 and Subtask3 (Subtask1-CR-JA, Subtask1-RR- JA, Subtask3-RR-JA). Our solution is based on a sliding- window approach using a Japanese BERT pre-trained masked- language model., which was used as a common architecture for addressing the specific subtasks. We additionally present a method that makes extensive use of medical knowledge for the same case identification subtask3-RR-JA.

In this paper, we discuss our contribution to the NII Testbeds and Community for Information Access Research (NTCIR) - 16 Real- MedNLP shared task. Our team (ZuKyo) participated in the English subtask: Few-resource Named Entity Recognition. The main challenge in this low-resource task was a low number of training documents annotated with a high number of tags and attributes. For our submissions, we used different general and domain-specific transfer learning approaches in combination with multiple data augmentation methods. In addition, we experimented with models enriched with biomedical concepts encoded as token-based input features

Nowadays, the interpretation of why a machine learning (ML) model makes certain inferences is as crucial as the accuracy of such inferences. Some ML models like the decision tree possess inherent interpretability that can be directly comprehended by humans. Others like artificial neural networks (ANN), however, rely on external methods to uncover the deduction mechanism. SHapley Additive exPlanations (SHAP) is one of such external methods, which requires a background dataset when interpreting ANNs. Generally, a background dataset consists of instances randomly sampled from the training dataset. However, the sampling size and its effect on SHAP remain to be unexplored. In our empirical study on the MIMIC-III dataset, we show that the two core explanations - SHAP values and variable rankings fluctuate when using different background datasets acquired from random sampling, indicating that users cannot unquestioningly trust the one-shot interpretation from SHAP. Luckily, such fluctuation decreases with the increase of the background dataset size. Also, we notice an U-shape in the stability assessment of SHAP variable rankings, demonstrating that SHAP is more reliable in ranking the most and least important variables compared to moderately important ones. Overall, our results suggest that users should take into account how background data affects SHAP results, with improved SHAP stability as the background sample size increases.

In recent decades, there has been an increase in polypharmacy, the concurrent administration of multiple drugs per patient. Studies have shown that polypharmacy is linked to adverse patient outcomes and there is interest in elucidating the exact causes behind this observation. In this paper, we are studying the relationship between drug prescriptions, drug-drug interactions (DDIs) and patient mortality. Our focus is not so much on the number of prescribed drugs, the typical metric in polypharmacy research, but rather on the specific combinations of drugs leading to a DDI. To learn the space of real-world drug combinations, we first assessed the drug prescription landscape of the UK Biobank, a large patient data registry. We observed distinct drug constellation patterns driven by the UK Biobank participants' disease status. We show that these drug prescription clusters matter in terms of the number and types of expected DDIs, and may possibly explain observed differences in health outcomes.

INTRODUCTION The human microbiota, the community of micro-organisms in different cavities, has been increasingly linked with inflammatory and neoplastic diseases. While investigation into the gut microbiome has been robust, the urinary microbiome has only recently been described. Investigation into the relationship between bladder cancer (BC) and the bladder and the intestinal microbiome may elucidate a pathophysiological relationship between the two. The bladder or the intestinal microbiome or the interplay between both may also act as a non-invasive biomarker for tumour behaviour. While these associations have not yet been fully investigated, urologists have been manipulating the bladder microbiome for treatment of BC for more than 40 years, treating high grade non-muscle invasive BC (NMIBC) with intravesical BCG immunotherapy. Neither the association between the microbiome sampled directly from bladder tissue and the response to BCG-therapy nor the association between response to BCG-therapy with the faecal microbiome has been studied until now. A prognostic tool prior to initiation of BCG-therapy is still needed. METHODS AND ANALYSIS In patients with NMIBC bladder samples will be collected during surgery (bladder microbiome assessment), faecal samples (microbiome assessment), instrumented urine and blood samples (biobank) will also be taken. We will analyse the microbial community by 16S rDNA gene amplicon sequencing. The difference in alpha diversity (diversity of species within each sample) and beta diversity (change in species diversity) between BCG-candidates will be assessed. Subgroup analysis will be performed which will lead to the development of a clinical prediction model estimating risk of BCG-response. ETHICS AND DISSEMINATION The study has been approved by the Cantonal Ethics Committee Zurich (2021-01783) and it is being conducted in accordance with the Declaration of Helsinki and Good Clinical Practice. Study results will be disseminated through pee

Inspired by Curriculum Learning, we propose a consecutive (i.e., image-to-text-to-text) generation framework where we divide the problem of radiology report generation into two steps. Contrary to generating the full radiology report from the image at once, the model generates global concepts from the image in the first step and then reforms them into finer and coherent texts using a transformer architecture. We follow the transformer-based sequence-to-sequence paradigm at each step. We improve upon the state-of-the-art on two benchmark datasets., Comment: Accepted to findings of EMNLP 2021

The continued digitalization of medicine has led to an increased availability of longitudinal patient data that allows the investigation of novel and known diseases in unprecedented detail. However, to accurately describe any underlying pathophysiology and allow inter-patient comparisons, individual patient trajectories have to be synchronized based on temporal markers. In this pilot study, we use longitudinal data from critically ill ICU COVID-19 patients to compare the commonly used alignment markers "onset of symptoms," "hospital admission," and "ICU admission" with a novel objective method based on the peak value of the inflammatory marker C-reactive protein (CRP). By applying our CRP-based method to align the progression of neutrophils and lymphocytes, we were able to define a pathophysiological window that improved mortality risk stratification in our COVID-19 patient cohort. Our data highlights that proper synchronization of longitudinal patient data is crucial for accurate interpatient comparisons and the definition of relevant subgroups. The use of objective temporal disease markers will facilitate both translational research efforts and multicenter trials.

High cell viability and recovered cell concentration are typical quality control requirements for single-cell processing and quality data. This protocol describes procedures for sampling, live-cell biobanking, preprocessing for single-cell RNA sequencing, and analysis of fine-needle aspiration (FNA) samples of the skin. The minimally invasive nature of FNA collection is more accepted by patients and allows for frequent longitudinal sampling, resulting in high-quality single-cell sequencing data that capture cellular heterogeneity in clinical samples.

Background: Drug-drug interactions (DDIs) refer to processes triggered by the administration of two or more drugs leading to side effects beyond those observed when drugs are administered by themselves. Due to the massive number of possible drug pairs, it is nearly impossible to experimentally test all combinations and discover previously unobserved side effects. Therefore, machine learning based methods are being used to address this issue. Methods: We propose a Siamese self-attention multi-modal neural network for DDI prediction that integrates multiple drug similarity measures that have been derived from a comparison of drug characteristics including drug targets, pathways and gene expression profiles. Results: Our proposed DDI prediction model provides multiple advantages: (1) It is trained end-to-end, overcoming limitations of models composed of multiple separate steps, (2) it offers model explainability via an Attention mechanism for identifying salient input features and (3) it achieves similar or better prediction performance (AUPR scores ranging from 0.77 to 0.92) compared to state-of-the-art DDI models when tested on various benchmark datasets. Novel DDI predictions are further validated using independent data resources. Conclusions: We find that a Siamese multi-modal neural network is able to accurately predict DDIs and that an Attention mechanism, typically used in the Natural Language Processing domain, can be beneficially applied to aid in DDI model explainability. Keywords: Attention; Deep learning; Drug–drug interactions; Prediction; Side effects

Der Begriff der künstlichen Intelligenz stammt aus den 50er-Jahren des 20. Jahrhunderts und umfasst eine Ansammlung von Technologien, welche einem Computer erlauben, typische Eigenschaften der menschlichen Intelligenz zu emulieren. Anfänglich wurden grosse Hoffnungen in diese Technologien gesetzt und früh wurde versucht, diese in der Medizin zu etablieren. Zu Beginn des 21. Jahrhunderts war es aufgrund der anfänglich ernüchternden Ergebnisse eher still in der Erforschung der KI in der Medizin. Mehrere wichtige Entwicklungen ebneten jedoch den Weg zum Durchbruch der Technologie.

Base editors are chimeric ribonucleoprotein complexes consisting of a DNA-targeting CRISPR-Cas module and a single-stranded DNA deaminase. They enable transition of C•G into T•A base pairs and vice versa on genomic DNA. While base editors have great potential as genome editing tools for basic research and gene therapy, their application has been hampered by a broad variation in editing efficiencies on different genomic loci. Here we perform an extensive analysis of adenine- and cytosine base editors on a library of 28,294 lentivirally integrated genetic sequences and establish BE-DICT, an attention-based deep learning algorithm capable of predicting base editing outcomes with high accuracy. BE-DICT is a versatile tool that in principle can be trained on any novel base editor variant, facilitating the application of base editing for research and therapy.

In digital medicine, patient data typically record health events over time (eg, through electronic health records, wearables, or other sensing technologies) and thus form unique patient trajectories. Patient trajectories are highly predictive of the future course of diseases and therefore facilitate effective care. However, digital medicine often uses only limited patient data, consisting of health events from only a single or small number of time points while ignoring additional information encoded in patient trajectories. To analyze such rich longitudinal data, new artificial intelligence (AI) solutions are needed. In this paper, we provide an overview of the recent efforts to develop trajectory-aware AI solutions and provide suggestions for future directions. Specifically, we examine the implications for developing disease models from patient trajectories along the typical workflow in AI: problem definition, data processing, modeling, evaluation, and interpretation. We conclude with a discussion of how such AI solutions will allow the field to build robust models for personalized risk scoring, subtyping, and disease pathway discovery.

Inspired by Curriculum Learning, we propose a consecutive (i.e., image-to-text-to-text) generation framework where we divide the problem of radiology report generation into two steps. Contrary to generating the full radiology report from the image at once, the model generates global concepts from the image in the first step and then reforms them into finer and coherent texts using transformer-based architecture. We follow the transformer-based sequence-to-sequence paradigm at each step. We improve upon the state-of-the-art on two benchmark datasets.

BACKGROUND: MUC16 is a mucin marker that is frequently mutated in melanoma, but whether MUC16 mutations could be useful as a surrogate biomarker for tumor mutation burden (TMB) remains unclear. METHODS: This study rigorously evaluates the MUC16 mutation as a clinical biomarker in cutaneous melanoma by utilizing genomic and clinical data from patient samples from The Cancer Genome Atlas (TCGA) and two independent validation cohorts. We further extended the analysis to studies with patients treated with immunotherapies. RESULTS: Analysis results showed that samples with MUC16 mutations had a higher TMB than the samples of wild-type, with strong statistical significance (P < 0.001) in all melanoma cohorts tested. Associations between MUC16 mutations and TMB remained statistically significant after adjusting for potential confounding factors in the TCGA cohort [OR, 9.28 (95% confidence interval (CI), 5.18-17.39); P < 0.001], Moffitt cohort [OR, 31.95 (95% CI, 8.71-163.90); P < 0.001], and Yale cohort [OR, 8.09 (95% CI, 3.12-23.79); P < 0.01]. MUC16 mutations were also found to be associated with overall survival in the TCGA [HR, 0.62; (95% CI, 0.45-0.85); P < 0.01] and Moffitt cohorts [HR, 0.49 (95% CI, 0.28-0.87); P = 0.014]. Strikingly, MUC16 is the only top frequently mutated gene for which prognostic significance was observed. MUC16 mutations were also found valuable in predicting anti-CTLA-4 and anti-PD-1 therapy responses. CONCLUSIONS: MUC16 mutation appears to be a useful predictive marker of global TMB and patient survival in melanoma. IMPACT: This is, to the best of our knowledge, the first systematic evaluation of MUC16 mutation as a clinical biomarker and a predictive biomarker for immunotherapy in melanoma.

Healthcare professionals have long envisioned using the enormous processing powers of computers to discover new facts and medical knowledge locked inside electronic health records. These vast medical archives contain time-resolved information about medical visits, tests and procedures, as well as outcomes, which together form individual patient journeys. By assessing the similarities among these journeys, it is possible to uncover clusters of common disease trajectories with shared health outcomes. The assignment of patient journeys to specific clusters may in turn serve as the basis for personalized outcome prediction and treatment selection. This procedure is a non-trivial computational problem, as it requires the comparison of patient data with multi-dimensional and multi-modal features that are captured at different times and resolutions. In this review, we provide a comprehensive overview of the tools and methods that are used in patient similarity analysis with longitudinal data and discuss its potential for improving clinical decision making.

Background: Drug-drug interactions (DDIs) refer to processes triggered by the administration of two or more drugs leading to side effects beyond those observed when drugs are administered by themselves. Due to the massive number of possible drug pairs, it is nearly impossible to experimentally test all combinations and discover previously unobserved side effects. Therefore, machine learning based methods are being used to address this issue. Methods: We propose a Siamese self-attention multi-modal neural network for DDI prediction that integrates multiple drug similarity measures that have been derived from a comparison of drug characteristics including drug targets, pathways and gene expression profiles. Results: Our proposed DDI prediction model provides multiple advantages: 1) It is trained end-to-end, overcoming limitations of models composed of multiple separate steps, 2) it offers model explainability via an Attention mechanism for identifying salient input features and 3) it achieves similar or better prediction performance (AUPR scores ranging from 0.77 to 0.92) compared to state-of-the-art DDI models when tested on various benchmark datasets. Novel DDI predictions are further validated using independent data resources. Conclusions: We find that a Siamese multi-modal neural network is able to accurately predict DDIs and that an Attention mechanism, typically used in the Natural Language Processing domain, can be beneficially applied to aid in DDI model explainability.

The amount of data that is stored in databases and must be analyzed is growing fast. Many analytical tasks are based on iterative methods that approximate optimal solutions. Propensity score matching is a technique that is used to reduce bias during cohort building. The main step is the propensity score computation, which is usually implemented via iterative methods such as gradient descent. Our goal is to support efficient and scalable propensity score computation over relations in a column-oriented database. To achieve this goal, we introduce shape-preserving iterations that update values in existing tuples until a fix point is reached. Shape-preserving iterations enable gradient descent over relations and, thus, propensity score matching. We also show how to create appropriate input relations for shape-preserving iterations with randomly initialized relations. The empirical evaluation compares in-database iterations with the native implementation in MonetDB where iterations are flattened.

Artificial intelligence (AI) systems are increasingly being used in healthcare, thanks to the high level of performance that these systems have proven to deliver. So far, clinical applications have focused on diagnosis and on prediction of outcomes. It is less clear in what way AI can or should support complex clinical decisions that crucially depend on patient preferences. In this paper, we focus on the ethical questions arising from the design, development and deployment of AI systems to support decision-making around cardio-pulmonary resuscitation leading to the determination of a patient's Do Not Attempt to Resuscitate (DNAR) status (also known as code status). The COVID-19 pandemic has made us keenly aware of the difficulties physicians encounter when they have to act quickly in stressful situations without knowing what their patient would have wanted. We discuss the results of an interview study conducted with healthcare professionals in a university hospital aimed at understanding the status quo of resuscitation decision processes while exploring a potential role for AI systems in decision-making around code status. Our data suggest that 1) current practices are fraught with challenges such as insufficient knowledge regarding patient preferences, time pressure and personal bias guiding care considerations and 2) there is considerable openness among clinicians to consider the use of AI-based decision support. We suggest a model for how AI can contribute to improve decision-making around resuscitation and propose a set of ethically relevant preconditions - conceptual, methodological and procedural - that need to be considered in further development and implementation efforts.

Background: Patients increasingly turn to search engines and online content before, or in place of, talking with a health professional. Low quality health information, which is common on the internet, presents risks to the patient in the form of misinformation and a possibly poorer relationship with their physician. To address this, the DISCERN criteria (developed at University of Oxford) are used to evaluate the quality of online health information. However, patients are unlikely to take the time to apply these criteria to the health websites they visit. Methods: We built an automated implementation of the DISCERN instrument (Brief version) using machine learning models. We compared the performance of a traditional model (Random Forest) with that of a hierarchical encoder attention-based neural network (HEA) model using two language embeddings, BERT and BioBERT. Results: The HEA BERT and BioBERT models achieved average F1-macro scores across all criteria of 0.75 and 0.74, respectively, outperforming the Random Forest model (average F1-macro = 0.69). Overall, the neural network based models achieved 81% and 86% average accuracy at 100% and 80% coverage, respectively, compared to 94% manual rating accuracy. The attention mechanism implemented in the HEA architectures not only provided 'model explainability' by identifying reasonable supporting sentences for the documents fulfilling the Brief DISCERN criteria, but also boosted F1 performance by 0.05 compared to the same architecture without an attention mechanism. Conclusions: Our research suggests that it is feasible to automate online health information quality assessment, which is an important step towards empowering patients to become informed partners in the healthcare process.

Base editors are chimeric ribonucleoprotein complexes consisting of a DNA-targeting CRISPR-Cas module and a single-stranded DNA deaminase. They enable conversion of C•G into T•A base pairs and vice versa on genomic DNA. While base editors have vast potential as genome editing tools for basic research and gene therapy, their application has been hampered by a broad variation in editing efficiencies on different genomic loci. Here we perform an extensive analysis of adenine- and cytosine base editors on thousands of lentivirally integrated genetic sequences and establish BE-DICT, an attention-based deep learning algorithm capable of predicting base editing outcomes with high accuracy. BE-DICT is a versatile tool that in principle can be trained on any novel base editor variant, facilitating the application of base editing for research and therapy.

Uropathogenic Escherichia coli (UPEC) is the primary causative agent of uncomplicated urinary tract infections (UTIs). UPEC fitness and virulence determinants have been evaluated in a variety of laboratory settings, including a well-established mouse model of UTI. However, the extent to which bacterial physiologies differ between experimental models and human infections remains largely understudied. To address this important issue, we compared the transcriptomes of three different UPEC isolates in human infection and under a variety of laboratory conditions, including LB culture, filter-sterilized urine culture, and the UTI mouse model. We observed high correlation in gene expression between the mouse model and human infection in all three strains examined (Pearson correlation coefficients of 0.86 to 0.87). Only 175 of 3,266 (5.4%) genes shared by all three strains had significantly different expression levels, with the majority of them (145 genes) downregulated in patients. Importantly, gene expression levels of both canonical virulence factors and metabolic machinery were highly similar between the mouse model and human infection, while the in vitro conditions displayed more substantial differences. Interestingly, comparison of gene expression between the mouse model and human infection hinted at differences in bladder oxygenation as well as nutrient composition. In summary, our work strongly validates the continued use of this mouse model for the study of the pathogenesis of human UTI.

The major drawback of melanoma therapy with BRAF and MAPK inhibitors is the innate and acquired drug resistance. We therefore explored alternative targets and developed a new compound, SAB298, that is a SRC-family kinase (SFK) inhibitor. The drug is cytotoxic to patient-derived melanoma cells regardless of oncogene expression and inhibits tumor growth in vivo. As expected, it inhibited SRC and PI3K activity, and had the additional property of ERBB2 inhibition, that lead to inactivation of the two ERK phosphatases PP2A and SHP2. In 57% of the melanoma cell lines tested, the consequent increase in ERK activity lead to proteolytic degradation of its substrate, the lineage specific transcription factor MITF, likely contributing to growth arrest. Treatment with a combination of SAB298 and AZD6244 (selumetinib), induced a synergistic growth inhibition, suggesting that the new compound could be used in the clinic as a substitute for, or in combination with MAPK inhibitors.

Heart failure (HF) is one of the leading causes of hospital admissions in the US. Readmission within 30 days after a HF hospitalization is both a recognized indicator for disease progression and a source of considerable financial burden to the healthcare system. Consequently, the identification of patients at risk for readmission is a key step in improving disease management and patient outcome. In this work, we used a large administrative claims dataset to (1) explore the systematic application of neural network-based models versus logistic regression for predicting 30 days all-cause readmission after discharge from a HF admission, and (2) to examine the additive value of patients’ hospitalization timelines on prediction performance. Based on data from 272,778 (49% female) patients with a mean (SD) age of 73 years (14) and 343,328 HF admissions (67% of total admissions), we trained and tested our predictive readmission models following a stratified 5-fold cross-validation scheme. Among the deep learning approaches, a recurrent neural network (RNN) combined with conditional random fields (CRF) model (RNNCRF) achieved the best performance in readmission prediction with 0.642 AUC (95% CI, 0.640–0.645). Other models, such as those based on RNN, convolutional neural networks and CRF alone had lower performance, with a non-timeline based model (MLP) performing worst. A competitive model based on logistic regression with LASSO achieved a performance of 0.643 AUC (95% CI, 0.640–0.646). We conclude that data from patient timelines improve 30 day readmission prediction, that a logistic regression with LASSO has equal performance to the best neural network model and that the use of administrative data result in competitive performance compared to published approaches based on richer clinical datasets.

Patients increasingly turn to search engines and online content before, or in place of, talking with a health professional. Low quality health information, which is common on the internet, presents risks to the patient in the form of misinformation and a possibly poorer relationship with their physician. To address this, the DISCERN criteria (developed at University of Oxford) are used to evaluate the quality of online health information. However, patients are unlikely to take the time to apply these criteria to the health websites they visit. We built an automated implementation of the DISCERN instrument (Brief version) using machine learning models. We compared the performance of a traditional model (Random Forest) with that of a hierarchical encoder attention-based neural network (HEA) model using two language embeddings, BERT and BioBERT. The HEA BERT and BioBERT models achieved average F1-macro scores across all criteria of 0.75 and 0.74, respectively, outperforming the Random Forest model (average F1-macro = 0.69). Overall, the neural network based models achieved 81% and 86% average accuracy at 100% and 80% coverage, respectively, compared to 94% manual rating accuracy. The attention mechanism implemented in the HEA architectures not only provided 'model explainability' by identifying reasonable supporting sentences for the documents fulfilling the Brief DISCERN criteria, but also boosted F1 performance by 0.05 compared to the same architecture without an attention mechanism. Our research suggests that it is feasible to automate online health information quality assessment, which is an important step towards empowering patients to become informed partners in the healthcare process.

Background: We describe a prototype implementation of a platform that could underlie a Precision Oncology Rapid Learning system. Results: We describe the prototype platform, and examine some important issues and details. In the Appendix we provide a complete walk-through of the prototype platform. Conclusions: The design choices made in this implementation rest upon ten constitutive hypotheses, which, taken together, define a particular view of how a rapid learning medical platform might be defined, organized, and implemented. Keywords: Natural language processing, Precision oncology, Controlled natural language, Nanopublication, Treatment reasoning, Rapid learning, Tumor boards, Targeted therapies, Cancer Commons, Deloitte Touche Tohmatsu (Firm)

Heart failure (HF) is one of the leading causes of hospital admissions in the US. Readmission within 30 days after a HF hospitalization is both a recognized indicator for disease progression and a source of considerable financial burden to the healthcare system. Consequently, the identification of patients at risk for readmission is a key step in improving disease management and patient outcome. In this work, we used a large administrative claims dataset to (1)explore the systematic application of neural network-based models versus logistic regression for predicting 30 days all-cause readmission after discharge from a HF admission, and (2)to examine the additive value of patients' hospitalization timelines on prediction performance. Based on data from 272,778 (49% female) patients with a mean (SD) age of 73 years (14) and 343,328 HF admissions (67% of total admissions), we trained and tested our predictive readmission models following a stratified 5-fold cross-validation scheme. Among the deep learning approaches, a recurrent neural network (RNN) combined with conditional random fields (CRF) model (RNNCRF) achieved the best performance in readmission prediction with 0.642 AUC (95% CI, 0.640-0.645). Other models, such as those based on RNN, convolutional neural networks and CRF alone had lower performance, with a non-timeline based model (MLP) performing worst. A competitive model based on logistic regression with LASSO achieved a performance of 0.643 AUC (95%CI, 0.640-0.646). We conclude that data from patient timelines improve 30 day readmission prediction for neural network-based models, that a logistic regression with LASSO has equal performance to the best neural network model and that the use of administrative data result in competitive performance compared to published approaches based on richer clinical datasets.