Search

Phenotypic plasticity, the ability to produce multiple phenotypes from a single genotype, represents an excellent model with which to examine the relationship between gene expression and phenotypes. Analyses of the molecular foundations of phenotypic plasticity are challenging, however, especially in the case of complex social phenotypes. Here we apply a machine learning approach to tackle this challenge by analyzing individual-level gene expression profiles of Polistes dominula paper wasps following the loss of a queen. We find that caste-associated gene expression profiles respond strongly to queen loss, and that this change is partly explained by attributes such as age but occurs even in individuals that appear phenotypically unaffected. These results demonstrate that large changes in gene expression may occur in the absence of outwardly detectable phenotypic changes, resulting here in a socially mediated de-differentiation of individuals at the transcriptomic level but not at the levels of ovarian development or behavior., Connecting genotypes to complex social behaviour is challenging. Taylor et al. use machine learning to show a strong response of caste-associated gene expression to queen loss, wherein individual wasp’s expression profiles become intermediate between queen and worker states, even in the absence of behavioural changes.

In daily clinical practice, clinicians integrate available data to ascertain the diagnostic and prognostic probability of a disease or clinical outcome for their patients. For patients with suspected or known cardiovascular disease, several anatomical and functional imaging techniques are commonly performed to aid this endeavor, including coronary computed tomography angiography (CCTA) and nuclear cardiology imaging. Continuous improvement in positron emission tomography (PET), single-photon emission computed tomography (SPECT), and CT hardware and software has resulted in improved diagnostic performance and wide implementation of these imaging techniques in daily clinical practice. However, the human ability to interpret, quantify, and integrate these data sets is limited. The identification of novel markers and application of machine learning (ML) algorithms, including deep learning (DL) to cardiovascular imaging techniques will further improve diagnosis and prognostication for patients with cardiovascular diseases. The goal of this position paper of the European Association of Nuclear Medicine (EANM) and the European Association of Cardiovascular Imaging (EACVI) is to provide an overview of the general concepts behind modern machine learning-based artificial intelligence, highlights currently prefered methods, practices, and computational models, and proposes new strategies to support the clinical application of ML in the field of cardiovascular imaging using nuclear cardiology (hybrid) and CT techniques.

Identifying the genes and mutations that drive the emergence of tumors is a critical step to improving our understanding of cancer and identifying new directions for disease diagnosis and treatment. Despite the large volume of genomics data, the precise detection of driver mutations and their carrying genes, known as cancer driver genes, from the millions of possible somatic mutations remains a challenge. Computational methods play an increasingly important role in discovering genomic patterns associated with cancer drivers and developing predictive models to identify these elements. Machine learning (ML), including deep learning, has been the engine behind many of these efforts and provides excellent opportunities for tackling remaining gaps in the field. Thus, this survey aims to perform a comprehensive analysis of ML-based computational approaches to identify cancer driver mutations and genes, providing an integrated, panoramic view of the broad data and algorithmic landscape within this scientific problem. We discuss how the interactions among data types and ML algorithms have been explored in previous solutions and outline current analytical limitations that deserve further attention from the scientific community. We hope that by helping readers become more familiar with significant developments in the field brought by ML, we may inspire new researchers to address open problems and advance our knowledge towards cancer driver discovery.

Ulcerative colitis (UC) is a complex immune-mediated disease in which the gut microbiota plays a central role, and may determine prognosis and disease progression. We aimed to assess whether a specific microbiota profile, as measured by a machine learning approach, can be associated with disease severity in patients with UC. In this prospective pilot study, consecutive patients with active or inactive UC and healthy controls (HCs) were enrolled. Stool samples were collected for fecal microbiota assessment analysis by 16S rRNA gene sequencing approach. A machine learning approach was used to predict the groups’ separation. Thirty-six HCs and forty-six patients with UC (20 active and 26 inactive) were enrolled. Alpha diversity was significantly different between the three groups (Shannon index: p-values: active UC vs HCs = 0.0005; active UC vs inactive UC = 0.0273; HCs vs inactive UC = 0.0260). In particular, patients with active UC showed the lowest values, followed by patients with inactive UC, and HCs. At species level, we found high levels of Bifidobacterium adolescentis and Haemophilus parainfluenzae in inactive UC and active UC, respectively. A specific microbiota profile was found for each group and was confirmed with sparse partial least squares discriminant analysis, a machine learning-supervised approach. The latter allowed us to observe a perfect class prediction and group separation using the complete information (full Operational Taxonomic Unit table), with a minimal loss in performance when using only 5% of features. A machine learning approach to 16S rRNA data identifies a bacterial signature characterizing different degrees of disease activity in UC. Follow-up studies will clarify whether such microbiota profiling are useful for diagnosis and management.

The human gut microbiome is a complex ecosystem that is closely related to the aging process. However, there is currently no reliable method to make full use of the metagenomics data of the gut microbiome to determine the age of the host. In this study, we considered the influence of geographical factors on the gut microbiome, and a total of 2604 filtered metagenomics data from the gut microbiome were used to construct an age prediction model. Then, we developed an ensemble model with multiple heterogeneous algorithms and combined species and pathway profiles for multi-view learning. By integrating gut microbiome metagenomics data and adjusting host confounding factors, the model showed high accuracy (R2 = 0.599, mean absolute error = 8.33 years). Besides, we further interpreted the model and identify potential biomarkers for the aging process. Among these identified biomarkers, we found that Finegoldia magna, Bifidobacterium dentium, and Clostridium clostridioforme had increased abundance in the elderly. Moreover, the utilization of amino acids by the gut microbiome undergoes substantial changes with increasing age which have been reported as the risk factors for age-associated malnutrition and inflammation. This model will be helpful for the comprehensive utilization of multiple omics data, and will allow greater understanding of the interaction between microorganisms and age to realize the targeted intervention of aging.

Objectives To develop and test the performance of computerized ultrasound image analysis using deep neural networks (DNNs) in discriminating between benign and malignant ovarian tumors and to compare its diagnostic accuracy with that of subjective assessment (SA) by an ultrasound expert. Methods We included 3077 (grayscale, n = 1927; power Doppler, n = 1150) ultrasound images from 758 women with ovarian tumors, who were classified prospectively by expert ultrasound examiners according to IOTA (International Ovarian Tumor Analysis) terms and definitions. Histological outcome from surgery (n = 634) or long‐term (≥ 3 years) follow‐up (n = 124) served as the gold standard. The dataset was split into a training set (n = 508; 314 benign and 194 malignant), a validation set (n = 100; 60 benign and 40 malignant) and a test set (n = 150; 75 benign and 75 malignant). We used transfer learning on three pre‐trained DNNs: VGG16, ResNet50 and MobileNet. Each model was trained, and the outputs calibrated, using temperature scaling. An ensemble of the three models was then used to estimate the probability of malignancy based on all images from a given case. The DNN ensemble classified the tumors as benign or malignant (Ovry‐Dx1 model); or as benign, inconclusive or malignant (Ovry‐Dx2 model). The diagnostic performance of the DNN models, in terms of sensitivity and specificity, was compared to that of SA for classifying ovarian tumors in the test set. Results At a sensitivity of 96.0%, Ovry‐Dx1 had a specificity similar to that of SA (86.7% vs 88.0%; P = 1.0). Ovry‐Dx2 had a sensitivity of 97.1% and a specificity of 93.7%, when designating 12.7% of the lesions as inconclusive. By complimenting Ovry‐Dx2 with SA in inconclusive cases, the overall sensitivity (96.0%) and specificity (89.3%) were not significantly different from using SA in all cases (P = 1.0). Conclusion Ultrasound image analysis using DNNs can predict ovarian malignancy with a diagnostic accuracy comparable to that of human expert examiners, indicating that these models may have a role in the triage of women with an ovarian tumor. © 2020 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology., This article's abstract has been translated into Spanish and Chinese. Follow the links from the abstract to view the translations.

Objective To validate a previously published machine learning model of delirium risk in hospitalized patients with coronavirus disease 2019 (COVID-19). Method Using data from six hospitals across two academic medical networks covering care occurring after initial model development, we calculated the predicted risk of delirium using a previously developed risk model applied to diagnostic, medication, laboratory, and other clinical features available in the electronic health record (EHR) at time of hospital admission. We evaluated the accuracy of these predictions against subsequent delirium diagnoses during that admission. Results Of the 5102 patients in this cohort, 716 (14%) developed delirium. The model's risk predictions produced a c-index of 0.75 (95% CI, 0.73–0.77) with 27.7% of cases occurring in the top decile of predicted risk scores. Model calibration was diminished compared to the initial COVID-19 wave. Conclusion This EHR delirium risk prediction model, developed during the initial surge of COVID-19 patients, produced consistent discrimination over subsequent larger waves; however, with changing cohort composition and delirium occurrence rates, model calibration decreased. These results underscore the importance of calibration, and the challenge of developing risk models for clinical contexts where standard of care and clinical populations may shift.

Machine learning has been widely used in the characterization of tumors recently. This article aims to explore the feasibility of the whole tumor fat-suppressed (FS) T2WI and ADC features-based least absolute shrinkage and selection operator (LASSO)-logistic predictive models in the differentiation of soft tissue neoplasms (STN). The clinical and MR findings of 160 cases with 161 histologically proven STN were reviewed, retrospectively, 75 with diffusion-weighted imaging (DWI with b values of 50, 400, and 800 s/mm2). They were divided into benign and malignant groups and further divided into training (70%) and validation (30%) cohorts. The MR FS T2WI and ADC features-based LASSO-logistic models were built and compared. The AUC of the FS T2WI features-based LASSO-logistic regression model for benign and malignant prediction was 0.65 and 0.75 for the training and validation cohorts. The model’s sensitivity, specificity, and accuracy of the validation cohort were 55%, 96%, and 76.6%. While the AUC of the ADC features-based model was 0.932 and 0.955 for the training and validation cohorts. The model’s sensitivity, specificity, and accuracy were 83.3%, 100%, and 91.7%. The performances of these models were also validated by decision curve analysis (DCA). The AUC of the whole tumor ADC features-based LASSO-logistic regression predictive model was larger than that of FS T2WI features (p = 0.017). The whole tumor fat-suppressed T2WI and ADC features-based LASSO-logistic predictive models both can serve as useful tools in the differentiation of STN. ADC features-based LASSO-logistic regression predictive model did better than that of FS T2WI features.

A majority of breast cancers (BC) are age-related and we seek to determine what cellular and molecular changes occur in breast tissue with age that make women more susceptible to cancer initiation. Immune-epithelial cell interactions are important during mammary gland development and the immune system plays an important role in BC progression. The composition of human immune cell populations is known to change in peripheral blood with age and in breast tissue during BC progression. Less is known about changes in immune populations in normal breast tissue and how their interactions with mammary epithelia change with age. We quantified densities of T cells, B cells, and macrophage subsets in pathologically normal breast tissue from 122 different women who ranged in age from 24 to 74 years old. Donor-matched peripheral blood from a subset of 20 donors was analyzed by flow cytometry. Tissue immune cell densities and localizations relative to the epithelium were quantified in situ with machine learning-based image analyses of multiplex immunohistochemistry-stained tissue sections. In situ results were corroborated with flow cytometry analyses of peri-epithelial immune cells from primary breast tissue preparations and transcriptome analyses of public data from bulk tissue reduction mammoplasties. Proportions of immune cell subsets in breast tissue and donor-matched peripheral blood were not correlated. Density (cells/mm2) of T and B lymphocytes in situ decreased with age. T cells and macrophages preferentially localized near or within epithelial bilayers, rather than the intralobular stroma. M2 macrophage density was higher than M1 macrophage density and this difference was due to higher density of M2 in the intralobular stroma. Transcriptional signature analyses suggested age-dependent decline in adaptive immune cell populations and functions and increased innate immune cell activity. T cells and macrophages are so intimately associated with the epithelia that they are embedded within the bilayer, suggesting an important role for immune-epithelial cell interactions. Age-associated decreased T cell density in peri-epithelial regions, and increased M2 macrophage density in intralobular stroma suggests the emergence of a tissue microenvironment that is simultaneously immune-senescent and immunosuppressive with age. Supplementary Information The online version contains supplementary material available at 10.1007/s10911-021-09495-2.

This study aims to explore biomarkers associated with vitiligo and analyze the pathological role of immune cell infiltration in the disease. We used the robust rank aggregation (RRA) method to integrate three vitiligo data sets downloaded from gene expression omnibus database, identify the differentially expressed genes (DEGs) and analyze the functional correlation. Then, the comprehensive strategy of combined weighted gene coexpression network analysis (WGCNA) and logical regression of the selection operator (LASSO), support vector machine recursive feature elimination (SVM-RFE), and random forest (RF) machine learning algorithm are employed to screen and biomarkers associated with vitiligo. Finally, the immune cell infiltration of vitiligo was evaluated by CIBERSORT, and the correlation between biomarkers and infiltrating immune cells was analyzed. Herein, we identified 131 robust DEGs, and enrichment analysis results showed that robust DEGs and melanogenesis were closely associated with vitiligo development and progression. TYR, TYRP1, DCT and LARP7 were identified as vitiligo-related biomarkers. Immune infiltration analysis demonstrated that CD4 T Cell, CD8 T Cell, Tregs, NK cells, dendritic cells, and macrophages were involved in vitiligo’s pathogenesis. In summary, we adopted a comprehensive strategy to screen biomarkers related to vitiligo and explore the critical role of immune cell infiltration in vitiligo. Abbreviations: TYR, Tyrosinase; TYRP1, Tyrosinase-related protein-1; DCT, dopachrome tautomerase; LARP7, La ribonucleoprotein domain family, member-7; RRA, robust rank aggregation; DEGs, differentially expressed genes; WGCNA, weighted gene coexpression network analysis; LASSO, logical regression of the selection operator; SVM-RFE, support vector machine recursive feature elimination; RF, random forest; GWAS, Genome-wide association study; FasL, Fas-Fas ligand; Tregs, T-regulatory cells; NK, natural killer; GEPCs, gene expression profiling chips; GO, gene ontology; GSEA, gene set enrichment analysis; FDR, false discovery rate; AUC, area under the curve; ROC, receiver-operating characteristic; BP, biological process; CC, cellular component; MF, molecular function., GraphicalAbstract

Backgrounds SARS‐CoV‐2 is affecting different countries all over the world, with significant variation in infection‐rate and death‐ratio. We have previously shown a presence of a possible relationship between different variables including the Bacillus Calmette–Guérin (BCG) vaccine, average age, gender, and malaria treatment, and the rate of spread, severity and mortality of COVID‐19 disease. This paper focuses on developing machine learning models for this relationship. Methods We have used real‐datasets collected from the Johns Hopkins University Center for Systems Science and Engineering and the European Centre for Disease Prevention and Control to develop a model from China data as the baseline country. From this model, we predicted and forecasted different countries' daily confirmed‐cases and daily death‐cases and examined if there was any possible effect of the variables mentioned above. Results The model was trained based on China data as a baseline model for daily confirmed‐cases and daily death‐cases. This machine learning application succeeded in modelling and forecasting daily confirmed‐cases and daily death‐cases. The modelling and forecasting of viral spread resulted in four different regions; these regions were dependent on the malarial treatments, BCG vaccination, weather conditions, and average age. However, the lack of social distancing resulted in variation in the effect of these factors, for example, double‐humped spread and mortality cases curves and sudden increases in the spread and mortality cases in different countries. The process of machine learning for time‐series prediction and forecasting, especially in the pandemic COVID‐19 domain, proved usefulness in modelling and forecasting the end status of the virus spreading based on specific regional and health support variables. Conclusion From the experimental results, we confirm that COVID‐19 has a very low spread in the African countries with all the four variables (average young age, hot weather, BCG vaccine and malaria treatment); a very high spread in European countries and the USA with no variable (old people, cold weather, no BCG vaccine and no malaria). The effect of the variables could be on the spread or the severity to the extent that the infected subject might not have symptoms or the case is mild and can be missed as a confirmed‐case. Social distancing decreases the effect of these factors.

Motivation Whole slide tissue images contain detailed data on the sub-cellular structure of cancer. Quantitative analyses of this data can lead to novel biomarkers for better cancer diagnosis and prognosis and can improve our understanding of cancer mechanisms. Such analyses are challenging to execute because of the sizes and complexity of whole slide image data and relatively limited volume of training data for machine learning methods. Results We propose and experimentally evaluate a multi-resolution deep learning method for breast cancer survival analysis. The proposed method integrates image data at multiple resolutions and tumor, lymphocyte and nuclear segmentation results from deep learning models. Our results show that this approach can significantly improve the deep learning model performance compared to using only the original image data. The proposed approach achieves a c-index value of 0.706 compared to a c-index value of 0.551 from an approach that uses only color image data at the highest image resolution. Furthermore, when clinical features (sex, age and cancer stage) are combined with image data, the proposed approach achieves a c-index of 0.773. Availability and implementation https://github.com/SBU-BMI/deep_survival_analysis

Background Decision curve analysis is a method to evaluate prediction models and diagnostic tests that was introduced in a 2006 publication. Decision curves are now commonly reported in the literature, but there remains widespread misunderstanding of and confusion about what they mean. Summary of commentary In this paper, we present a didactic, step-by-step introduction to interpreting a decision curve analysis and answer some common questions about the method. We argue that many of the difficulties with interpreting decision curves can be solved by relabeling the y-axis as “benefit” and the x-axis as “preference.” A model or test can be recommended for clinical use if it has the highest level of benefit across a range of clinically reasonable preferences. Conclusion Decision curves are readily interpretable if readers and authors follow a few simple guidelines.

Although DNA methylation variation of autosomal CpGs provides robust age predictive biomarkers, no male-specific age predictor exists based on Y-CpGs yet. Since sex chromosomes play an important role in aging, a Y-chromosome-based age predictor would allow studying male-specific aging effects and would also be useful in forensics. Here, we used blood-based DNA methylation microarray data of 1,057 males from six cohorts aged 15-87 and identified 75 Y-CpGs with an interquartile range of ≥0.1. Of these, 22 and six were significantly hyper- and hypomethylated with age (p(cor)

Rationale: This paper aims to show how the focus on eradicating bias from Machine Learning decision-support systems in medical diagnosis diverts attention from the hermeneutic nature of medical decision-making and the productive role of bias. We want to show how an introduction of Machine Learning systems alters the diagnostic process. Reviewing the negative conception of bias and incorporating the mediating role of Machine Learning systems in the medical diagnosis are essential for an encompassing, critical and informed medical decision-making. Methods: This paper presents a philosophical analysis, employing the conceptual frameworks of hermeneutics and technological mediation, while drawing on the case of Machine Learning algorithms assisting doctors in diagnosis. This paper unravels the non-neutral role of algorithms in the doctor's decision-making and points to the dialogical nature of interaction not only with the patients but also with the technologies that co-shape the diagnosis. Findings: Following the hermeneutical model of medical diagnosis, we review the notion of bias to show how it is an inalienable and productive part of diagnosis. We show how Machine Learning biases join the human ones to actively shape the diagnostic process, simultaneously expanding and narrowing medical attention, highlighting certain aspects, while disclosing others, thus mediating medical perceptions and actions. Based on that, we demonstrate how doctors can take Machine Learning systems on board for an enhanced medical diagnosis, while being aware of their non-neutral role. Conclusions: We show that Machine Learning systems join doctors and patients in co-designing a triad of medical diagnosis. We highlight that it is imperative to examine the hermeneutic role of the Machine Learning systems. Additionally, we suggest including not only the patient, but also colleagues to ensure an encompassing diagnostic process, to respect its inherently hermeneutic nature and to work productively with the existing human and machine biases.

AimsAtrial fibrillation (AF) is sustained by re-entrant activation patterns. Ablation strategies have been proposed that target regions of tissue that may support re-entrant activation patterns. We aimed to characterize the tissue properties associated with regions that tether re-entrant activation patterns in a validated virtual patient cohort.Methods and resultsAtrial fibrillation patient-specific models (seven paroxysmal and three persistent) were generated and validated against local activation time (LAT) measurements during an S1–S2 pacing protocol from the coronary sinus and high right atrium, respectively. Atrial models were stimulated with burst pacing from three locations in the proximity of each pulmonary vein to initiate re-entrant activation patterns. Five atria exhibited sustained activation patterns for at least 80 s. Models with short maximum action potential durations (APDs) were associated with sustained activation. Phase singularities were mapped across the atria sustained activation patterns. Regions with a low maximum conduction velocity (CV) were associated with tethering of phase singularities. A support vector machine (SVM) was trained on maximum local conduction velocity and action potential duration to identify regions that tether phase singularities. The SVM identified regions of tissue that could support tethering with 91% accuracy. This accuracy increased to 95% when the SVM was also trained on surface area.ConclusionIn a virtual patient cohort, local tissue properties, that can be measured (CV) or estimated (APD; using effective refractory period as a surrogate) clinically, identified regions of tissue that tether phase singularities. Combing CV and APD with atrial surface area further improved the accuracy in identifying regions that tether phase singularities.

Early risk assessments and interventions for metabolic syndrome (MetS) are limited because of a lack of effective biomarkers. In the present study, several candidate genes were selected as a blood-based transcriptomic signature for MetS. We collected so far the largest MetS-associated peripheral blood high-throughput transcriptomics data and put forward a novel feature selection strategy by combining weighted gene co-expression network analysis, protein-protein interaction network analysis, LASSO regression and random forest approaches. Two gene modules and 51 hub genes as well as a 9-hub-gene signature associated with metabolic syndrome were identified. Then, based on this 9-hub-gene signature, we performed logistic analysis and subsequently established a web nomogram calculator for metabolic syndrome risk (https://xjtulgz.shinyapps.io/DynNomapp/). This 9-hub-gene signature showed excellent classification and calibration performance (AUC = 0.968 in training set, AUC = 0.883 in internal validation set, AUC = 0.861 in external validation set) as well as ideal potential clinical benefit.

Rationale: Platinum-based chemotherapy is one of treatment mainstay for patients with advanced lung squamous cell carcinoma (LUSC) but it is still a "one-size fits all" approach. Here, we aimed to investigate the predictive and monitoring role of circulating cell-free DNA (cfDNA) profiling for the outcome of first-line chemotherapy in patients with advanced LUSC. Methods: Peripheral blood samples of 155 patients from a phase IV trial and 42 cases from an external real-world cohort were prospectively collected. We generated a copy number variations-based classifier via machine learning algorithm to integrate molecular profiling of cfDNA, named RESPONSE SCORE (RS) to predict the treatment outcome. To monitor the treatment efficacy, cfDNA samples collected at different time points were subjected to an ultra-deep sequencing platform. Results: The results showed that patients with high RS showed substantially higher objective response rate than those with low RS in training set (P < 0.001), validation set (P < 0.001) and real-world cohort (P = 0.019). Furthermore, a significant difference was observed in both progression-free survival (training set, P < 0.001; validation set: P < 0.001; real-world cohort: P = 0.019) and overall survival (training set, P < 0.001; validation set: P = 0.037) between high and low RS group. Notably, variant allele frequency (VAF) calculated from an ultra-deep sequencing platform significantly reduced in patients experienced a complete or partial response after 2 cycles of chemotherapy (P < 0.001), while it significantly increased in these of non-responder (P < 0.001). Moreover, VAF undetectable after 2 cycles of chemotherapy was correlated with markedly better objective response rate (P < 0.001) and progression-free survival (P < 0.001) than those with detectable VAF. Conclusions: These findings indicated that the RS, a circulating cfDNA sequencing-based stratification index, could help to guide first-line chemotherapy in advanced LUSC. The change of VAF is valuable to monitor the treatment response.

Cardiac remodeling is recognized as an important aspect of cardiovascular disease (CVD) progression. Machine learning (ML) techniques were applied to basic clinical parameters and electrocardiographic features, in order to detect abnormal left ventricular geometry (LVG) even before the onset of left ventricular hypertrophy (LVH), in a population without established CVD. The authors enrolled 528 patients with and without essential hypertension, but no other indications of CVD. All patients underwent a full echocardiographic evaluation and were classified into 3 groups; normal geometry (NG), concentric remodeling without LVH (CR), and LVH. Abnormal LVG was identified as increased relative wall thickness (RWT) and/or left ventricular mass index (LVMi). The authors trained supervised ML models to classify patients with abnormal LVG and calculated SHAP values to perform feature importance and interaction analysis. Hypertension, age, body mass index over the Sokolow‐Lyon voltage, QRS‐T angle, and QTc duration were some of the most important features. Our model was able to distinguish NG from CR+LVH combined, with 87% accuracy on an unseen test set, 75% specificity, 97% sensitivity, and area under the receiver operating curve (AUC/ROC) equal to 0.91. The authors also trained our model to classify NG and CR (NG + CR) against those with LVH, with 89% test set accuracy, 93% specificity, 67% sensitivity, and an AUC/ROC value of 0.89, for a 0.4 decision threshold. Our ML algorithm effectively detects abnormal LVG even at early stages. Innovative solutions are needed to improve risk stratification of patients without established CVD, and ML may enable progress in this direction., We used machine learning techniques and found a combination of clinical and ECG features that enable prediction of abnormal left ventricular geometry or left ventricular hypertrophy (LVH). Concentric remodeling appears to be separable from LVH through a random forest method.

Early detection of diseases is critical in infants. This study evaluates the usefulness of web searches in predicting diseases in order to encourage guardians to consult a doctor promptly if their children are ill. We collected six months of search queries from Yahoo! JAPAN Search between October 2016 and March 2017. Using a machine learning model, we investigated the accuracy of the search query’s ability to predict the diagnosis of biliary atresia and hypertrophic pyloric stenosis. Both diseases were modeled with an accuracy of approximately 80%, and symptoms related to the disease were significant features in the model. These findings suggest the possibility of detecting diseases from web search queries performed by guardians. Through future research, we intend to propose a method that uses web search queries for early detection of these diseases by providing appropriate and timely information to support the guardians of patients.

Barriers to identifying autism spectrum disorder (ASD) in young children in a timely manner have led to calls for novel screening and assessment strategies. Combining computational methods with clinical expertise presents an opportunity for identifying patterns within large clinical datasets that can inform new assessment paradigms. The present study describes an analytic approach used to identify key features predictive of ASD in young children, drawn from large amounts of data from comprehensive diagnostic evaluations. A team of expert clinicians used these predictive features to design a set of assessment activities allowing for observation of these core behaviors. The resulting brief assessment underlies several novel approaches to the identification of ASD that are the focus of ongoing research.

Introduction While the visual inspection of histopathology images by expert pathologists remains the golden standard method for grading of prostate cancer the quest for developing automated algorithms for the job is set and deep-learning techniques have emerged on top of other approaches. Methods Two pre-trained deep-learning networks, obtained with transfer learning from two general purpose classification networks - AlexNet and GoogleNet, originally trained on a proprietary dataset of prostate cancer were used to classify 6000 cropped images from Gleason2019 Challenge. Results The average agreement between the two networks and the six pathologists was found to be substantial for AlexNet and moderate for GoogleNet. When tested against the majority vote of the six pathologists the agreement was perfect and moderate for AlexNet, and GoogleNet, respectively. Despite our expectations, the average inter-pathologist agreement was moderate, while between the two networks it was substantial. Resulted accuracy for AlexNet and GoogleNet when tested against the majority vote as ground truth was of 85.51% and 74.75%, respectively. This result was higher than the score obtained on the dataset that they were trained on, showing their generalization capabilities. Conclusions Both the agreement and the accuracy indicate a better performance of AlexNet over GoogleNet, making it suitable for clinical deployment thus could potentially contribute to faster, more accurate and with higher reproducibility prostate cancer diagnosis.

Approximately 30% of medulloblastoma (MB) patients exhibit metastasis at initial diagnosis, which often leads to a poor prognosis. Here, by using univariate Cox regression analysis, two machine learning methods (Lasso-penalized Cox regression and random survival forest-variable hunting (RSF-VH)), and multivariate Cox regression analysis, we established two metastasis-related prognostic models, including the 47-mRNA-based model based on the Lasso method and the 21-mRNA-based model based on the RSF-VH method. In terms of the results of the receiver operating characteristic (ROC) curve analyses, we selected the 47-mRNA metastasis-associated model with the higher area under the curve (AUC). The 47-mRNA-based prognostic model could classify MB patients into two subgroups with different prognoses. The ROC analyses also suggested that the 47-mRNA metastasis-associated model may have a better predictive ability than MB subgroup. Multivariable Cox regression analysis demonstrated that the 47-mRNA-based model was independent of other clinical characteristics. In addition, a nomogram comprising the 47-mRNA-based model was built. The results of ROC analyses suggested that the nomogram had good discrimination ability. Our 47-mRNA metastasis-related prognostic model and nomogram might be an efficient and valuable tool for overall survival (OS) prediction and provide information for individualized treatment decisions in patients with MB.

The large number of available MRI sequences means patients cannot realistically undergo them all, so the range of sequences to be acquired during a scan are protocolled based on clinical details. Adapting this to unexpected findings identified early on in the scan requires experience and vigilance. We investigated whether deep learning of the images acquired in the first few minutes of a scan could provide an automated early alert of abnormal features. Anatomy sequences from 375 CMR scans were used as a training set. From these, we annotated 1500 individual slices and used these to train a convolutional neural network to perform automatic segmentation of the cardiac chambers, great vessels and any pleural effusions. 200 scans were used as a testing set. The system then assembled a 3D model of the thorax from which it made clinical measurements to identify important abnormalities. The system was successful in segmenting the anatomy slices (Dice 0.910) and identified multiple features which may guide further image acquisition. Diagnostic accuracy was 90.5% and 85.5% for left and right ventricular dilatation, 85% for left ventricular hypertrophy and 94.4% for ascending aorta dilatation. The area under ROC curve for diagnosing pleural effusions was 0.91. We present proof-of-concept that a neural network can segment and derive accurate clinical measurements from a 3D model of the thorax made from transaxial anatomy images acquired in the first few minutes of a scan. This early information could lead to dynamic adaptive scanning protocols, and by focusing scanner time appropriately and prioritizing cases for supervision and early reporting, improve patient experience and efficiency. Electronic supplementary material The online version of this article (10.1007/s10554-020-02050-w) contains supplementary material, which is available to authorized users.

Motivation There is growing interest in the biomedical research community to incorporate retrospective data, available in healthcare systems, to shed light on associations between different biomarkers. Understanding the association between various types of biomedical data, such as genetic, blood biomarkers, imaging, etc. can provide a holistic understanding of human diseases. To formally test a hypothesized association between two types of data in Electronic Health Records (EHRs), one requires a substantial sample size with both data modalities to achieve a reasonable power. Current association test methods only allow using data from individuals who have both data modalities. Hence, researchers cannot take advantage of much larger EHR samples that includes individuals with at least one of the data types, which limits the power of the association test. Results We present a new method called the Semi-paired Association Test (SAT) that makes use of both paired and unpaired data. In contrast to classical approaches, incorporating unpaired data allows SAT to produce better control of false discovery and to improve the power of the association test. We study the properties of the new test theoretically and empirically, through a series of simulations and by applying our method on real studies in the context of Chronic Obstructive Pulmonary Disease. We are able to identify an association between the high-dimensional characterization of Computed Tomography chest images and several blood biomarkers as well as the expression of dozens of genes involved in the immune system. Availability and implementation Code is available on https://github.com/batmanlab/Semi-paired-Association-Test. Supplementary information Supplementary data are available at Bioinformatics online.

Background The current COVID-19 pandemic is unprecedented; under resource-constrained settings, predictive algorithms can help to stratify disease severity, alerting physicians of high-risk patients; however, there are only few risk scores derived from a substantially large electronic health record (EHR) data set, using simplified predictors as input. Objective The objectives of this study were to develop and validate simplified machine learning algorithms that predict COVID-19 adverse outcomes; to evaluate the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and calibration of the algorithms; and to derive clinically meaningful thresholds. Methods We performed machine learning model development and validation via a cohort study using multicenter, patient-level, longitudinal EHRs from the Optum COVID-19 database that provides anonymized, longitudinal EHR from across the United States. The models were developed based on clinical characteristics to predict 28-day in-hospital mortality, intensive care unit (ICU) admission, respiratory failure, and mechanical ventilator usages at inpatient setting. Data from patients who were admitted from February 1, 2020, to September 7, 2020, were randomly sampled into development, validation, and test data sets; data collected from September 7, 2020, to November 15, 2020, were reserved as the postdevelopment prospective test data set. Results Of the 3.7 million patients in the analysis, 585,867 patients were diagnosed or tested positive for SARS-CoV-2, and 50,703 adult patients were hospitalized with COVID-19 between February 1 and November 15, 2020. Among the study cohort (n=50,703), there were 6204 deaths, 9564 ICU admissions, 6478 mechanically ventilated or EMCO patients, and 25,169 patients developed acute respiratory distress syndrome or respiratory failure within 28 days since hospital admission. The algorithms demonstrated high accuracy (AUC 0.89, 95% CI 0.89-0.89 on the test data set [n=10,752]), consistent prediction through the second wave of the pandemic from September to November (AUC 0.85, 95% CI 0.85-0.86) on the postdevelopment prospective test data set [n=14,863], great clinical relevance, and utility. Besides, a comprehensive set of 386 input covariates from baseline or at admission were included in the analysis; the end-to-end pipeline automates feature selection and model development. The parsimonious model with only 10 input predictors produced comparably accurate predictions; these 10 predictors (age, blood urea nitrogen, SpO2, systolic and diastolic blood pressures, respiration rate, pulse, temperature, albumin, and major cognitive disorder excluding stroke) are commonly measured and concordant with recognized risk factors for COVID-19. Conclusions The systematic approach and rigorous validation demonstrate consistent model performance to predict even beyond the period of data collection, with satisfactory discriminatory power and great clinical utility. Overall, the study offers an accurate, validated, and reliable prediction model based on only 10 clinical features as a prognostic tool to stratifying patients with COVID-19 into intermediate-, high-, and very high-risk groups. This simple predictive tool is shared with a wider health care community, to enable service as an early warning system to alert physicians of possible high-risk patients, or as a resource triaging tool to optimize health care resources.

Background: Mobile phone–delivered life skills programs are an emerging and promising way to promote mental health and prevent substance use among adolescents, but little is known about how adolescents actually use them. Objective: The aim of this study is to determine engagement with a mobile phone–based life skills program and its different components, as well as the associations of engagement with adolescent characteristics and intended substance use and mental health outcomes. Methods: We performed secondary data analysis on data from the intervention group (n=750) from a study that compared a mobile phone–based life skills intervention for adolescents recruited in secondary and upper secondary school classes with an assessment-only control group. Throughout the 6-month intervention, participants received 1 SMS text message prompt per week that introduced a life skills topic or encouraged participation in a quiz or individual life skills training or stimulated sharing messages with other program participants through a friendly contest. Decision trees were used to identify predictors of engagement (use and subjective experience). The stability of these decision trees was assessed using a resampling method and by graphical representation. Finally, associations between engagement and intended substance use and mental health outcomes were examined using logistic and linear regression analyses. Results: The adolescents took part in half of the 50 interactions (mean 23.6, SD 15.9) prompted by the program, with SMS text messages being the most used and contests being the least used components. Adolescents who did not drink in a problematic manner and attended an upper secondary school were the ones to use the program the most. Regarding associations between engagement and intended outcomes, adolescents who used the contests more frequently were more likely to be nonsmokers at follow-up than those who did not (odds ratio 0.86, 95% CI 0.76-0.98; P=.02). In addition, adolescents who read the SMS text messages more attentively were less likely to drink in a problematic manner at follow-up (odds ratio 0.43, 95% CI 1.29-3.41; P=.003). Finally, participants who used the program the most and least were more likely to increase their well-being from baseline to 6-month follow-up compared with those with average engagement (βs=.39; t586=2.66; P=.008; R2=0.24). Conclusions: Most of the adolescents participating in a digital life skills program that aimed to prevent substance use and promote mental health engaged with the intervention. However, measures to increase engagement in problem drinkers should be considered. Furthermore, efforts must be made to ensure that interventions are engaging and powerful across different educational levels. First results indicate that higher engagement with digital life skills programs could be associated with intended outcomes. Future studies should apply further measures to improve the reach of lower-engaged participants at follow-up to establish such associations with certainty., Journal of Medical Internet Research, 24 (1), ISSN:1438-8871

The benefit of regular physical activity and exercise training for the prevention of cardiovascular and metabolic diseases is undisputed. Many molecular mechanisms mediating exercise effects have been deciphered. Personalised exercise prescription can help patients in achieving their individual greatest benefit from an exercise-based cardiovascular rehabilitation programme. Yet, we still struggle to provide truly personalised exercise prescriptions to our patients. In this position paper, we address novel basic and translational research concepts that can help us understand the principles underlying the inter-individual differences in the response to exercise, and identify early on who would most likely benefit from which exercise intervention. This includes hereditary, non-hereditary and sex-specific concepts. Recent insights have helped us to take on a more holistic view, integrating exercise-mediated molecular mechanisms with those influenced by metabolism and immunity. Unfortunately, while the outline is recognisable, many details are still lacking to turn the understanding of a concept into a roadmap ready to be used in clinical routine. This position paper therefore also investigates perspectives on how the advent of 'big data' and the use of animal models could help unravel inter-individual responses to exercise parameters and thus influence hypothesis-building for translational research in exercise-based cardiovascular rehabilitation. The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: TSB is supported by a Medical Research Council UK New Investigator award (MR/S025472/1). EMVC is supported by Fund for Scientific Research Flanders (Senior Clinical Investigator fellowship) and Koning Boudewijnstichting (Fund Joseph Oscar Waldmann-Berteau 2015). NK receives project-specific funding from the German Centre for Cardiovascular Research (DZHK; 81X2100238 and 81X2100243), the German Foundation of Heart Research (F/39/17) and the German Diabetes Foundation (FP-0421-2018). ABG, VA, MB, VC, MD, DH, HMCK and PL do not receive funding pertinent to this work.

Background There is a strong demand for an accurate and objective means of assessing acute pain among hospitalized patients to help clinicians provide pain medications at a proper dosage and in a timely manner. Heart rate variability (HRV) comprises changes in the time intervals between consecutive heartbeats, which can be measured through acquisition and interpretation of electrocardiography (ECG) captured from bedside monitors or wearable devices. As increased sympathetic activity affects the HRV, an index of autonomic regulation of heart rate, ultra–short-term HRV analysis can provide a reliable source of information for acute pain monitoring. In this study, widely used HRV time and frequency domain measurements are used in acute pain assessments among postoperative patients. The existing approaches have only focused on stimulated pain in healthy subjects, whereas, to the best of our knowledge, there is no work in the literature building models using real pain data and on postoperative patients. Objective The objective of our study was to develop and evaluate an automatic and adaptable pain assessment algorithm based on ECG features for assessing acute pain in postoperative patients likely experiencing mild to moderate pain. Methods The study used a prospective observational design. The sample consisted of 25 patient participants aged 18 to 65 years. In part 1 of the study, a transcutaneous electrical nerve stimulation unit was employed to obtain baseline discomfort thresholds for the patients. In part 2, a multichannel biosignal acquisition device was used as patients were engaging in non-noxious activities. At all times, pain intensity was measured using patient self-reports based on the Numerical Rating Scale. A weak supervision framework was inherited for rapid training data creation. The collected labels were then transformed from 11 intensity levels to 5 intensity levels. Prediction models were developed using 5 different machine learning methods. Mean prediction accuracy was calculated using leave-one-out cross-validation. We compared the performance of these models with the results from a previously published research study. Results Five different machine learning algorithms were applied to perform a binary classification of baseline (BL) versus 4 distinct pain levels (PL1 through PL4). The highest validation accuracy using 3 time domain HRV features from a BioVid research paper for baseline versus any other pain level was achieved by support vector machine (SVM) with 62.72% (BL vs PL4) to 84.14% (BL vs PL2). Similar results were achieved for the top 8 features based on the Gini index using the SVM method, with an accuracy ranging from 63.86% (BL vs PL4) to 84.79% (BL vs PL2). Conclusions We propose a novel pain assessment method for postoperative patients using ECG signal. Weak supervision applied for labeling and feature extraction improves the robustness of the approach. Our results show the viability of using a machine learning algorithm to accurately and objectively assess acute pain among hospitalized patients. International Registered Report Identifier (IRRID) RR2-10.2196/17783

Background Accurate, objective pain assessment is required in the health care domain and clinical settings for appropriate pain management. Automated, objective pain detection from physiological data in patients provides valuable information to hospital staff and caregivers to better manage pain, particularly for patients who are unable to self-report. Galvanic skin response (GSR) is one of the physiologic signals that refers to the changes in sweat gland activity, which can identify features of emotional states and anxiety induced by varying pain levels. This study used different statistical features extracted from GSR data collected from postoperative patients to detect their pain intensity. To the best of our knowledge, this is the first work building pain models using postoperative adult patients instead of healthy subjects. Objective The goal of this study was to present an automatic pain assessment tool using GSR signals to predict different pain intensities in noncommunicative, postoperative patients. Methods The study was designed to collect biomedical data from postoperative patients reporting moderate to high pain levels. We recruited 25 participants aged 23-89 years. First, a transcutaneous electrical nerve stimulation (TENS) unit was employed to obtain patients' baseline data. In the second part, the Empatica E4 wristband was worn by patients while they were performing low-intensity activities. Patient self-report based on the numeric rating scale (NRS) was used to record pain intensities that were correlated with objectively measured data. The labels were down-sampled from 11 pain levels to 5 different pain intensities, including the baseline. We used 2 different machine learning algorithms to construct the models. The mean decrease impurity method was used to find the top important features for pain prediction and improve the accuracy. We compared our results with a previously published research study to estimate the true performance of our models. Results Four different binary classification models were constructed using each machine learning algorithm to classify the baseline and other pain intensities (Baseline [BL] vs Pain Level [PL] 1, BL vs PL2, BL vs PL3, and BL vs PL4). Our models achieved higher accuracy for the first 3 pain models than the BioVid paper approach despite the challenges in analyzing real patient data. For BL vs PL1, BL vs PL2, and BL vs PL4, the highest prediction accuracies were achieved when using a random forest classifier (86.0, 70.0, and 61.5, respectively). For BL vs PL3, we achieved an accuracy of 72.1 using a k-nearest-neighbor classifier. Conclusions We are the first to propose and validate a pain assessment tool to predict different pain levels in real postoperative adult patients using GSR signals. We also exploited feature selection algorithms to find the top important features related to different pain intensities. International Registered Report Identifier (IRRID) RR2-10.2196/17783

As a surrogate for human tactile cognition, an artificial tactile perception and cognition system are proposed to produce smooth/soft and rough tactile sensations by its user's tactile feeling; and named this system as “tactile avatar”. A piezoelectric tactile sensor is developed to record dynamically various physical information such as pressure, temperature, hardness, sliding velocity, and surface topography. For artificial tactile cognition, the tactile feeling of humans to various tactile materials ranging from smooth/soft to rough are assessed and found variation among participants. Because tactile responses vary among humans, a deep learning structure is designed to allow personalization through training based on individualized histograms of human tactile cognition and recording physical tactile information. The decision error in each avatar system is less than 2% when 42 materials are used to measure the tactile data with 100 trials for each material under 1.2N of contact force with 4cm s−1 of sliding velocity. As a tactile avatar, the machine categorizes newly experienced materials based on the tactile knowledge obtained from training data. The tactile sensation showed a high correlation with the specific user's tendency. This approach can be applied to electronic devices with tactile emotional exchange capabilities, as well as advanced digital experiences., This article suggests a tactile avatar system as a surrogate for human tactile cognition producing smooth/soft and rough tactile sensation. Because tactile responses vary among humans, a deep learning network is designed considering the cognitive processing of independent tactile features engaged by humans. As a result, the tactile tendency of a specific user is successfully mimicked by this system.

Summary: Background: Long non-coding RNAs (lncRNAs) have recently emerged as essential biomarkers of cancer progression. However, studies are limited regarding lncRNAs correlated with recurrence and fluorouracil-based adjuvant chemotherapy (ACT) in stage II/III colorectal cancer (CRC). Methods: 1640 stage II/III CRC patients were enrolled from 15 independent datasets and a clinical in-house cohort. 10 prevalent machine learning algorithms were collected and then combined into 76 combinations. 109 published transcriptome signatures were also retrieved. qRT-PCR assay was performed to verify our model. Findings: We comprehensively identified 27 stably recurrence-related lncRNAs from multi-center cohorts. According to these lncRNAs, a consensus machine learning-derived lncRNA signature (CMDLncS) that exhibited best power for predicting recurrence risk was determined from 76 kinds of algorithm combinations. A high CMDLncS indicated unfavorable recurrence and mortality rates. CMDLncS not only could work independently of common clinical traits (e.g., AJCC stage) and molecular features (e.g., microsatellite state, KRAS mutation), but also presented dramatically better performance than these variables. qRT-PCR results from 173 patients further verified our in-silico findings and assessed its feasible in different centers. Comparisons of CMDLncS with 109 published transcriptome signatures further demonstrated its predictive superiority. Additionally, patients with high CMDLncS benefited more from fluorouracil-based ACT and were characterized by activation of stromal and epithelial-mesenchymal transition, while patients with low CMDLncS suggested the sensitivity to bevacizumab and displayed enhanced immune activation. Interpretation: CMDLncS provides an attractive platform for identifying patient at high risk of recurrence and could optimize precision treatment to improve the clinical outcomes in stage II/III CRC. Funding: This study was supported by the National Natural Science Foundation of China (81,972,663); Henan Province Young and Middle‐Aged Health Science and Technology Innovation Talent Project (YXKC2020037); and Henan Provincial Health Commission Joint Youth Project (SB201902014).

OBJECTIVES: The International Society for Bipolar Disorders Big Data Task Force assembled leading researchers in the field of bipolar disorder (BD), machine learning, and big data with extensive experience to evaluate the rationale of machine learning and big data analytics strategies for BD.METHOD: A task force was convened to examine and integrate findings from the scientific literature related to machine learning and big data based studies to clarify terminology and to describe challenges and potential applications in the field of BD. We also systematically searched PubMed, Embase, and Web of Science for articles published up to January 2019 that used machine learning in BD.RESULTS: The results suggested that big data analytics has the potential to provide risk calculators to aid in treatment decisions and predict clinical prognosis, including suicidality, for individual patients. This approach can advance diagnosis by enabling discovery of more relevant data-driven phenotypes, as well as by predicting transition to the disorder in high-risk unaffected subjects. We also discuss the most frequent challenges that big data analytics applications can face, such as heterogeneity, lack of external validation and replication of some studies, cost and non-stationary distribution of the data, and lack of appropriate funding.CONCLUSION: Machine learning-based studies, including atheoretical data-driven big data approaches, provide an opportunity to more accurately detect those who are at risk, parse-relevant phenotypes as well as inform treatment selection and prognosis. However, several methodological challenges need to be addressed in order to translate research findings to clinical settings.

Emphysema is visible on computed tomography (CT) as low-density lesions representing the destruction of the pulmonary alveoli. To train a machine learning model on the emphysema extent in CT images, labeled image data is needed. The provision of these labels requires trained readers, who are a limited resource. The purpose of the study was to test the reading time, inter-observer reliability and validity of the multi-reader–multi-split method for acquiring CT image labels from radiologists. The approximately 500 slices of each stack of lung CT images were split into 1-cm chunks, with 17 thin axial slices per chunk. The chunks were randomly distributed to 26 readers, radiologists and radiology residents. Each chunk was given a quick score concerning emphysema type and severity in the left and right lung separately. A cohort of 102 subjects, with varying degrees of visible emphysema in the lung CT images, was selected from the SCAPIS pilot, performed in 2012 in Gothenburg, Sweden. In total, the readers created 9050 labels for 2881 chunks. Image labels were compared with regional annotations already provided at the SCAPIS pilot inclusion. The median reading time per chunk was 15 s. The inter-observer Krippendorff’s alpha was 0.40 and 0.53 for emphysema type and score, respectively, and higher in the apical part than in the basal part of the lungs. The multi-split emphysema scores were generally consistent with regional annotations. In conclusion, the multi-reader–multi-split method provided reasonably valid image labels, with an estimation of the inter-observer reliability. Electronic supplementary material The online version of this article (10.1007/s10278-020-00378-2) contains supplementary material, which is available to authorized users.

Over the past 2 decades, polypharmacy has become the de-facto standard of acute treatment in psychiatry where patients with psychiatric disorders receive a multiple medication regimen. There is growing evidence for a potential link between major psychiatric disorders and inflammatory processes. Combining these two aspects aims at avoiding polypharmacy attempts among patients with inflammatory activation through alternative treatment strategies. In this study, we addressed the following questions: (1) to what extent can polypharmacy be explained through the factors “diagnosis”, “previous history”, “severity at baseline”, “age”, “gender”, and “psychiatrist in charge”; (2) what are the differences between polypharmacy and monotherapy regarding efficacy and side effect profiles; and (3) what amount of between-patient variance is explainable by the natural antibody immunoglobulin M (IgM) within each diagnostic group. This naturalistic longitudinal study was comprised of 279 patients under therapy with a clinical diagnosis of depressive (ICD-10: “F3x.x”;n = 195) or schizophrenic disorders (ICD-10: “F2x.x”;n = 84). The study protocol included (1) assessment of previous history by the SADS Syndrome Check List SSCL-16 (lifetime version); (2) repeated measurements over 5 weeks assessing the time course of improvement by the Hamilton Depression Scale HAM-D and the Positive and Negative Syndrome Scale PANSS, along with medication and unwanted side effects through the Medication and Side Effects Inventory MEDIS; and (3) the collection of blood samples from which DNA and serum were extracted. The association between inflammatory response system and psychiatric disorders was detailed by fitting multi-layer Neural Net (NN) models to the observed data (“supervised learning”). The same approach was used to set up prediction models of side effects. Our data showed that polypharmacy was omnipresent. Yet the various polypharmacy regimens had no advantage over monotherapy: we even found slightly larger baseline score reductions under monotherapy, independent of primary diagnoses and for comparable baseline severities. Most patients experienced unwanted side effects. The close link between side effects and treatment regimen was revealed by a linear model in which the mere number of drugs explained a significant (p r = 0.746 (p = 0.0004) between global schizophrenia scores and IgM levels, along with a correct prediction of response of 94.4%, thus explaining 55.7% of the observed between-patient variance. For the F3 patients, our NN model identified a 19.6% subgroup exhibiting a significant correlation ofr = 0.644 (p = 0.00003) between global depression scores and IgM levels, along a correct prediction of response of 89.6%, thus explaining 41.4% of the observed between-patient variance. Polypharmacy is omnipresent in today’s acute treatment of psychiatric disorders. Given the large proportion of patients with unwanted side effects and the strong correlation between side effects and the number of drugs, polypharmacy approaches are not equally suited for every patient. In terms of efficacy, there are no advantages of polypharmacy over monotherapy. Most notably, our study appears to have cleared the way for the reliable identification of a subgroup of patients for whom the inflammatory response system is a promising target of therapeutic intervention.

Purpose Accurately forecasting the occurrence of future covid-19-related cases across relaxed (Sweden) and stringent (USA and Canada) policy contexts has a renewed sense of urgency. Moreover, there is a need for a multidimensional county-level approach to monitor the second wave of covid-19 in the USA. Method We use an artificial intelligence framework based on timeline of policy interventions that triangulated results based on the three approaches—Bayesian susceptible-infected-recovered (SIR), Kalman filter, and machine learning. Results Our findings suggest three important insights. First, the effective growth rate of covid-19 infections dropped in response to the approximate dates of key policy interventions. We find that the change points for spreading rates approximately coincide with the timelines of policy interventions across respective countries. Second, forecasted trend until mid-June in the USA was downward trending, stable, and linear. Sweden is likely to be heading in the other direction. That is, Sweden’s forecasted trend until mid-June appears to be non-linear and upward trending. Canada appears to fall somewhere in the middle—the trend for the same period is flat. Third, a Kalman filter based robustness check indicates that by mid-June the USA will likely have close to two million virus cases, while Sweden will likely have over 44,000 covid-19 cases. Conclusion We show that drop in effective growth rate of covid-19 infections was sharper in the case of stringent policies (USA and Canada) but was more gradual in the case of relaxed policy (Sweden). Our study exhorts policy makers to take these results into account as they consider the implications of relaxing lockdown measures.

Aging has pronounced effects on blood laboratory biomarkers used in the clinic. Prior studies have largely investigated one biomarker or population at a time, limiting a comprehensive view of biomarker variation and aging across different populations. Here we develop a supervised machine learning approach to study aging using 356 blood biomarkers measured in 67,563 individuals across diverse populations. Our model predicts age with a mean absolute error (MAE), or average magnitude of prediction errors, in held-out data of 4.76 years and an R2 value of 0.92. Age prediction was highly accurate for the pediatric cohort (MAE = 0.87, R2 = 0.94) but inaccurate for ages 65+ (MAE = 4.30, R2 = 0.25). Variability was observed in which biomarkers carry predictive power across age groups, genders, and race/ethnicity groups, and novel candidate biomarkers of aging were identified for specific age ranges (e.g. Vitamin E, ages 18-44). We show that predictors for one age group may fail to generalize to other groups and investigate non-linearity in biomarkers near adulthood. As populations worldwide undergo major demographic changes, it is increasingly important to catalogue biomarker variation across age groups and discover new biomarkers to distinguish chronological and biological aging.

Motivation Recent developments in technology have enabled researchers to collect multiple OMICS datasets for the same individuals. The conventional approach for understanding the relationships between the collected datasets and the complex trait of interest would be through the analysis of each OMIC dataset separately from the rest, or to test for associations between the OMICS datasets. In this work we show that integrating multiple OMICS datasets together, instead of analysing them separately, improves our understanding of their in-between relationships as well as the predictive accuracy for the tested trait. Several approaches have been proposed for the integration of heterogeneous and high-dimensional (p≫n) data, such as OMICS. The sparse variant of canonical correlation analysis (CCA) approach is a promising one that seeks to penalize the canonical variables for producing sparse latent variables while achieving maximal correlation between the datasets. Over the last years, a number of approaches for implementing sparse CCA (sCCA) have been proposed, where they differ on their objective functions, iterative algorithm for obtaining the sparse latent variables and make different assumptions about the original datasets. Results Through a comparative study we have explored the performance of the conventional CCA proposed by Parkhomenko et al., penalized matrix decomposition CCA proposed by Witten and Tibshirani and its extension proposed by Suo et al. The aforementioned methods were modified to allow for different penalty functions. Although sCCA is an unsupervised learning approach for understanding of the in-between relationships, we have twisted the problem as a supervised learning one and investigated how the computed latent variables can be used for predicting complex traits. The approaches were extended to allow for multiple (more than two) datasets where the trait was included as one of the input datasets. Both ways have shown improvement over conventional predictive models that include one or multiple datasets. Availability and implementation https://github.com/theorod93/sCCA. Supplementary information Supplementary data are available at Bioinformatics online.

Background: Osteocalcin is related to energy metabolism, memory and the acute stress response, suggesting a relationship between bone and the brain. The need to explore the effect of osteocalcin on acute ischemic stroke is therefore urgent. Results: Patients with better outcomes had higher serum osteocalcin levels than those whose NIHSS scores did not improve. Multivariable logistic regression analysis showed acceptable performance (area under the curve = 0.766). The effect of osteocalcin on the promotion of neuron survival was confirmed by Cell Counting Kit-8 experiments. In addition, osteocalcin could decrease proline hydroxylase 1 and inhibit the degradation of gasdermin D. Conclusions: We propose that osteocalcin can improve outcome after acute ischemic stroke in the acute period. By downregulating proline hydroxylase 1, osteocalcin leads glucose metabolism to the pentose phosphate pathway and therefore promotes neuronal survival through inhibiting pyroptosis. Methods: Demographic data and laboratory results were obtained from patients with ischemic stroke in the acute period for analysis. A receiver operating characteristic curve was used to assess the discrimination of the prediction model. The potential effect of osteocalcin on cerebral ischemia and osteocalcin mechanism were explored in cultured primary rat cerebral cortical neurons treated with oxygen-glucose deprivation and reoxygenation.

BackgroundOptimal detection and surveillance of bladder cancer (BCa) rely primarily on the cystoscopic visualization of bladder lesions. AI-assisted cystoscopy may improve image recognition and accelerate data acquisition.ObjectiveTo provide a comprehensive review of machine learning (ML), deep learning (DL) and convolutional neural network (CNN) applications in cystoscopic image recognition.Evidence acquisitionA detailed search of original articles was performed using the PubMed-MEDLINE database to identify recent English literature relevant to ML, DL and CNN applications in cystoscopic image recognition.Evidence synthesisIn total, two articles and one conference abstract were identified addressing the application of AI methods in cystoscopic image recognition. These investigations showed accuracies exceeding 90% for tumor detection; however, future work is necessary to incorporate these methods into AI-aided cystoscopy and compared to other tumor visualization tools. Furthermore, we present results from the RaVeNNA-4pi consortium initiative which has extracted 4200 frames from 62 videos, analyzed them with the U-Net network and achieved an average dice score of 0.67. Improvements in its precision can be achieved by augmenting the video/frame database.ConclusionAI-aided cystoscopy has the potential to outperform urologists at recognizing and classifying bladder lesions. To ensure their real-life implementation, however, these algorithms require external validation to generalize their results across other data sets.

Despite considerable research efforts, pancreatic cancer is associated with a dire prognosis and a 5-year survival rate of only 10%. Early symptoms of the disease are mostly nonspecific. The premise of improved survival through early detection is that more individuals will benefit from potentially curative treatment. Artificial intelligence (AI) methodology has emerged as a successful tool for risk stratification and identification in general health care. In response to the maturity of AI, Kenner Family Research Fund conducted the 2020 AI and Early Detection of Pancreatic Cancer Virtual Summit (www.pdac-virtualsummit.org) in conjunction with the American Pancreatic Association, with a focus on the potential of AI to advance early detection efforts in this disease. This comprehensive presummit article was prepared based on information provided by each of the interdisciplinary participants on one of the 5 following topics: Progress, Problems, and Prospects for Early Detection; AI and Machine Learning; AI and Pancreatic Cancer-Current Efforts; Collaborative Opportunities; and Moving Forward-Reflections from Government, Industry, and Advocacy. The outcome from the robust Summit conversations, to be presented in a future white paper, indicate that significant progress must be the result of strategic collaboration among investigators and institutions from multidisciplinary backgrounds, supported by committed funders.

The purpose of this study was to predict the survival time of patients with malignant glioma after radiotherapy with high accuracy by considering additional clinical factors and optimize the prescription dose and treatment duration for individual patient by using a machine learning model. A total of 35 patients with malignant glioma were included in this study. The candidate features included 12 clinical features and 192 dose–volume histogram (DVH) features. The appropriate input features and parameters of the support vector machine (SVM) were selected using the genetic algorithm based on Akaike’s information criterion, i.e. clinical, DVH, and both clinical and DVH features. The prediction accuracy of the SVM models was evaluated through a leave-one-out cross-validation test with residual error, which was defined as the absolute difference between the actual and predicted survival times after radiotherapy. Moreover, the influences of various values of prescription dose and treatment duration on the predicted survival time were evaluated. The prediction accuracy was significantly improved with the combined use of clinical and DVH features compared with the separate use of both features (P

Background: There is huge variability in the way that individuals with tinnitus respond to interventions. These experiential variations, together with a range of associated etiologies, contribute to tinnitus being a highly heterogeneous condition. Despite this heterogeneity, a “one size fits all” approach is taken when making management recommendations. Although there are various management approaches, not all are equally effective. Psychological approaches such as cognitive behavioral therapy have the most evidence base. Managing tinnitus is challenging due to the significant variations in tinnitus experiences and treatment successes. Tailored interventions based on individual tinnitus profiles may improve outcomes. Predictive models of treatment success are, however, lacking.\ud \ud Objective: This study aimed to use exploratory data mining techniques (ie, decision tree models) to identify the variables associated with the treatment success of internet-based cognitive behavioral therapy (ICBT) for tinnitus.\ud \ud Methods: Individuals (N=228) who underwent ICBT in 3 separate clinical trials were included in this analysis. The primary outcome variable was a reduction of 13 points in tinnitus severity, which was measured by using the Tinnitus Functional Index following the intervention. The predictor variables included demographic characteristics, tinnitus and hearing-related variables, and clinical factors (ie, anxiety, depression, insomnia, hyperacusis, hearing disability, cognitive function, and life satisfaction). Analyses were undertaken by using various exploratory machine learning algorithms to identify the most influencing variables. In total, 6 decision tree models were implemented, namely the classification and regression tree (CART), C5.0, GB, XGBoost, AdaBoost algorithm and random forest models. The Shapley additive explanations framework was applied to the two optimal decision tree models to determine relative predictor importance.\ud \ud Results: Among the six decision tree models, the CART (accuracy: mean 70.7%, SD 2.4%; sensitivity: mean 74%, SD 5.5%; specificity: mean 64%, SD 3.7%; area under the receiver operating characteristic curve [AUC]: mean 0.69, SD 0.001) and gradient boosting (accuracy: mean 71.8%, SD 1.5%; sensitivity: mean 78.3%, SD 2.8%; specificity: 58.7%, SD 4.2%; AUC: mean 0.68, SD 0.02) models were found to be the best predictive models. Although the other models had acceptable accuracy (range 56.3%-66.7%) and sensitivity (range 68.6%-77.9%), they all had relatively weak specificity (range 31.1%-50%) and AUCs (range 0.52-0.62). A higher education level was the most influencing factor for ICBT outcomes. The CART decision tree model identified 3 participant groups who had at least an 85% success probability following the undertaking of ICBT.\ud \ud Conclusions: Decision tree models, especially the CART and gradient boosting models, appeared to be promising in predicting ICBT outcomes. Their predictive power may be improved by using larger sample sizes and including a wider range of predictive factors in future studies.

Background The COVID-19 pandemic has highlighted the inability of health systems to leverage existing system infrastructure in order to rapidly develop and apply broad analytical tools that could inform state- and national-level policymaking, as well as patient care delivery in hospital settings. The COVID-19 pandemic has also led to highlighted systemic disparities in health outcomes and access to care based on race or ethnicity, gender, income-level, and urban-rural divide. Although the United States seems to be recovering from the COVID-19 pandemic owing to widespread vaccination efforts and increased public awareness, there is an urgent need to address the aforementioned challenges. Objective This study aims to inform the feasibility of leveraging broad, statewide datasets for population health–driven decision-making by developing robust analytical models that predict COVID-19–related health care resource utilization across patients served by Indiana’s statewide Health Information Exchange. Methods We leveraged comprehensive datasets obtained from the Indiana Network for Patient Care to train decision forest-based models that can predict patient-level need of health care resource utilization. To assess these models for potential biases, we tested model performance against subpopulations stratified by age, race or ethnicity, gender, and residence (urban vs rural). Results For model development, we identified a cohort of 96,026 patients from across 957 zip codes in Indiana, United States. We trained the decision models that predicted health care resource utilization by using approximately 100 of the most impactful features from a total of 1172 features created. Each model and stratified subpopulation under test reported precision scores >70%, accuracy and area under the receiver operating curve scores >80%, and sensitivity scores approximately >90%. We noted statistically significant variations in model performance across stratified subpopulations identified by age, race or ethnicity, gender, and residence (urban vs rural). Conclusions This study presents the possibility of developing decision models capable of predicting patient-level health care resource utilization across a broad, statewide region with considerable predictive performance. However, our models present statistically significant variations in performance across stratified subpopulations of interest. Further efforts are necessary to identify root causes of these biases and to rectify them.

Motivation Large-scale screenings of cancer cell lines with detailed molecular profiles against libraries of pharmacological compounds are currently being performed in order to gain a better understanding of the genetic component of drug response and to enhance our ability to recommend therapies given a patient's molecular profile. These comprehensive screens differ from the clinical setting in which (i) medical records only contain the response of a patient to very few drugs, (ii) drugs are recommended by doctors based on their expert judgment and (iii) selecting the most promising therapy is often more important than accurately predicting the sensitivity to all potential drugs. Current regression models for drug sensitivity prediction fail to account for these three properties. Results We present a machine learning approach, named Kernelized Rank Learning (KRL), that ranks drugs based on their predicted effect per cell line (patient), circumventing the difficult problem of precisely predicting the sensitivity to the given drug. Our approach outperforms several state-of-the-art predictors in drug recommendation, particularly if the training dataset is sparse, and generalizes to patient data. Our work phrases personalized drug recommendation as a new type of machine learning problem with translational potential to the clinic. Availability and implementation The Python implementation of KRL and scripts for running our experiments are available at https://github.com/BorgwardtLab/Kernelized-Rank-Learning. Supplementary information Supplementary data are available at Bioinformatics online., Bioinformatics, 34 (16), ISSN:1367-4803, ISSN:1460-2059

Background Recently, machine learning (ML) has been transforming our daily lives by enabling intelligent voice assistants, personalized support for purchase decisions, and efficient credit card fraud detection. In addition to its everyday applications, ML holds the potential to improve medicine as well, especially with regard to diagnostics in clinics. In a world characterized by population growth, demographic change, and the global COVID-19 pandemic, ML systems offer the opportunity to make diagnostics more effective and efficient, leading to a high interest of clinics in such systems. However, despite the high potential of ML, only a few ML systems have been deployed in clinics yet, as their adoption process differs significantly from the integration of prior health information technologies given the specific characteristics of ML. Objective This study aims to explore the factors that influence the adoption process of ML systems for medical diagnostics in clinics to foster the adoption of these systems in clinics. Furthermore, this study provides insight into how these factors can be used to determine the ML maturity score of clinics, which can be applied by practitioners to measure the clinic status quo in the adoption process of ML systems. Methods To gain more insight into the adoption process of ML systems for medical diagnostics in clinics, we conducted a qualitative study by interviewing 22 selected medical experts from clinics and their suppliers with profound knowledge in the field of ML. We used a semistructured interview guideline, asked open-ended questions, and transcribed the interviews verbatim. To analyze the transcripts, we first used a content analysis approach based on the health care–specific framework of nonadoption, abandonment, scale-up, spread, and sustainability. Then, we drew on the results of the content analysis to create a maturity model for ML adoption in clinics according to an established development process. Results With the help of the interviews, we were able to identify 13 ML-specific factors that influence the adoption process of ML systems in clinics. We categorized these factors according to 7 domains that form a holistic ML adoption framework for clinics. In addition, we created an applicable maturity model that could help practitioners assess their current state in the ML adoption process. Conclusions Many clinics still face major problems in adopting ML systems for medical diagnostics; thus, they do not benefit from the potential of these systems. Therefore, both the ML adoption framework and the maturity model for ML systems in clinics can not only guide future research that seeks to explore the promises and challenges associated with ML systems in a medical setting but also be a practical reference point for clinicians.

Background: The vaccination uptake rates of the human papillomavirus (HPV) vaccine remain low despite the fact that the effectiveness of HPV vaccines has been established for more than a decade. Vaccine hesitancy is in part due to false information about HPV vaccines on social media. Combating false HPV vaccine information is a reasonable step to addressing vaccine hesitancy. Objective: Given the substantial harm of false HPV vaccine information, there is an urgent need to identify false social media messages before it goes viral. The goal of the study is to develop a systematic and generalizable approach to identifying false HPV vaccine information on social media. Methods: This study used machine learning and natural language processing to develop a series of classification models and causality mining methods to identify and examine true and false HPV vaccine-related information on Twitter. Results: We found that the convolutional neural network model outperformed all other models in identifying tweets containing false HPV vaccine-related information (F score=91.95). We also developed completely unsupervised causality mining models to identify HPV vaccine candidate effects for capturing risk perceptions of HPV vaccines. Furthermore, we found that false information contained mostly loss-framed messages focusing on the potential risk of vaccines covering a variety of topics using more diverse vocabulary, while true information contained both gain- and loss-framed messages focusing on the effectiveness of vaccines covering fewer topics using relatively limited vocabulary. Conclusions: Our research demonstrated the feasibility and effectiveness of using predictive models to identify false HPV vaccine information and its risk perceptions on social media. Technology Services and the Cancer Center at Illinois at the University of Illinois at Urbana-Champaign; National Institutes of HealthUnited States Department of Health & Human ServicesNational Institutes of Health (NIH) - USA [1R21CA237483] Published version We thank Chieh-Li (Julian) Chin and Wen-Yuh Su for their assistance in the data collection and preparation for this study. We also thank the Technology Services and the Cancer Center at Illinois at the University of Illinois at Urbana-Champaign for their support of our study. This study was supported by a grant from the National Institutes of Health (#1R21CA237483). Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Institutes of Health.

In coeliac disease (CeD), immune‐mediated small intestinal damage is precipitated by gluten, leading to variable symptoms and complications, occasionally including aggressive T‐cell lymphoma. Diagnosis, based primarily on histopathological examination of duodenal biopsies, is confounded by poor concordance between pathologists and minimal histological abnormality if insufficient gluten is consumed. CeD pathogenesis involves both CD4+ T‐cell‐mediated gluten recognition and CD8+ and γδ T‐cell‐mediated inflammation, with a previous study demonstrating a permanent change in γδ T‐cell populations in CeD. We leveraged this understanding and explored the diagnostic utility of bulk T‐cell receptor (TCR) sequencing in assessing duodenal biopsies in CeD. Genomic DNA extracted from duodenal biopsies underwent sequencing for TCR‐δ (TRD) (CeD, n = 11; non‐CeD, n = 11) and TCR‐γ (TRG) (CeD, n = 33; non‐CeD, n = 21). We developed a novel machine learning‐based analysis of the TCR repertoire, clustering samples by diagnosis. Leave‐one‐out cross‐validation (LOOCV) was performed to validate the classification algorithm. Using TRD repertoire, 100% (22/22) of duodenal biopsies were correctly classified, with a LOOCV accuracy of 91%. Using TCR‐γ (TRG) repertoire, 94.4% (51/54) of duodenal biopsies were correctly classified, with LOOCV of 87%. Duodenal biopsy TRG repertoire analysis permitted accurate classification of biopsies from patients with CeD following a strict gluten‐free diet for at least 6 months, who would be misclassified by current tests. This result reflects permanent changes to the duodenal γδ TCR repertoire in CeD, even in the absence of gluten consumption. Our method could complement or replace histopathological diagnosis in CeD and might have particular clinical utility in the diagnostic testing of patients unable to tolerate dietary gluten, and for assessing duodenal biopsies with equivocal features. This approach is generalisable to any TCR/BCR locus and any sequencing platform, with potential to predict diagnosis or prognosis in conditions mediated or modulated by the adaptive immune response. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons, Ltd. on behalf of The Pathological Society of Great Britain and Ireland.

Background Technology-based computational strategies that leverage social network site (SNS) data to detect substance use are promising screening tools but rely on the presence of sufficient data to detect risk if it is present. A better understanding of the association between substance use and SNS participation may inform the utility of these technology-based screening tools. Objective This paper aims to examine associations between substance use and Instagram posts and to test whether such associations differ as a function of age, gender, and race/ethnicity. Methods Participants with an Instagram account were recruited primarily via Clickworker (N=3117). With participant permission and Instagram’s approval, participants’ Instagram photo posts were downloaded with an application program interface. Participants’ past-year substance use was measured with an adapted version of the National Institute on Drug Abuse Quick Screen. At-risk drinking was defined as at least one past-year instance having “had more than a few alcoholic drinks a day,” drug use was defined as any use of nonprescription drugs, and prescription drug use was defined as any nonmedical use of prescription medications. We used logistic regression to examine the associations between substance use and any Instagram posts and negative binomial regression to examine the associations between substance use and number of Instagram posts. We examined whether age (18-25, 26-38, 39+ years), gender, and race/ethnicity moderated associations in both logistic and negative binomial models. All differences noted were significant at the .05 level. Results Compared with no at-risk drinking, any at-risk drinking was associated with both a higher likelihood of any Instagram posts and a higher number of posts, except among Hispanic/Latino individuals, in whom at-risk drinking was associated with a similar number of posts. Compared with no drug use, any drug use was associated with a higher likelihood of any posts but was associated with a similar number of posts. Compared with no prescription drug use, any prescription drug use was associated with a similar likelihood of any posts and was associated with a lower number of posts only among those aged 39 years and older. Of note, main effects showed that being female compared with being male and being Hispanic/Latino compared with being White were significantly associated with both a greater likelihood of any posts and a greater number of posts. Conclusions Researchers developing computational substance use risk detection models using Instagram or other SNS data may wish to consider our findings showing that at-risk drinking and drug use were positively associated with Instagram participation, while prescription drug use was negatively associated with Instagram participation for middle- and older-aged adults. As more is learned about SNS behaviors among those who use substances, researchers may be better positioned to successfully design and interpret innovative risk detection approaches.

Background Respiratory virus infections are significant causes of morbidity and mortality, and may induce host metabolite alterations by infecting respiratory epithelial cells. We investigated the use of liquid chromatography quadrupole time-of-flight mass spectrometry (LC/Q-TOF) combined with machine learning for the diagnosis of influenza infection. Methods We analyzed nasopharyngeal swab samples by LC/Q-TOF to identify distinct metabolic signatures for diagnosis of acute illness. Machine learning models were performed for classification, followed by Shapley additive explanation (SHAP) analysis to analyze feature importance and for biomarker discovery. Findings A total of 236 samples were tested in the discovery phase by LC/Q-TOF, including 118 positive samples (40 influenza A 2009 H1N1, 39 influenza H3 and 39 influenza B) as well as 118 age and sex-matched negative controls with acute respiratory illness. Analysis showed an area under the receiver operating characteristic curve (AUC) of 1.00 (95% confidence interval [95% CI] 0.99, 1.00), sensitivity of 1.00 (95% CI 0.86, 1.00) and specificity of 0.96 (95% CI 0.81, 0.99). The metabolite most strongly associated with differential classification was pyroglutamic acid. Independent validation of a biomarker signature based on the top 20 differentiating ion features was performed in a prospective cohort of 96 symptomatic individuals including 48 positive samples (24 influenza A 2009 H1N1, 5 influenza H3 and 19 influenza B) and 48 negative samples. Testing performed using a clinically-applicable targeted approach, liquid chromatography triple quadrupole mass spectrometry, showed an AUC of 1.00 (95% CI 0.998, 1.00), sensitivity of 0.94 (95% CI 0.83, 0.98), and specificity of 1.00 (95% CI 0.93, 1.00). Limitations include lack of sample suitability assessment, and need to validate these findings in additional patient populations. Interpretation This metabolomic approach has potential for diagnostic applications in infectious diseases testing, including other respiratory viruses, and may eventually be adapted for point-of-care testing. Funding None.