Medical Visual Question Answering (VQA) is a combination of medical artificial intelligence and popular VQA challenges. Given a medical image and a clinically relevant question in natural language, the medical VQA system is expected to predict a plausible and convincing answer. Although the general-domain VQA has been extensively studied, the medical VQA still needs specific investigation and exploration due to its task features. In the first part of this survey, we collect and discuss the publicly available medical VQA datasets up-to-date about the data source, data quantity, and task feature. In the second part, we review the approaches used in medical VQA tasks. We summarize and discuss their techniques, innovations, and potential improvements. In the last part, we analyze some medical-specific challenges for the field and discuss future research directions. Our goal is to provide comprehensive and helpful information for researchers interested in the medical visual question answering field and encourage them to conduct further research in this field.

The world has recently been exposed to a fierce attack from many viral diseases, such as Covid-19, that exhausted medical systems around the world. Such attack had a negative impact not only on the health status of people or the high death rate, but also had a bad impact on the economic situation, which affected all countries of the world especially the poor and the developing ones. Monkeypox is one of the latest viral diseases that may cause a pandemic in the near future if not dealt and diagnosed with appropriately. This paper provides a new strategy for diagnosing monkeypox, which is called; Accurate Monkeypox Diagnosing Strategy (AMDS). The proposed AMDS consists of two phases, which are; (i) pre-processing and (ii) classification. During the pre-processing phase, the most effective feature are selected using Binary Tiki-Taka Algorithm (BTTA). On the other hand, in the classification phase, ensemble classification is used for diagnosing new cases, which combines evidence from three different new classifiers, namely; (a) Layered K-Nearest Neighbors (LKNN), (b) Statistical Naïve Bayes (SNB), and (c) Deep Learning Classifier (DLC). Moreover, the decisions of the proposed classifiers are merged in a new voting scheme called Fuzzified Voting Scheme (FVS). AMDS has been compared against recent diagnostic strategies. Experimental results have proven that AMDS outperforms other monkeypox diagnostic strategies as it introduces the most accurate diagnosis according to two different datasets.

Nowadays, it is really important and crucial to follow the new biomedical knowledge that is presented in scientific literature. To this end, Information Extraction pipelines can help to automatically extract meaningful relations from textual data that further require additional checks by domain experts. In the last two decades, a lot of work has been performed for extracting relations between phenotype and health concepts, however, the relations with food entities which are one of the most important environmental concepts have never been explored. In this study, we propose FooDis, a novel Information Extraction pipeline that employs state-of-the-art approaches in Natural Language Processing to mine abstracts of biomedical scientific papers and automatically suggests potential cause or treat relations between food and disease entities in different existing semantic resources. A comparison with already known relations indicates that the relations predicted by our pipeline match for 90% of the food-disease pairs that are common in our results and the NutriChem database, and 93% of the common pairs in the DietRx platform. The comparison also shows that the FooDis pipeline can suggest relations with high precision. The FooDis pipeline can be further used to dynamically discover new relations between food and diseases that should be checked by domain experts and further used to populate some of the existing resources used by NutriChem and DietRx.

Training deep neural network classifiers for electrocardiograms (ECGs) requires sufficient data. However, imbalanced datasets pose a major problem for the training process and hence data augmentation is commonly performed. Generative adversarial networks (GANs) can create synthetic ECG data to augment such imbalanced datasets. This review aims at identifying the present literature concerning synthetic ECG signal generation using GANs to provide a comprehensive overview of architectures, quality evaluation metrics, and classification performances. Thirty publications from the years 2019 to 2022 were selected from three separate databases. Nine publications used a quality evaluation metric neglecting classification, eleven performed a classification but omitted a quality evaluation metric, and ten publications performed both. Twenty different quality evaluation metrics were observed. Overall, the classification performance of databases augmented with synthetically created ECG signals increased by 7 % to 98 % in accuracy and 6 % to 97 % in sensitivity. In conclusion, synthetic ECG signal generation using GANs represents a promising tool for data augmentation of imbalanced datasets. Consistent quality evaluation of generated signals remains challenging. Hence, future work should focus on the establishment of a gold standard for quality evaluation metrics for GANs.

Brain segmentation is often the first and most critical step in quantitative analysis of the brain for many clinical applications, including fetal imaging. Different aspects challenge the segmentation of the fetal brain in magnetic resonance imaging (MRI), such as the non-standard position of the fetus owing to his/her movements during the examination, rapid brain development, and the limited availability of imaging data. In recent years, several segmentation methods have been proposed for automatically partitioning the fetal brain from MR images. These algorithms aim to define regions of interest with different shapes and intensities, encompassing the entire brain, or isolating specific structures. Deep learning techniques, particularly convolutional neural networks (CNNs), have become a state-of-the-art approach in the field because they can provide reliable segmentation results over heterogeneous datasets. Here, we review the deep learning algorithms developed in the field of fetal brain segmentation and categorize them according to their target structures. Finally, we discuss the perceived research gaps in the literature of the fetal domain, suggesting possible future research directions that could impact the management of fetal MR images.

Electronic systems are increasingly present in the healthcare system and are often related to improved medical care. However, the widespread use of these technologies ended up building a relationship of dependence that can disrupt the doctor-patient relationship. In this context, digital scribes are automated clinical documentation systems that capture the physician-patient conversation and then generate the documentation for the appointment, enabling the physician to engage with the patient entirely. We have performed a systematic literature review on intelligent solutions for automatic speech recognition (ASR) with automatic documentation during a medical interview. The scope included only original research on systems that could detect speech and transcribe it in a natural and structured fashion simultaneously with the doctor-patient interaction, excluding speech-to-text-only technologies. The search resulted in a total of 1995 titles, with eight articles remaining after filtering for the inclusion and exclusion criteria. The intelligent models mainly consisted of an ASR system with natural language processing capability, a medical lexicon, and structured text output. None of the articles had a commercially available product at the time of the publication and reported limited real-life experience. So far, none of the applications has been prospectively validated and tested in large-scale clinical studies. Nonetheless, these first reports suggest that automatic speech recognition may be a valuable tool in the future to facilitate medical registration in a faster and more reliable manner. Improving transparency, accuracy, and empathy could drastically change how patients and doctors experience a medical visit. Unfortunately, clinical data on the usability and benefits of such applications is almost non-existent. We believe that future work in this area is necessary and needed.

BACKGROUND: Artificial intelligence (AI) technology has clustered patients based on clinical features into sub-clusters to stratify high-risk and low-risk groups to predict outcomes in lung cancer after radiotherapy and has gained much more attention in recent years. Given that the conclusions vary considerably, this meta-analysis was conducted to investigate the combined predictive effect of AI models on lung cancer. METHODS: This study was performed according to PRISMA guidelines. PubMed, ISI Web of Science, and Embase databases were searched for relevant literature. Outcomes, including overall survival (OS), disease-free survival (DFS), progression-free survival (PFS), and local control (LC), were predicted using AI models in patients with lung cancer after radiotherapy, and were used to calculate the pooled effect. Quality, heterogeneity, and publication bias of the included studies were also evaluated. RESULTS: Eighteen articles with 4719 patients were eligible for this meta-analysis. The combined hazard ratios (HRs) of the included studies for OS, LC, PFS, and DFS of lung cancer patients were 2.55 (95 % confidence interval (CI) = 1.73-3.76), 2.45 (95 % CI = 0.78-7.64), 3.84 (95 % CI = 2.20-6.68), and 2.66 (95 % CI = 0.96-7.34), respectively. The combined area under the receiver operating characteristics curve (AUC) of the included articles on OS and LC in patients with lung cancer was 0.75 (95 % CI = 0.67-0.84), and 0.80 (95%CI = 0.0.68-0.95), respectively. CONCLUSION: The clinical feasibility of predicting outcomes using AI models after radiotherapy in patients with lung cancer was demonstrated. Large-scale, prospective, multicenter studies should be conducted to more accurately predict the outcomes in patients with lung cancer.

Electronic health records (EHR) are sparse, noisy, and private, with variable vital measurements and stay lengths. Deep learning models are the current state of the art in many machine learning domain; however, the EHR data is not a suitable training input for most of them. In this paper, we introduce RIMD, a novel deep learning model that consists of a decay mechanism, modular recurrent networks, and a custom loss function that learns minor classes. The decay mechanism learns from patterns in sparse data. The modular network allows multiple recurrent networks to pick only relevant input based on the attention score at a given timestamp. Finally, the custom class balance loss function is responsible for learning minor classes based on samples provided in training. This novel model is used to evaluate predictions for early mortality identification, length of stay, and acute respiratory failure on MIMIC-III dataset. Experiment results indicate that the proposed models outperform similar models in F1-Score, AUROC, and PRAUC scores.

BACKGROUND: Cervical cancer is one of the most common cancers in women with an incidence of around 6.5 % of all the cancer in women worldwide. Early detection and adequate treatment according to staging improve the patient's life expectancy. Outcome prediction models might aid treatment decisions, but a systematic review on prediction models for cervical cancer patients is not available. DESIGN: We performed a systematic review for prediction models in cervical cancer following PRISMA guidelines. Key features that were used for model training and validation, the endpoints were extracted from the article and data were analyzed. Selected articles were grouped based on prediction endpoints i.e. Group1: Overall survival, Group2: progression-free survival; Group3: recurrence or distant metastasis; Group4: treatment response; Group5: toxicity or quality of life. We developed a scoring system to evaluate the manuscript. As per our criteria, studies were divided into four groups based on scores obtained in our scoring system, the Most significant study (Score > 60 %); Significant study (60 % > Score > 50 %); Moderately Significant study (50 % > Score > 40 %); least significant study (score < 40 %). A meta-analysis was performed for all the groups separately. RESULTS: The first line of search selected 1358 articles and finally 39 articles were selected as eligible for inclusion in the review. As per our assessment criteria, 16, 13 and 10 studies were found to be the most significant, significant and moderately significant respectively. The intra-group pooled correlation coefficient for Group1, Group2, Group3, Group4, and Group5 were 0.76 [0.72, 0.79], 0.80 [0.73, 0.86], 0.87 [0.83, 0.90], 0.85 [0.77, 0.90], 0.88 [0.85, 0.90] respectively. All the models were found to be good (prediction accuracy [c-index/AUC/R2] >0.7) in endpoint prediction. CONCLUSIONS: Prediction models of cervical cancer toxicity, local or distant recurrence and survival prediction show promising results with reasonable prediction accuracy [c-index/AUC/R2 > 0.7]. These models should also be validated on external data and evaluated in prospective clinical studies.

Clinical Practice Guidelines (CPGs) include recommendations aimed at optimising patient care, informed by a review of the available clinical evidence. To achieve their potential benefits, CPG should be readily available at the point of care. This can be done by translating CPG recommendations into one of the languages for Computer-Interpretable Guidelines (CIGs). This is a difficult task for which the collaboration of clinical and technical staff is crucial. However, in general CIG languages are not accessible to non-technical staff. We propose to support the modelling of CPG processes (and hence the authoring of CIGs) based on a transformation, from a preliminary specification in a more accessible language into an implementation in a CIG language. In this paper, we approach this transformation following the Model-Driven Development (MDD) paradigm, in which models and transformations are key elements for software development. To demonstrate the approach, we implemented and tested an algorithm for the transformation from the BPMN language for business processes to the PROforma CIG language. This implementation uses transformations defined in the ATLAS Transformation Language. Additionally, we conducted a small experiment to assess the hypothesis that a language such as BPMN can facilitate the modelling of CPG processes by clinical and technical staff.

Medical dialog systems have the potential to assist e-medicine in improving access to healthcare services, improving patient treatment quality, and lowering medical expenses. In this research, we describe a knowledge-grounded conversation generation model that demonstrates how large-scale medical information in the form of knowledge graphs can aid in language comprehension and generation in medical dialog systems. Generic responses are often produced by existing generative dialog systems, resulting in monotonous and uninteresting conversations. To solve this problem, we combine various pre-trained language models with a medical knowledge base (UMLS) to generate clinically correct and human-like medical conversations using the recently released MedDialog-EN dataset. The medical-specific knowledge graph contains broadly 3 types of medical-related information, including disease, symptom and laboratory test. We perform reasoning over the retrieved knowledge graph by reading the triples in each graph using MedFact attention, which allows us to use semantic information from the graphs for better response generation. In order to preserve medical information, we employ a policy network, which effectively injects relevant entities associated with each dialog into the response. We also study how transfer learning can significantly improve the performance by utilizing a relatively small corpus, created by extending the recently released CovidDialog dataset, containing the dialogs for diseases that are symptoms of Covid-19. Empirical results on the MedDialog corpus and the extended CovidDialog dataset demonstrate that our proposed model significantly outperforms the state-of-the-art methods in terms of both automatic evaluation and human judgment.

While deep learning has displayed excellent performance in a broad spectrum of application areas, neural networks still struggle to recognize what they have not seen, i.e., out-of-distribution (OOD) inputs. In the medical field, building robust models that are able to detect OOD images is highly critical, as these rare images could show diseases or anomalies that should be detected. In this study, we use wireless capsule endoscopy (WCE) images to present a novel patch-based self-supervised approach comprising three stages. First, we train a triplet network to learn vector representations of WCE image patches. Second, we cluster the patch embeddings to group patches in terms of visual similarity. Third, we use the cluster assignments as pseudolabels to train a patch classifier and use the Out-of-Distribution Detector for Neural Networks (ODIN) for OOD detection. The system has been tested on the Kvasir-capsule, a publicly released WCE dataset. Empirical results show an OOD detection improvement compared to baseline methods. Our method can detect unseen pathologies and anomalies such as lymphangiectasia, foreign bodies and blood with AUROC>0.6. This work presents an effective solution for OOD detection models without needing labeled images.

Deep learning has become a thriving force in the computer aided diagnosis of liver cancer, as it solves extremely complicated challenges with high accuracy over time and facilitates medical experts in their diagnostic and treatment procedures. This paper presents a comprehensive systematic review on deep learning techniques applied for various applications pertaining to liver images, challenges faced by the clinicians in liver tumour diagnosis and how deep learning bridges the gap between clinical practice and technological solutions with an in-depth summary of 113 articles. Since, deep learning is an emerging revolutionary technology, recent state-of-the-art research implemented on liver images are reviewed with more focus on classification, segmentation and clinical applications in the management of liver diseases. Additionally, similar review articles in literature are reviewed and compared. The review is concluded by presenting the contemporary trends and unaddressed research issues in the field of liver tumour diagnosis, offering directions for future research in this field.

OBJECTIVE: The proper handling of missing values is critical to delivering reliable estimates and decisions, especially in high-stakes fields such as clinical research. In response to the increasing diversity and complexity of data, many researchers have developed deep learning (DL)-based imputation techniques. We conducted a systematic review to evaluate the use of these techniques, with a particular focus on the types of data, intending to assist healthcare researchers from various disciplines in dealing with missing data. MATERIALS AND METHODS: We searched five databases (MEDLINE, Web of Science, Embase, CINAHL, and Scopus) for articles published prior to February 8, 2023 that described the use of DL-based models for imputation. We examined selected articles from four perspectives: data types, model backbones (i.e., main architectures), imputation strategies, and comparisons with non-DL-based methods. Based on data types, we created an evidence map to illustrate the adoption of DL models. RESULTS: Out of 1822 articles, a total of 111 were included, of which tabular static data (29%, 32/111) and temporal data (40%, 44/111) were the most frequently investigated. Our findings revealed a discernible pattern in the choice of model backbones and data types, for example, the dominance of autoencoder and recurrent neural networks for tabular temporal data. The discrepancy in imputation strategy usage among data types was also observed. The "integrated" imputation strategy, which solves the imputation task simultaneously with downstream tasks, was most popular for tabular temporal data (52%, 23/44) and multi-modal data (56%, 5/9). Moreover, DL-based imputation methods yielded a higher level of imputation accuracy than non-DL methods in most studies. CONCLUSION: The DL-based imputation models are a family of techniques, with diverse network structures. Their designation in healthcare is usually tailored to data types with different characteristics. Although DL-based imputation models may not be superior to conventional approaches across all datasets, it is highly possible for them to achieve satisfactory results for a particular data type or dataset. There are, however, still issues with regard to portability, interpretability, and fairness associated with current DL-based imputation models.

Every day, 800 women and 6700 newborns die from complications related to pregnancy or childbirth. A well-trained midwife can prevent most of these maternal and newborn deaths. Data science models together with logs generated by users of online learning applications for midwives can help improve their learning competencies. In this work, we evaluate various forecasting methods to determine the future interest of users for the different types of content available in the Safe Delivery App, a digital training tool for skilled birth attendants, broken down by profession and region. This first attempt at health content demand forecasting for midwifery learning shows that DeepAR can accurately anticipate content demand in operational settings, and could therefore be used to offer users personalized content and to provide an adaptive learning journey.

BACKGROUND AND AIM: Cognitive Computing systems are the intelligent systems that thinks, understands and augments the capabilities of human brain by blending the technologies of Artificial Intelligence, Machine Learning and Natural Language Processing. In recent days, maintenance or enhancement of health by preclusion, prognosis, and analysis of diseases has become a challenging task. The increasing diseases and its causes becomes a big question before humanity. Limited risk analysis, meticulous training process, and automated critical decision-making are some of the issues of cognitive computing. To overcome this issue, cognitive computing in healthcare works like a medical prodigy which anticipates the disease or illness of the human being and helps the doctors with technological facts to take the timely action. The main aim of this survey article is to explore the present and futuristic technological trends of cognitive computing in healthcare. In this work, different cognitive computing applications are reviewed, and the best application is recommended to the clinicians. Based on this recommendation, the clinicians are able to monitor and analyze the physical health of patients. METHODS: This article presents the systematic literature on the different aspects of cognitive computing in healthcare. Nearly seven online databases such as SCOPUS, IEEE Xplore, Google Scholar, DBLP, Web of Science, Springer and PubMed were screened and the published articles related to cognitive computing in healthcare is collected from 2014 to 2021. In total, 75 articles were selected, examined and their pros and cons are analyzed. The analysis is done with respect to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. RESULTS: The basic findings of this review article and their significance for theory and practice are mindmaps portraying the cognitive computing platforms, cognitive applications in healthcare, and use cases of cognitive computing in healthcare. A detailed discussion section highlighting the present issues, future research directions and recent applications of cognitive computing in healthcare. Accuracy analysis of different cognitive systems conclude that the Medical Sieve achieves 0.95 and Watson For Oncology (WFO) achieves 0.93 and hence proves to be the prominent computing systems for healthcare. CONCLUSIONS: Cognitive computing, an evolving technology in healthcare augments the clinical thought process and enable the doctors to make the right diagnosis and preserve the patient's health in good condition. These systems provides timely care, optimal and cost-effective treatment. This article provides an extensive survey of the importance of cognitive computing in the health sector by highlighting the platforms, techniques, tools, algorithms, applications, and use cases. This survey also explores about the works in the literature on present issues and proposes the future research directions of applying cognitive systems in healthcare.

Diabetic Retinopathy (DR) is the most popular debilitating impairment of diabetes and it progresses symptom-free until a sudden loss of vision occurs. Understanding the progression of DR is a pressing issue in clinical research and practice. In this systematic review of articles on Machine Learning (ML) based risk prediction models for DR progression, ever since the use of Artificial Intelligence (AI) for DR detection, there have been more cross-sectional studies with different algorithms of use of AI, there haven't been many longitudinal studies for the AI based risk prediction models. This paper proposes a novel review to fill in the gaps identified in current reviews and facilitate other researchers with current research solutions for developing AI-based risk prediction models for DR progression and closely related problems; synthesize the current results from these studies and identify research challenges, limitations and gaps to inform the selection of machine learning techniques and predictors to build novel prediction models. Additionally, this paper suggested six (6) deep AI-related technical and critical discussion of the adopted strategies and approaches. The Systematic Literature Review (SLR) methodology was employed to gather relevant studies. We searched IEEE Xplore, PubMed, Springer Link, Google Scholar, and Science Direct electronic databases for papers published from January 2017 to 30th April 2023. Thirteen (13) studies were chosen on the basis of their relevance to the review questions and satisfying the selection criteria. However, findings from the literature review exposed some critical research gaps that need to be addressed in future research to improve on the performance of risk prediction models for DR progression.

BACKGROUND: Hospital-acquired pressure injuries (HAPIs) constitute a significant challenge harming thousands of people worldwide yearly. While various tools and methods are used to identify pressure injuries, artificial intelligence (AI) and decision support systems (DSS) can help to reduce HAPIs risks by proactively identifying patients at risk and preventing them before harming patients. OBJECTIVE: This paper comprehensively reviews AI and DSS applications for HAPIs prediction using Electronic Health Records (EHR), including a systematic literature review and bibliometric analysis. METHODS: A systematic literature review was conducted through PRISMA and bibliometric analysis. In February 2023, the search was performed using four electronic databases: SCOPIS, PubMed, EBSCO, and PMCID. Articles on using AI and DSS in the management of PIs were included. RESULTS: The search approach yielded 319 articles, 39 of which have been included and classified into 27 AI-related and 12 DSS-related categories. The years of publication varied from 2006 to 2023, with 40% of the studies taking place in the US. Most studies focused on using AI algorithms or DSS for HAPIs prediction in inpatient units using various types of data such as electronic health records, PI assessment scales, and expert knowledge-based and environmental data to identify the risk factors associated with HAPIs development. CONCLUSIONS: There is insufficient evidence in the existing literature concerning the real impact of AI or DSS on making decisions for HAPIs treatment or prevention. Most studies reviewed are solely hypothetical and retrospective prediction models, with no actual application in healthcare settings. The accuracy rates, prediction results, and intervention procedures suggested based on the prediction, on the other hand, should inspire researchers to combine both approaches with larger-scale data to bring a new venue for HAPIs prevention and to investigate and adopt the suggested solutions to the existing gaps in AI and DSS prediction methods.

The SARS-CoV-2 pandemic highlighted the need for software tools that could facilitate patient triage regarding potential disease severity or even death. In this article, an ensemble of Machine Learning (ML) algorithms is evaluated in terms of predicting the severity of their condition using plasma proteomics and clinical data as input. An overview of AI-based technical developments to support COVID-19 patient management is presented outlining the landscape of relevant technical developments. Based on this review, the use of an ensemble of ML algorithms that analyze clinical and biological data (i.e., plasma proteomics) of COVID-19 patients is designed and deployed to evaluate the potential use of AI for early COVID-19 patient triage. The proposed pipeline is evaluated using three publicly available datasets for training and testing. Three ML "tasks" are defined, and several algorithms are tested through a hyperparameter tuning method to identify the highest-performance models. As overfitting is one of the typical pitfalls for such approaches (mainly due to the size of the training/validation datasets), a variety of evaluation metrics are used to mitigate this risk. In the evaluation procedure, recall scores ranged from 0.6 to 0.74 and F1-score from 0.62 to 0.75. The best performance is observed via Multi-Layer Perceptron (MLP) and Support Vector Machines (SVM) algorithms. Additionally, input data (proteomics and clinical data) were ranked based on corresponding Shapley additive explanation (SHAP) values and evaluated for their prognosticated capacity and immuno-biological credence. This "interpretable" approach revealed that our ML models could discern critical COVID-19 cases predominantly based on patient's age and plasma proteins on B cell dysfunction, hyper-activation of inflammatory pathways like Toll-like receptors, and hypo-activation of developmental and immune pathways like SCF/c-Kit signaling. Finally, the herein computational workflow is corroborated in an independent dataset and MLP superiority along with the implication of the abovementioned predictive biological pathways are corroborated. Regarding limitations of the presented ML pipeline, the datasets used in this study contain less than 1000 observations and a significant number of input features hence constituting a high-dimensional low-sample (HDLS) dataset which could be sensitive to overfitting. An advantage of the proposed pipeline is that it combines biological data (plasma proteomics) with clinical-phenotypic data. Thus, in principle, the presented approach could enable patient triage in a timely fashion if used on already trained models. However, larger datasets and further systematic validation are needed to confirm the potential clinical value of this approach. The code is available on Github: https://github.com/inab-certh/Predicting-COVID-19-severity-through-interpretable-AI-analysis-of-plasma-proteomics.

This paper presents ArrhyMon, a self-attention-based LSTM-FCN model for arrhythmia classification from ECG signal inputs. ArrhyMon targets to detect and classify six different types of arrhythmia apart from normal ECG patterns. To the best of our knowledge, ArrhyMon is the first end-to-end classification model that successfully targets the classification of six detailed arrhythmia types and compared to previous work does not require additional preprocessing and/or feature extraction operations separate from the classification model. ArrhyMon's deep learning model is designed to capture and exploit both global and local features embedded in ECG sequences by integrating fully convolutional network (FCN) layers and a self-attention-based long and short-term memory (LSTM) architecture. Moreover, to enhance its practicality, ArrhyMon incorporates a deep ensemble-based uncertainty model that generates a confidence-level measure for each classification result. We evaluate ArrhyMon's effectiveness using three publicly available arrhythmia datasets (i.e., MIT-BIH, Physionet Cardiology Challenge 2017 and 2020/2021) to show that ArrhyMon achieves state-of-the-art classification performance (average accuracy 99.63%), and that confidence measures show close correlation with subjective diagnosis made from practitioners.

OBJECTIVE: Many Computer Aided Prognostic (CAP) systems based on machine learning techniques have been proposed in the field of oncology. The objective of this systematic review was to assess and critically appraise the methodologies and approaches used in predicting the prognosis of gynecological cancers using CAPs. METHODS: Electronic databases were used to systematically search for studies utilizing machine learning methods in gynecological cancers. Study risk of bias (ROB) and applicability were assessed using the PROBAST tool. 139 studies met the inclusion criteria, of which 71 predicted outcomes for ovarian cancer patients, 41 predicted outcomes for cervical cancer patients, 28 predicted outcomes for uterine cancer patients, and 2 predicted outcomes for gynecological malignancies broadly. RESULTS: Random forest (22.30 %) and support vector machine (21.58 %) classifiers were used most commonly. Use of clinicopathological, genomic and radiomic data as predictors was observed in 48.20 %, 51.08 % and 17.27 % of studies, respectively, with some studies using multiple modalities. 21.58 % of studies were externally validated. Twenty-three individual studies compared ML and non-ML methods. Study quality was highly variable and methodologies, statistical reporting and outcome measures were inconsistent, preventing generalized commentary or meta-analysis of performance outcomes. CONCLUSION: There is significant variability in model development when prognosticating gynecological malignancies with respect to variable selection, machine learning (ML) methods and endpoint selection. This heterogeneity prevents meta-analysis and conclusions regarding the superiority of ML methods. Furthermore, PROBAST-mediated ROB and applicability analysis demonstrates concern for the translatability of existing models. This review identifies ways that this can be improved upon in future works to develop robust, clinically translatable models within this promising field.

In this paper, we study human-AI collaboration protocols, a design-oriented construct aimed at establishing and evaluating how humans and AI can collaborate in cognitive tasks. We applied this construct in two user studies involving 12 specialist radiologists (the knee MRI study) and 44 ECG readers of varying expertise (the ECG study), who evaluated 240 and 20 cases, respectively, in different collaboration configurations. We confirm the utility of AI support but find that XAI can be associated with a "white-box paradox", producing a null or detrimental effect. We also find that the order of presentation matters: AI-first protocols are associated with higher diagnostic accuracy than human-first protocols, and with higher accuracy than both humans and AI alone. Our findings identify the best conditions for AI to augment human diagnostic skills, rather than trigger dysfunctional responses and cognitive biases that can undermine decision effectiveness.

Medical risk detection is an important topic and a challenging task to improve the performance of clinical practices in Intensive Care Units (ICU). Although many bio-statistical learning and deep learning approaches have provided patient-specific mortality predictions, these existing methods lack interpretability that is crucial to gain adequate insight on why such predictions would work. In this paper, we introduce cascading theory to model the physiological domino effect and provide a novel approach to dynamically simulate the deterioration of patients' conditions. We propose a general DEep CAscading Framework (DECAF) to predict the potential risks of all physiological functions at each clinical stage. Compared with other feature-based and/or score-based models, our approach has a range of desirable properties, such as being interpretable, applicable with multi prediction tasks, and learnable from medical common sense and/or clinical experience knowledge. Experiments on a medical dataset (MIMIC-III) of 21,828 ICU patients show that DECAF reaches up to 89.30 % on AUROC, which surpasses the best competing methods for mortality prediction.

BACKGROUND: Indigenous peoples often have higher rates of morbidity and mortality associated with cardiometabolic disease (CMD) than non-Indigenous people and this may be even more so in urban areas. The use of electronic health records and expansion of computing power has led to mainstream use of artificial intelligence (AI) to predict the onset of disease in primary health care (PHC) settings. However, it is unknown if AI and in particular machine learning is used for risk prediction of CMD in Indigenous peoples. METHODS: We searched peer-reviewed literature using terms associated with AI machine learning, PHC, CMD, and Indigenous peoples. RESULTS: We identified 13 suitable studies for inclusion in this review. Median total number of participants was 19,270 (range 911-2,994,837). The most common algorithms used in machine learning in this setting were support vector machine, random forest, and decision tree learning. Twelve studies used the area under the receiver operating characteristic curve (AUC) to measure performance. Two studies reported an AUC of >0.9. Six studies had an AUC score between 0.9 and 0.8, 4 studies had an AUC score between 0.8 and 0.7. 1 study reported an AUC score between 0.7 and 0.6. Risk of bias was observed in 10 (77 %) studies. CONCLUSION: AI machine learning and risk prediction models show moderate to excellent discriminatory ability over traditional statistical models in predicting CMD. This technology could help address the needs of urban Indigenous peoples by predicting CMD early and more rapidly than conventional methods.

BACKGROUND: Amyotrophic Lateral Sclerosis (ALS) is a fatal neurodegenerative disorder characterised by the progressive loss of motor neurons in the brain and spinal cord. The fact that ALS's disease course is highly heterogeneous, and its determinants not fully known, combined with ALS's relatively low prevalence, renders the successful application of artificial intelligence (AI) techniques particularly arduous. OBJECTIVE: This systematic review aims at identifying areas of agreement and unanswered questions regarding two notable applications of AI in ALS, namely the automatic, data-driven stratification of patients according to their phenotype, and the prediction of ALS progression. Differently from previous works, this review is focused on the methodological landscape of AI in ALS. METHODS: We conducted a systematic search of the Scopus and PubMed databases, looking for studies on data-driven stratification methods based on unsupervised techniques resulting in (A) automatic group discovery or (B) a transformation of the feature space allowing patient subgroups to be identified; and for studies on internally or externally validated methods for the prediction of ALS progression. We described the selected studies according to the following characteristics, when applicable: variables used, methodology, splitting criteria and number of groups, prediction outcomes, validation schemes, and metrics. RESULTS: Of the starting 1604 unique reports (2837 combined hits between Scopus and PubMed), 239 were selected for thorough screening, leading to the inclusion of 15 studies on patient stratification, 28 on prediction of ALS progression, and 6 on both stratification and prediction. In terms of variables used, most stratification and prediction studies included demographics and features derived from the ALSFRS or ALSFRS-R scores, which were also the main prediction targets. The most represented stratification methods were K-means, and hierarchical and expectation-maximisation clustering; while random forests, logistic regression, the Cox proportional hazard model, and various flavours of deep learning were the most widely used prediction methods. Predictive model validation was, albeit unexpectedly, quite rarely performed in absolute terms (leading to the exclusion of 78 eligible studies), with the overwhelming majority of included studies resorting to internal validation only. CONCLUSION: This systematic review highlighted a general agreement in terms of input variable selection for both stratification and prediction of ALS progression, and in terms of prediction targets. A striking lack of validated models emerged, as well as a general difficulty in reproducing many published studies, mainly due to the absence of the corresponding parameter lists. While deep learning seems promising for prediction applications, its superiority with respect to traditional methods has not been established; there is, instead, ample room for its application in the subfield of patient stratification. Finally, an open question remains on the role of new environmental and behavioural variables collected via novel, real-time sensors.

OBJECTIVE: High-resolution ultrasound is an emerging tool for diagnosing carpal tunnel syndrome caused by the compression of the median nerve at the wrist. This systematic review and meta-analysis aimed to explore and summarize the performance of deep learning algorithms in the automatic sonographic assessment of the median nerve at the carpal tunnel level. METHODS: PubMed, Medline, Embase, and Web of Science were searched from the earliest records to May 2022 for studies investigating the utility of deep neural networks in the evaluation of the median nerve in carpal tunnel syndrome. The quality of the included studies was evaluated using the Quality Assessment Tool for Diagnostic Accuracy Studies. The outcome variables included precision, recall, accuracy, F-score, and Dice coefficient. RESULTS: In total, seven articles were included, comprising 373 participants. The deep learning and related algorithms comprised U-Net, phase-based probabilistic active contour, MaskTrack, ConvLSTM, DeepNerve, DeepSL, ResNet, Feature Pyramid Network, DeepLab, Mask R-CNN, region proposal network, and ROI Align. The pooled values of precision and recall were 0.917 (95 % confidence interval [CI], 0.873-0.961) and 0.940 (95 % CI, 0.892-0.988), respectively. The pooled accuracy and Dice coefficient were 0.924 (95 % CI, 0.840-1.008) and 0.898 (95 % CI, 0.872-0.923), respectively, whereas the summarized F-score was 0.904 (95 % CI, 0.871-0.937). CONCLUSION: The deep learning algorithm enables automated localization and segmentation of the median nerve at the carpal tunnel level in ultrasound imaging with acceptable accuracy and precision. Future research is expected to validate the performance of deep learning algorithms in detecting and segmenting the median nerve along its entire length as well as across datasets obtained from various ultrasound manufacturers.

BACKGROUND: DNA methylation biomarkers have great potential in improving prognostic classification systems for patients with cancer. Machine learning (ML)-based analytic techniques might help overcome the challenges of analyzing high-dimensional data in relatively small sample sizes. This systematic review summarizes the current use of ML-based methods in epigenome-wide studies for the identification of DNA methylation signatures associated with cancer prognosis. METHODS: We searched three electronic databases including PubMed, EMBASE, and Web of Science for articles published until 2 January 2023. ML-based methods and workflows used to identify DNA methylation signatures associated with cancer prognosis were extracted and summarized. Two authors independently assessed the methodological quality of included studies by a seven-item checklist adapted from 'A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies (PROBAST)' and from the 'Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK). Different ML methods and workflows used in included studies were summarized and visualized by a sunburst chart, a bubble chart, and Sankey diagrams, respectively. RESULTS: Eighty-three studies were included in this review. Three major types of ML-based workflows were identified. 1) unsupervised clustering, 2) supervised feature selection, and 3) deep learning-based feature transformation. For the three workflows, the most frequently used ML techniques were consensus clustering, least absolute shrinkage and selection operator (LASSO), and autoencoder, respectively. The systematic review revealed that the performance of these approaches has not been adequately evaluated yet and that methodological and reporting flaws were common in the identified studies using ML techniques. CONCLUSIONS: There is great heterogeneity in ML-based methodological strategies used by epigenome-wide studies to identify DNA methylation markers associated with cancer prognosis. In theory, most existing workflows could not handle the high multi-collinearity and potentially non-linearity interactions in epigenome-wide DNA methylation data. Benchmarking studies are needed to compare the relative performance of various approaches for specific cancer types. Adherence to relevant methodological and reporting guidelines are urgently needed.

Prevention and treatment of complications are the backbone of medical care, particularly in critical care settings. Early detection and prompt intervention can potentially prevent complications from occurring and improve outcomes. In this study, we use four longitudinal vital signs variables of intensive care unit patients, focusing on predicting acute hypertensive episodes (AHEs). These episodes represent elevations in blood pressure and may result in clinical damage or indicate a change in a patient's clinical situation, such as an elevation in intracranial pressure or kidney failure. Prediction of AHEs may allow clinicians to anticipate changes in the patient's condition and respond early on to prevent these from occurring. Temporal abstraction was employed to transform the multivariate temporal data into a uniform representation of symbolic time intervals, from which frequent time-intervals-related patterns (TIRPs) are mined and used as features for AHE prediction. A novel TIRP metric for classification, called coverage, is introduced that measures the coverage of a TIRP's instances in a time window. For comparison, several baseline models were applied on the raw time series data, including logistic regression and sequential deep learning models, are used. Our results show that using frequent TIRPs as features outperforms the baseline models, and the use of the coverage, metric outperforms other TIRP metrics. Two approaches to predicting AHEs in real-life application conditions are evaluated: using a sliding window to continuously predict whether a patient would experience an AHE within a specific prediction time period ahead, our models produced an AUC-ROC of 82%, but with low AUPRC. Alternatively, predicting whether an AHE would generally occur during the entire admission resulted in an AUC-ROC of 74%.

Medical image segmentation is a fundamental and critical step in many clinical approaches. Semi-supervised learning has been widely applied to medical image segmentation tasks since it alleviates the heavy burden of acquiring expert-examined annotations and takes the advantage of unlabeled data which is much easier to acquire. Although consistency learning has been proven to be an effective approach by enforcing an invariance of predictions under different distributions, existing approaches cannot make full use of region-level shape constraint and boundary-level distance information from unlabeled data. In this paper, we propose a novel uncertainty-guided mutual consistency learning framework to effectively exploit unlabeled data by integrating intra-task consistency learning from up-to-date predictions for self-ensembling and cross-task consistency learning from task-level regularization to exploit geometric shape information. The framework is guided by the estimated segmentation uncertainty of models to select out relatively certain predictions for consistency learning, so as to effectively exploit more reliable information from unlabeled data. Experiments on two publicly available benchmark datasets showed that: (1) Our proposed method can achieve significant performance improvement by leveraging unlabeled data, with up to 4.13% and 9.82% in Dice coefficient compared to supervised baseline on left atrium segmentation and brain tumor segmentation, respectively. (2) Compared with other semi-supervised segmentation methods, our proposed method achieve better segmentation performance under the same backbone network and task settings on both datasets, demonstrating the effectiveness and robustness of our method and potential transferability for other medical image segmentation tasks.

Clinical event sequences consist of hundreds of clinical events that represent records of patient care in time. Developing accurate predictive models of such sequences is of a great importance for supporting a variety of models for interpreting/classifying the current patient condition, or predicting adverse clinical events and outcomes, all aimed to improve patient care. One important challenge of learning predictive models of clinical sequences is their patient-specific variability. Based on underlying clinical conditions, each patient's sequence may consist of different sets of clinical events (observations, lab results, medications, procedures). Hence, simple population-wide models learned from event sequences for many different patients may not accurately predict patient-specific dynamics of event sequences and their differences. To address the problem, we propose and investigate multiple new event sequence prediction models and methods that let us better adjust the prediction for individual patients and their specific conditions. The methods developed in this work pursue refinement of population-wide models to subpopulations, self-adaptation, and a meta-level model switching that is able to adaptively select the model with the best chance to support the immediate prediction. We analyze and test the performance of these models on clinical event sequences of patients in MIMIC-III database.

Predicting the trajectory of a disease at an early stage can aid physicians in offering effective treatment, prompt care to patients, and also avoid misdiagnosis. However, forecasting patient trajectories is challenging due to long-range dependencies, irregular intervals between consecutive admissions, and non-stationarity data. To address these challenges, we propose a novel method called Clinical-GAN, a Transformer-based Generative Adversarial Networks (GAN) to forecast the patients' medical codes for subsequent visits. First, we represent the patients' medical codes as a time-ordered sequence of tokens akin to language models. Then, a Transformer mechanism is used as a Generator to learn from existing patients' medical history and is trained adversarially against a Transformer-based Discriminator. We address the above mentioned challenges based on our data modeling and Transformer-based GAN architecture. Additionally, we enable the local interpretation of the model's prediction using a multi-head attention mechanism. We evaluated our method using a publicly available dataset, Medical Information Mart for Intensive Care IV v1.0 (MIMIC-IV), with more than 500,000 visits completed by around 196,000 adult patients over an 11-year period from 2008-2019. Clinical-GAN significantly outperforms baseline methods and existing works, as demonstrated through various experiments. Source code is at https://github.com/vigi30/Clinical-GAN.

Secondary hypertension is associated with higher risks of target organ damage and cardiovascular and cerebrovascular disease events. Early aetiology identification can eliminate aetiologies and control blood pressure. However, inexperienced doctors often fail to diagnose secondary hypertension, and comprehensively screening for all causes of high blood pressure increases health care costs. To date, deep learning has rarely been involved in the differential diagnosis of secondary hypertension. Relevant machine learning methods cannot combine textual information such as chief complaints with numerical information such as the laboratory examination results in electronic health records (EHRs), and the use of all features increases health care costs. To reduce redundant examinations and accurately identify secondary hypertension, we propose a two-stage framework that follows clinical procedures. The framework carries out an initial diagnosis process in the first stage, on which basis patients are recommended for disease-related examinations, followed by differential diagnoses of different diseases based on the different characteristics observed in the second stage. We convert the numerical examination results into descriptive sentences, thus blending textual and numerical characteristics. Medical guidelines are introduced through label embedding and attention mechanisms to obtain interactive features. Our model was trained and evaluated using a cross-sectional dataset containing 11,961 patients with hypertension from January 2013 to December 2019. The F1 scores of our model were 0.912, 0.921, 0.869 and 0.894 for primary aldosteronism, thyroid disease, nephritis and nephrotic syndrome and chronic kidney disease, respectively, which are four kinds of secondary hypertension with high incidence rates. The experimental results show that our model can powerfully use the textual and numerical data contained in EHRs to provide effective decision support for the differential diagnosis of secondary hypertension.

With mHealth apps, data can be recorded in real life, which makes them useful, for example, as an accompanying tool in treatments. However, such datasets, especially those based on apps with usage on a voluntary basis, are often affected by fluctuating engagement and by high user dropout rates. This makes it difficult to exploit the data using machine learning techniques and raises the question of whether users have stopped using the app. In this extended paper, we present a method to identify phases with varying dropout rates in a dataset and predict for each. We also present an approach to predict what period of inactivity can be expected for a user in the current state. We use change point detection to identify the phases, show how to deal with uneven misaligned time series and predict the user's phase using time series classification. In addition, we examine how the evolution of adherence develops in individual clusters of individuals. We evaluated our method on the data of an mHealth app for tinnitus, and show that our approach is appropriate for the study of adherence in datasets with uneven, unaligned time series of different lengths and with missing values.

The task of detecting common and unique characteristics among different cancer subtypes is an important focus of research that aims to improve personalized therapies. Unlike current approaches mainly based on predictive techniques, our study aims to improve the knowledge about the molecular mechanisms that descriptively led to cancer, thus not requiring previous knowledge to be validated. Here, we propose an approach based on contrast set mining to capture high-order relationships in cancer transcriptomic data. In this way, we were able to extract valuable insights from several cancer subtypes in the form of highly specific genetic relationships related to functional pathways affected by the disease. To this end, we have divided several cancer gene expression databases by the subtype associated with each sample to detect which gene groups are related to each cancer subtype. To demonstrate the potential and usefulness of the proposed approach we have extensively analysed RNA-Seq gene expression data from breast, kidney, and colon cancer subtypes. The possible role of the obtained genetic relationships was further evaluated through extensive literature research, while its prognosis was assessed via survival analysis, finding gene expression patterns related to survival in various cancer subtypes. Some gene associations were described in the literature as potential cancer biomarkers while other results have been not described yet and could be a starting point for future research.

Complexity of medical treatments can vary from prescribing medicine for a specific ailment to managing a complex set of simultaneous medical issues. In the latter case, doctors are assisted by clinical guidelines which outline standard medical procedures, tests, treatments, etc. To facilitate the use of such guidelines, they can be digitized as processes and adopted in complex process engines offering additional help to health providers such as decision support while monitoring active treatments so as to detect flaws in treatment procedures and suggest possible reactions on them. For example, a patient may present symptoms of multiple diseases simultaneously (requiring multiple clinical guidelines to be followed), while also being allergic to some often-used drugs (requiring additional constraints to be respected). This can easily lead to treating a patient based on a set of process specifications which are not fully compatible with each other. While a scenario like that commonly occurs in practice, research in that direction has thus far given little consideration to how to specify multiple clinical guidelines and how to automatically combine their specifications in the context of the monitoring task. In our previous work (Alman et al., 20222), we presented a conceptual framework for handling the above cases in the context of monitoring. In this paper, we present the algorithms necessary for implementing key components of this conceptual framework. More specifically, we provide formal languages for representing clinical guideline specifications and formalize a solution for monitoring the interplay of such specifications expressed as a combination of (data-aware) Petri nets and temporal logic rules. The proposed solution seamlessly handles combination of the input process specifications and provides both early conflict detection and decision support during process execution. We also discuss a proof-of-concept implementation of our approach and present the results of extensive scalability experiments.

Sepsis is one of the most challenging health conditions worldwide, with relatively high incidence and mortality rates. It is shown that preventing sepsis is the key to avoid potentially irreversible organ dysfunction. However, data-driven early identification of sepsis is challenging as sepsis shares signs and symptoms with other health conditions. This paper adopts a temporal pattern mining approach to identify frequent temporal and evolving patterns of physiological and biological biomarkers in sepsis patients. We show that using these frequent patterns as features for classifying sepsis and non-sepsis patients can improve the prediction accuracy and performance up to 7%. Most of the temporal modeling approaches adopted in the sepsis literature are based on deep learning methods. Although these approaches produce high accuracy, they generally have limited model explainability and interpretability. Using the adopted methods in this study, we could identify the most important features contributing to the patients' sepsis incidence, such as fluctuations in platelet, lactate, and creatinine, or evolution of patterns including renal and metabolic organ systems, and consequently, enhance the findings' clinical interpretability.

The wide adoption of electronic health records (EHRs) offers immense potential as a source of support for clinical research. However, previous studies focused on extracting only a limited set of medical concepts to support information extraction in the cancer domain for the Spanish language. Building on the success of deep learning for processing natural language texts, this paper proposes a transformer-based approach to extract named entities from breast cancer clinical notes written in Spanish and compares several language models. To facilitate this approach, a schema for annotating clinical notes with breast cancer concepts is presented, and a corpus for breast cancer is developed. Results indicate that both BERT-based and RoBERTa-based language models demonstrate competitive performance in clinical Named Entity Recognition (NER). Specifically, BETO and multilingual BERT achieve F-scores of 93.71% and 94.63%, respectively. Additionally, RoBERTa Biomedical attains an F-score of 95.01%, while RoBERTa BNE achieves an F-score of 94.54%. The findings suggest that transformers can feasibly extract information in the clinical domain in the Spanish language, with the use of models trained on biomedical texts contributing to enhanced results. The proposed approach takes advantage of transfer learning techniques by fine-tuning language models to automatically represent text features and avoiding the time-consuming feature engineering process.

The medication recommendation (MR) or medication combination prediction task aims to predict effective prescriptions given accurate patient representations derived from electronic health records (EHRs), which contributes to improving the quality of clinical decision-making, especially for patients with multi-morbidity. Although in recent years deep learning technology has achieved great success in MR, the performance of current multi-label based MR solutions is unsatisfactory. They mainly focus on improving the patient representation module and modeling the medication label dependencies such as drug-drug interaction (DDI) correlation and co-occurrence relationship. However, the hierarchical dependency among medication labels and diversity of difficulty among MR training examples lack sufficient consideration. In this paper, we propose a framework of Curriculum learning Enhanced Hierarchical multi-label classification for MR (CEHMR). Motivated by the category hierarchy of medications which organizes standard medication codes in a hierarchical structure, we utilize it to provide more trustworthy prior knowledge for modeling label dependency. Specifically, we design a hierarchical multi-label classifier with a learnable gate fusion layer, to simultaneously capture the level-independent (local) and level-dependent (global) hierarchical information in the medication hierarchy. In addition, to overcome the diversity of training example difficulties, and progressively achieve a smoother training process, we introduce a bootstrap-based curriculum learning strategy. Hence, the example difficulty can be measured based on the predictive performance of the MR model, and then all training examples would be retrained from easy to hard under the guidance of a predefined training scheduler. Experiments on the real-world medical MIMIC-III database demonstrate that the proposed framework can achieve state-of-the-art performance compared with seven representative baselines, and extensive ablation studies validate the effectiveness of each component of CEHMR.

Bacterial resistance to antibiotics has been rapidly increasing, resulting in low antibiotic effectiveness even treating common infections. The presence of resistant pathogens in environments such as a hospital Intensive Care Unit (ICU) exacerbates the critical admission-acquired infections. This work focuses on the prediction of antibiotic resistance in Pseudomonas aeruginosa nosocomial infections at the ICU, using Long Short-Term Memory (LSTM) artificial neural networks as the predictive method. The analyzed data were extracted from the Electronic Health Records (EHR) of patients admitted to the University Hospital of Fuenlabrada from 2004 to 2019 and were modeled as Multivariate Time Series. A data-driven dimensionality reduction method is built by adapting three feature importance techniques from the literature to the considered data and proposing an algorithm for selecting the most appropriate number of features. This is done using LSTM sequential capabilities so that the temporal aspect of features is taken into account. Furthermore, an ensemble of LSTMs is used to reduce the variance in performance. Our results indicate that the patient's admission information, the antibiotics administered during the ICU stay, and the previous antimicrobial resistance are the most important risk factors. Compared to other conventional dimensionality reduction schemes, our approach is able to improve performance while reducing the number of features for most of the experiments. In essence, the proposed framework achieve, in a computationally cost-efficient manner, promising results for supporting decisions in this clinical task, characterized by high dimensionality, data scarcity, and concept drift.

INTRODUCTION: Artificial Intelligence-based Medical Devices (AI-based MDs) are experiencing exponential growth in healthcare. This study aimed to investigate whether current studies assessing AI contain the information required for health technology assessment (HTA) by HTA bodies. METHODS: We conducted a systematic literature review based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses methodology to extract articles published between 2016 and 2021 related to the assessment of AI-based MDs. Data extraction focused on study characteristics, technology, algorithms, comparators, and results. AI quality assessment and HTA scores were calculated to evaluate whether the items present in the included studies were concordant with the HTA requirements. We performed a linear regression for the HTA and AI scores with the explanatory variables of the impact factor, publication date, and medical specialty. We conducted a univariate analysis of the HTA score and a multivariate analysis of the AI score with an alpha risk of 5 %. RESULTS: Of 5578 retrieved records, 56 were included. The mean AI quality assessment score was 67 %; 32 % of articles had an AI quality score ≥ 70 %, 50 % had a score between 50 % and 70 %, and 18 % had a score under 50 %. The highest quality scores were observed for the study design (82 %) and optimisation (69 %) categories, whereas the scores were lowest in the clinical practice category (23 %). The mean HTA score was 52 % for all seven domains. 100 % of the studies assessed clinical effectiveness, whereas only 9 % evaluated safety, and 20 % evaluated economic issues. There was a statistically significant relationship between the impact factor and the HTA and AI scores (both p = 0.046). DISCUSSION: Clinical studies on AI-based MDs have limitations and often lack adapted, robust, and complete evidence. High-quality datasets are also required because the output data can only be trusted if the inputs are reliable. The existing assessment frameworks are not specifically designed to assess AI-based MDs. From the perspective of regulatory authorities, we suggest that these frameworks should be adapted to assess the interpretability, explainability, cybersecurity, and safety of ongoing updates. From the perspective of HTA agencies, we highlight that transparency, professional and patient acceptance, ethical issues, and organizational changes are required for the implementation of these devices. Economic assessments of AI should rely on a robust methodology (business impact or health economic models) to provide decision-makers with more reliable evidence. CONCLUSION: Currently, AI studies are insufficient to cover HTA prerequisites. HTA processes also need to be adapted because they do not consider the important specificities of AI-based MDs. Specific HTA workflows and accurate assessment tools should be designed to standardise evaluations, generate reliable evidence, and create confidence.

Medical information extraction consists of a group of natural language processing (NLP) tasks, which collaboratively convert clinical text to pre-defined structured formats. This is a critical step to exploit electronic medical records (EMRs). Given the recent thriving NLP technologies, model implementation and performance seem no longer an obstacle, whereas the bottleneck locates on a high-quality annotated corpus and the whole engineering workflow. This study presents an engineering framework consisting of three tasks, i.e., medical entity recognition, relation extraction and attribute extraction. Within this framework, the whole workflow is demonstrated from EMR data collection through model performance evaluation. Our annotation scheme is designed to be comprehensive and compatible between the multiple tasks. With the EMRs from a general hospital in Ningbo, China, and the manual annotation by experienced physicians, our corpus is of large scale and high quality. Built upon this Chinese clinical corpus, the medical information extraction system show performance that approaches human annotation. The annotation scheme, (a subset of) the annotated corpus, and the code are all publicly released, to facilitate further research.

The Human Phenotype Ontology (HPO) is a dictionary of >15,000 clinical phenotypic terms with defined semantic relationships, developed to standardize phenotypic analysis. Over the last decade, the HPO has been used to accelerate the implementation of precision medicine into clinical practice. In addition, recent research in representation learning, specifically in graph embedding, has led to notable progress in automated prediction via learned features. Here, we present a novel approach to phenotype representation by incorporating phenotypic frequencies based on 53 million full-text health care notes from >1.5 million individuals. We demonstrate the efficacy of our proposed phenotype embedding technique by comparing our work to existing phenotypic similarity-measuring methods. Using phenotype frequencies in our embedding technique, we are able to identify phenotypic similarities that surpass current computational models. Furthermore, our embedding technique exhibits a high degree of agreement with domain experts' judgment. By transforming complex and multidimensional phenotypes from the HPO format into vectors, our proposed method enables efficient representation of these phenotypes for downstream tasks that require deep phenotyping. This is demonstrated in a patient similarity analysis and can further be applied to disease trajectory and risk prediction.

Text-Based Medical Image Retrieval (TBMIR) has been known to be successful in retrieving medical images with textual descriptions. Usually, these descriptions are very brief and cannot express the whole visual content of the image in words, hence negatively affect the retrieval performance. One of the solutions offered in the literature is to form a Bayesian Network thesaurus taking advantage of some medical terms extracted from the image datasets. Despite the interestingness of this solution, it is not efficient as it is highly related to the co-occurrence measure, the layer arrangement and the arc directions. A significant drawback of the co-occurrence measure is the generation of a lot of uninteresting co-occurring terms. Several studies applied the association rules mining and its measures to discover the correlation between the terms. In this paper, we propose a new efficient association Rule Based Bayesian Network (R2BN) model for TBMIR using updated medically-dependent features (MDF) based on Unified Medical Language System (UMLS). The MDF are a set of medical terms that refers to the imaging modalities, the image color, the searched object dimension, etc. The proposed model presents the association rules mined from MDF in the form of Bayesian Network model. Then, it exploits the association rule measures (support, confidence, and lift) to prune the Bayesian Network model for efficient computation. The proposed R2BN model is combined with a literature probabilistic model to predict the relevance of an image to a given query. Experiments are carried out with ImageCLEF medical retrieval task collections from 2009 to 2013. Results show that our proposed model enhances significantly the image retrieval accuracy compared to the state-of-the-art retrieval models.

BACKGROUND: Stroke is one of the leading causes of death and disability worldwide. The National Institutes of Health Stroke Scale (NIHSS) scores in electronic health records (EHRs), which quantitatively describe patients' neurological deficits in evidence-based treatment, are crucial in stroke-related clinical investigations. However, the free-text format and lack of standardization inhibit their effective use. Automatically extracting the scale scores from the clinical free text so that its potential value in real-world studies is realized has become an important goal. OBJECTIVE: This study aims to develop an automated method to extract scale scores from the free text of EHRs. METHODS: We propose a two-step pipeline method to identify NIHSS items and numerical scores and validate its feasibility using a freely accessible critical care database: MIMIC-III (Medical Information Mart for Intensive Care III). First, we utilize MIMIC-III to create an annotated corpus. Then, we investigate possible machine learning methods for two subtasks, NIHSS item and score recognition and item-score relation extraction. In the evaluation, we conduct both task-specific and end-to-end evaluations and compare our method with the rule-based method using precision, recall and F1 scores as evaluation metrics. RESULTS: We use all available discharge summaries of stroke cases in MIMIC-III. The annotated NIHSS corpus contains 312 cases, 2929 scale items, 2774 scores and 2733 relations. The results show that the best F1-score of our method was 0.9006, which was attained by combining BERT-BiLSTM-CRF and Random Forest, and it outperformed the rule-based method (F1-score = 0.8098). In the end-to-end task, our method could successfully recognize the item "1b level of consciousness questions", the score "1" and their relation "('1b level of consciousness questions', '1', 'has value')" from the sentence "1b level of consciousness questions: said name = 1", while the rule-based method could not. CONCLUSIONS: The two-step pipeline method we propose is an effective approach to identify NIHSS items, scores and their relations. With its help, clinical investigators can easily retrieve and access structured scale data, thereby supporting stroke-related real-world studies.

Protein-protein interaction is one of the ways viruses interact with their hosts. Therefore, identifying protein interactions between viruses and hosts helps explain how virus proteins work, how they replicate, and how they cause disease. SARS-CoV-2 is a new type of virus that emerged from the coronavirus family in 2019 and caused a worldwide pandemic. Detection of human proteins interacting with this novel virus strain plays an important role in monitoring the cellular process of virus-associated infection. Within the scope of the study, a natural language processing-based collective learning method is proposed for the prediction of potential SARS-CoV-2-human PPIs. Protein language models were obtained with the prediction-based word2Vec and doc2Vec embedding methods and the frequency-based tf-idf method. Known interactions were represented by proposed language models and traditional feature extraction methods (conjoint triad and repeat pattern), and their performances were compared. The interaction data were trained with support vector machine, artificial neural network (ANN), k-nearest neighbor (KNN), naive Bayes (NB), decision tree (DT), and ensemble algorithms. Experimental results show that protein language models are a promising protein representation method for protein-protein interaction prediction. The term frequency-inverse document frequency-based language model performed the SARS-CoV-2 protein-protein interaction estimation with an error of 1.4%. Additionally, the decisions of high-performing learning models for different feature extraction methods were combined with a collective voting approach to make new interaction predictions. For 10,000 human proteins, 285 new potential interactions were predicted, with models combining decisions.

Despite the advancements in the diagnosis of early-stage cirrhosis, the accuracy in the diagnosis using ultrasound is still challenging owing to the presence of various image artifacts, which results in poor visual quality of the textural and lower-frequency components. In this study, we propose an end-to-end multistep network called CirrhosisNet that includes two transfer-learned convolutional neural networks for semantic segmentation and classification tasks. It uses a uniquely designed image, called an aggregated micropatch (AMP), as an input image to the classification network, thereby assessing whether the liver is in a cirrhotic stage. With a prototype AMP image, we synthesized a bunch of AMP images while retaining the textural features. This synthesis significantly increases the number of insufficient cirrhosis-labeled images, thereby circumventing overfitting issues and optimizing network performance. Furthermore, the synthesized AMP images contained unique textural patterns, mostly generated on the boundaries between adjacent micropatches (μ-patches) during their aggregation. These newly created boundary patterns provide rich information regarding the texture features of the ultrasound image, thereby making cirrhosis diagnosis more accurate and sensitive. The experimental results demonstrated that our proposed AMP image synthesis is extremely effective in expanding the dataset of cirrhosis images, thus diagnosing liver cirrhosis with considerably high accuracy. We achieved an accuracy of 99.95 %, a sensitivity of 100 %, and a specificity of 99.9 % on the Samsung Medical Center dataset using 8 × 8 pixels-sized μ-patches. The proposed approach provides an effective solution to deep-learning models with limited-training data, such as medical imaging tasks.

Classification is one of the most significant subfields of data mining that has been successfully applied to various applications. The literature has expended substantial effort to present more efficient and accurate classification models. Despite the diversity of the proposed models, they were all created using the same methodology, and their learning processes ignored a fundamental issue. In all existing classification model learning processes, a continuous distance-based cost function is optimized to estimate the unknown parameters. The classification problem's objective function is discrete. Consequently, applying a continuous cost function to a classification problem with a discrete objective function is illogical or inefficient. This paper proposes a novel classification methodology utilizing a discrete cost function in the learning process. To this end, one of the most popular intelligent classification models, the multilayer perceptron (MLP), is used to implement the proposed methodology. Theoretically, the classification performance of the proposed discrete learning-based MLP (DIMLP) model is not dissimilar to that of its continuous learning-based counterpart. Nevertheless, in this study, to demonstrate the efficacy of the DIMLP model, it was applied to several breast cancer classification datasets, and its classification rate was compared to that of the conventional continuous learning-based MLP model. The empirical results indicate that the proposed DIMLP model outperforms the MLP model across all datasets. The results demonstrate that the presented DIMLP classification model achieves an average classification rate of 94.70 %, a 6.95 % improvement over the classification rate of the traditional MLP model, which was 88.54 %. Therefore, the classification approach proposed in this study can be utilized as an alternative learning process in intelligent classification methods for medical decision-making and other classification applications, particularly when more accurate results are required.

Accurate airway segmentation from computed tomography (CT) images is critical for planning navigation bronchoscopy and realizing a quantitative assessment of airway-related chronic obstructive pulmonary disease (COPD). Existing methods face difficulty in airway segmentation, particularly for the small branches of the airway. These difficulties arise due to the constraints of limited labeling and failure to meet clinical use requirements in COPD. We propose a two-stage framework with a novel 3D contextual transformer for segmenting the overall airway and small airway branches using CT images. The method consists of two training stages sharing the same modified 3D U-Net network. The novel 3D contextual transformer block is integrated into both the encoder and decoder path of the network to effectively capture contextual and long-range information. In the first training stage, the proposed network segments the overall airway with the overall airway mask. To improve the performance of the segmentation result, we generate the intrapulmonary airway branch label, and train the network to focus on producing small airway branches in the second training stage. Extensive experiments were performed on in-house and multiple public datasets. Quantitative and qualitative analyses demonstrate that our proposed method extracts significantly more branches and longer lengths of the airway tree while accomplishing state-of-the-art airway segmentation performance. The code is available at https://github.com/zhaozsq/airway_segmentation.

The outbreak of COVID-19 pandemic poses new challenges to research community to investigate novel mechanisms for monitoring as well as controlling its further spread via crowded scenes. Moreover, the contemporary methods of COVID-19 preventions are enforcing strict protocols in the public places. The emergence of robust computer vision-enabled applications leverages intelligent frameworks for monitoring of the pandemic deterrence in public places. The employment of COVID-19 protocols via wearing face masks by human is an effective procedure that is implemented in several countries across the world. It is a challenging task for authorities to manually monitor these protocols particularly in densely crowded public gatherings such as, shopping malls, railway stations, airports, religious places etc. Thus, to overcome these issues, the proposed research aims to design an operative method that automatically detects the violation of face mask regulation for COVID-19 pandemic. In this research work, we expound a novel technique for COVID-19 protocol desecration via video summarization in the crowded scenes (CoSumNet). Our approach automatically yields short summaries from crowded video scenes (i.e., with and without mask human). Besides, the CoSumNet can be deployed in crowded places that may assist the controlling agencies to take appropriate actions to enforce the penalty to the protocol violators. To evaluate the efficacy of the approach, the CoSumNet is trained on a benchmark "Face Mask Detection ∼12K Images Dataset" and validated through various real-time CCTV videos. The CoSumNet demonstrates superior performance of 99.98 % and 99.92 % detection accuracy in the seen and unseen scenarios respectively. Our method offers promising performance in cross-datasets environments as well as on a variety of face masks. Furthermore, the model can convert the longer videos to short summaries in nearly 5-20 s approximately.

Clinical practice guidelines (CPGs) are patient management tools that synthesize medical knowledge into an actionable format. CPGs are disease specific with limited applicability to the management of complex patients suffering from multimorbidity. For the management of these patients, CPGs need to be augmented with secondary medical knowledge coming from a variety of knowledge repositories. The operationalization of this knowledge is key to increasing CPGs' uptake in clinical practice. In this work, we propose an approach to operationalizing secondary medical knowledge inspired by graph rewriting. We assume that the CPGs can be represented as task network models, and provide an approach for representing and applying codified medical knowledge to a specific patient encounter. We formally define revisions that model and mitigate adverse interactions between CPGs and we use a vocabulary of terms to instantiate these revisions. We demonstrate the application of our approach using synthetic and clinical examples. We conclude by identifying areas for future work with the vision of developing a theory of mitigation that will facilitate the development of comprehensive decision support for the management of multimorbid patients.

Current models on Explainable Artificial Intelligence (XAI) have shown a lack of reliability when evaluating feature-relevance for deep neural biomarker classifiers. The inclusion of reliable saliency-maps for obtaining trustworthy and interpretable neural activity is still insufficiently mature for practical applications. These limitations impede the development of clinical applications of Deep Learning. To address, these limitations we propose the RemOve-And-Retrain (ROAR) algorithm which supports the recovery of highly relevant features from any pre-trained deep neural network. In this study we evaluated the ROAR methodology and algorithm for the Face Emotion Recognition (FER) task, which is clinically applicable in the study of Autism Spectrum Disorder (ASD). We trained a Convolutional Neural Network (CNN) from electroencephalography (EEG) signals and assessed the relevance of FER-elicited EEG features from individuals diagnosed with and without ASD. Specifically, we compared the ROAR reliability from well-known relevance maps such as Layer-Wise Relevance Propagation, PatternNet, Pattern-Attribution, and Smooth-Grad Squared. This study is the first to bridge previous neuroscience and ASD research findings to feature-relevance calculation for EEG-based emotion recognition with CNN in typically-development (TD) and in ASD individuals.

BACKGROUND: In everyday medical practice, the results of cardiac ultrasound examinations are generally recorded in unstructured text, from which extracting relevant information is an important and challenging task. This paper presents a generally applicable language and corpus-independent text mining method for extracting and structuring numerical measurement results and their descriptions from echocardiography reports. METHOD: The developed method is based on generally applicable text mining preprocessing activities, it automatically identifies and standardizes the descriptions of the cardiac ultrasound measures, and it stores the extracted and standardized measurement descriptions with their measurement results in a structured form for later usage. The method does not contain any regular expression-based search and does not rely on information about the structure of the document. RESULTS: The method has been tested on a document set containing more than 20,000 echocardiographic reports by examining the efficiency of extracting 12 echocardiography parameters considered important by experts. The method extracted and structured the echocardiography parameters under the study with good sensitivity (lowest value: 0.775, highest value: 1.0, average: 0.904) and excellent specificity (for all cases 1.0). The F1 score ranged between 0.873 and 1.0, and its average value was 0.948. CONCLUSION: The presented case study has shown that the proposed method can extract measurement results from echocardiography documents with high confidence without performing a direct search or having detailed information about the data recording habits. Furthermore, it effectively handles spelling errors, abbreviations and the highly varied terminology used in descriptions. As it does not rely on any information related to the structure or the language of the documents or data recording habits, it can be applied for processing any free-text written medical texts.

Cardiovascular diseases account for 17 million deaths per year worldwide. Of these, 25% are categorized as sudden cardiac death, which can be related to ventricular tachycardia (VT). This type of arrhythmia can be caused by focal activation sources outside the sinus node. Catheter ablation of these foci is a curative treatment in order to inactivate the abnormal triggering activity. However, the localization procedure is usually time-consuming and requires an invasive procedure in the catheter lab. To facilitate and expedite the treatment, we present two novel localization support techniques based on convolutional neural networks (CNNs) that address these clinical needs. In contrast to existing methods, our approaches were designed to be independent of the patient-specific geometry and directly applicable to surface ECG signals, while also delivering a binary transmural position. Moreover, one of the method's outputs can be interpreted as several ranked solutions. The CNNs were trained on a dataset containing only simulated data and evaluated both on simulated test data and clinical data. On a novel large and open simulated dataset, the median test error was below 3 mm. The median localization error on the unseen clinical data ranged from 32 mm to 41 mm without optimizing the pre-processing and CNN to the clinical data. Interpreting the output of one of the approaches as ranked solutions, the best median error of the top-3 solutions decreased to 20 mm on the clinical data. The transmural position was correctly detected in up to 82% of all clinical cases. These results demonstrate a proof of principle to utilize CNNs to localize the activation source without the intrinsic need for patient-specific geometrical information. Furthermore, providing multiple solutions can assist physicians in identifying the true activation source amongst more than one possible location. With further optimization to clinical data, these methods have high potential to accelerate clinical interventions, replace certain steps within these procedures and consequently reduce procedural risk and improve VT patient outcomes.

In this paper we investigate which airborne pollutants have a short-term causal effect on cardiovascular and respiratory disease using the Ancestral Probabilities (AP) procedure, a novel Bayesian approach for deriving the probabilities of causal relationships from observational data. The results are largely consistent with EPA assessments of causality, however, in a few cases AP suggests that some pollutants thought to cause cardiovascular or respiratory disease are associated due purely to confounding. The AP procedure utilizes maximal ancestral graph (MAG) models to represent and assign probabilities to causal relationships while accounting for latent confounding. The algorithm does so locally by marginalizing over models with and without causal features of interest. Before applying AP to real data, we evaluate it in a simulation study and investigate the benefits of providing background knowledge. Overall, the results suggest that AP is an effective tool for causal discovery.

Image texture analysis has for decades represented a promising opportunity for cancer assessment and disease progression evaluation, evolving in a discipline, i.e., radiomics. However, the road to a complete translation into clinical practice is still hampered by intrinsic limitations. As purely supervised classification models fail in devising robust imaging-based biomarkers for prognosis, cancer subtyping approaches would benefit from the employment of distant supervision, for instance exploiting survival/recurrence information. In this work, we assessed, tested, and validated the domain-generality of our previously proposed Distant Supervised Cancer Subtyping model on Hodgkin Lymphoma. We evaluate the model performance on two independent datasets coming from two hospitals, comparing and analyzing the results. Although successful and consistent, the comparison confirmed the instability of radiomics due to an across-center lack of reproducibility, leading to explainable results in one center and poor interpretability in the other. We thus propose a Random Forest-based Explainable Transfer Model for testing the domain-invariance of imaging biomarkers extracted from retrospective cancer subtyping. In doing so, we tested the predictive ability of cancer subtyping in a validation and perspective setting, which led to successful results and supported the domain-generality of the proposed approach. On the other hand, the extraction of decision rules enables to draw of risk factors and robust biomarkers to inform clinical decisions. This work shows the potentialities of the Distant Supervised Cancer Subtyping model to be further evaluated in larger multi-center datasets, to reliably translate radiomics into medical practice. The code is available at this GitHub repository.

Deep learning models for scoring sleep stages based on single-channel EEG have been proposed as a promising method for remote sleep monitoring. However, applying these models to new datasets, particularly from wearable devices, raises two questions. First, when annotations on a target dataset are unavailable, which different data characteristics affect the sleep stage scoring performance the most and by how much? Second, when annotations are available, which dataset should be used as the source of transfer learning to optimize performance? In this paper, we propose a novel method for computationally quantifying the impact of different data characteristics on the transferability of deep learning models. Quantification is accomplished by training and evaluating two models with significant architectural differences, TinySleepNet and U-Time, under various transfer configurations in which the source and target datasets have different recording channels, recording environments, and subject conditions. For the first question, the environment had the highest impact on sleep stage scoring performance, with performance degrading by over 14% when sleep annotations were unavailable. For the second question, the most useful transfer sources for TinySleepNet and the U-Time models were MASS-SS1 and ISRUC-SG1, containing a high percentage of N1 (the rarest sleep stage) relative to the others. The frontal and central EEGs were preferred for TinySleepNet. The proposed approach enables full utilization of existing sleep datasets for training and planning model transfer to maximize the sleep stage scoring performance on a target problem when sleep annotations are limited or unavailable, supporting the realization of remote sleep monitoring.

Evolutionary algorithms have been successfully employed to find the best structure for many learning algorithms including neural networks. Due to their flexibility and promising results, Convolutional Neural Networks (CNNs) have found their application in many image processing applications. The structure of CNNs greatly affects the performance of these algorithms both in terms of accuracy and computational cost, thus, finding the best architecture for these networks is a crucial task before they are employed. In this paper, we develop a genetic programming approach for the optimization of CNN structure in diagnosing COVID-19 cases via X-ray images. A graph representation for CNN architecture is proposed and evolutionary operators including crossover and mutation are specifically designed for the proposed representation. The proposed architecture of CNNs is defined by two sets of parameters, one is the skeleton which determines the arrangement of the convolutional and pooling operators and their connections and one is the numerical parameters of the operators which determine the properties of these operators like filter size and kernel size. The proposed algorithm in this paper optimizes the skeleton and the numerical parameters of the CNN architectures in a co-evolutionary scheme. The proposed algorithm is used to identify covid-19 cases via X-ray images.

Medical experts may use Artificial Intelligence (AI) systems with greater trust if these are supported by 'contextual explanations' that let the practitioner connect system inferences to their context of use. However, their importance in improving model usage and understanding has not been extensively studied. Hence, we consider a comorbidity risk prediction scenario and focus on contexts regarding the patients' clinical state, AI predictions about their risk of complications, and algorithmic explanations supporting the predictions. We explore how relevant information for such dimensions can be extracted from Medical guidelines to answer typical questions from clinical practitioners. We identify this as a question answering (QA) task and employ several state-of-the-art Large Language Models (LLM) to present contexts around risk prediction model inferences and evaluate their acceptability. Finally, we study the benefits of contextual explanations by building an end-to-end AI pipeline including data cohorting, AI risk modeling, post-hoc model explanations, and prototyped a visual dashboard to present the combined insights from different context dimensions and data sources, while predicting and identifying the drivers of risk of Chronic Kidney Disease (CKD) - a common type-2 diabetes (T2DM) comorbidity. All of these steps were performed in deep engagement with medical experts, including a final evaluation of the dashboard results by an expert medical panel. We show that LLMs, in particular BERT and SciBERT, can be readily deployed to extract some relevant explanations to support clinical usage. To understand the value-add of the contextual explanations, the expert panel evaluated these regarding actionable insights in the relevant clinical setting. Overall, our paper is one of the first end-to-end analyses identifying the feasibility and benefits of contextual explanations in a real-world clinical use case. Our findings can help improve clinicians' usage of AI models.

Early melanoma diagnosis is the most important factor in the treatment of skin cancer and can effectively reduce mortality rates. Recently, Generative Adversarial Networks have been used to augment data, prevent overfitting and improve the diagnostic capacity of models. However, its application remains a challenging task due to the high levels of inter and intra-class variance seen in skin images, limited amounts of data, and model instability. We present a more robust Progressive Growing of Adversarial Networks based on residual learning, which is highly recommended to ease the training of deep networks. The stability of the training process was increased by receiving additional inputs from preceding blocks. The architecture is able to produce plausible photorealistic synthetic 512 × 512 skin images, even with small dermoscopic and non-dermoscopic skin image datasets as problem domains. In this manner, we tackle the lack of data and the imbalance problems. Additionally, the proposed approach leverages a skin lesion boundary segmentation algorithm and transfer learning to enhance the diagnosis of melanoma. Inception score and Matthews Correlation Coefficient were used to measure the performance of the models. The architecture was evaluated qualitatively and quantitatively through the use of an extensive experimental study on sixteen datasets, illustrating its effectiveness in the diagnosis of melanoma. Finally, four state-of-the-art data augmentation techniques applied in five convolutional neural network models were significantly outperformed. The results indicated that a bigger number of trainable parameters will not necessarily obtain a better performance in melanoma diagnosis.

Medical Subject Headings (MeSH) is a hierarchically structured thesaurus created by the National Library of Medicine of USA. Each year the vocabulary gets revised, bringing forth different types of changes. Those of particular interest are the ones that introduce new descriptors in the vocabulary either brand new or those who come up as a product of a complex change. These new descriptors often lack ground truth articles and rendering learning models that require supervision not applicable. Furthermore, this problem is characterized by its multi label nature and the fine-grained character of the descriptors that play the role of classes, requiring expert supervision and a lot of human resources. In this work, we alleviate these issues through retrieving insights from provenance information about those descriptors present in MeSH to create a weakly labeled train set for them. At the same time, we make use of a similarity mechanism to further filter the weak labels obtained through the descriptor information mentioned earlier. Our method, called WeakMeSH, was applied on a large-scale subset of the BioASQ 2018 data set consisting of 900 thousand biomedical articles. The performance of our method was evaluated on BioASQ 2020 against several other approaches that had given competitive results in similar problems in the past, or apply alternative transformations against the proposed one, as well as some variants that showcase the importance of each different component of our proposed approach. Finally, an analysis was performed on the different MeSH descriptors each year to assess the applicability of our method on the thesaurus.

The onset of cancer disease is a traumatic experience for both patients and their families that suddenly change the patient's life and is accompanied by important physical, emotional, and psycho-social problems. The complexity of this scenario has been exacerbated by the COVID-19 pandemic which dramatically affected the continuity of the provision of optimal care to chronic patients. Telemedicine can support the management of oncology care paths by furnishing a suite of effective and efficient tools to monitor the therapies of cancer patients. In particular, this is a suitable setting for therapies that are administered at home. In this paper, we present an AI-based system, called Arianna, designed and implemented to support and monitor patients treated by the professionals belonging to the Breast Cancer Unit Network (BCU-Net) along the entire clinical path of breast cancer treatment. We describe in this work the three modules composing the Arianna system (the tools for patients and clinicians, and the symbolic AI-based module). The system has been validated in a qualitative way and we demonstrated how the Arianna solution reached a high level of acceptability by all types of end-users by making it suitable for a concrete integration into the daily practice of the BCU-Net.

Symbolic learning is the logic-based approach to machine learning, and its mission is to provide algorithms and methodologies to extract logical information from data and express it in an interpretable way. Interval temporal logic has been recently proposed as a suitable tool for symbolic learning, specifically via the design of an interval temporal logic decision tree extraction algorithm. In order to improve their performances, interval temporal decision trees can be embedded into interval temporal random forests, mimicking the corresponding schema at the propositional level. In this article we consider a dataset of cough and breath sample recordings of volunteer subjects, labeled with their COVID-19 status, originally collected by the University of Cambridge. By interpreting such recordings as multivariate time series, we study the problem of their automated classification using interval temporal decision trees and forests. While this problem has been approached with the same dataset as well as with other datasets, in all cases, non-symbolic learning methods (usually, deep learning-based) have been applied to solve it; in this article we apply a symbolic approach, and show that it does not only outperform the state-of-the-art obtained with the same dataset, but its results are also superior to those of most non-symbolic techniques applied on other datasets. As an added bonus, thanks to the symbolic nature of our approach, we are also able to extract explicit knowledge to help physicians characterize typical COVID-positive cough and breath.

Nowadays it is increasingly important in many applications to understand how different factors influence a variable of interest in a predictive modeling process. This task becomes particularly important in the context of Explainable Artificial Intelligence. Knowing the relative impact of each variable on the output allows us to acquire more information about the problem and about the output provided by a model. This paper proposes a new methodology, XAIRE, that determines the relative importance of input variables in a prediction environment, considering multiple prediction models in order to increase generality and avoid bias inherent in a particular learning algorithm. Concretely, we present an ensemble-based methodology that promotes the aggregation of results from several prediction methods to obtain a relative importance ranking. Also, statistical tests are considered in the methodology in order to reveal significant differences between the relative importance of the predictor variables. As a case study, XAIRE is applied to the arrival of patients in a Hospital Emergency Department, which has resulted in one of the largest sets of different predictor variables in the literature. Results show the extracted knowledge related to the relative importance of the predictors involved in the case study.

The increasing reliance on mobile health for managing disease conditions has opened a new frontier in digital health, thus, the need for understanding what constitutes positive and negative sentiments of the various apps. This paper relies on Embedded Deep Neural Networks (E-DNN), Kmeans, and Latent Dirichlet Allocation (LDA) for predicting the sentiments of diabetes mobile apps users and identifying the themes and sub-themes of positive and negative sentimental users. A total of 38,640 comments from 39 diabetes mobile apps obtained from the google play store are analyzed and accuracy of 87.67 % ± 2.57 % was obtained from a 10-fold leave-one-out cross-validation. This accuracy is 2.95 % - 18.71 % better than other predominant algorithms used for sentiment analysis and 3.47 % - 20.17 % better than the results obtained by previous researchers. The study also identified the challenges of diabetes mobile apps usage to include safety and security issues, outdated information for diabetes management, clumsy user interface, and difficulty controlling operations. The positives of the apps are ease of operation, lifestyle management, effectiveness in communication and control, and data management capabilities.

BACKGROUND: Proper maintenance of hypnosis is crucial for ensuring the safety of patients undergoing surgery. Accordingly, indicators, such as the Bispectral index (BIS), have been developed to monitor hypnotic levels. However, the black-box nature of the algorithm coupled with the hardware makes it challenging to understand the underlying mechanisms of the algorithms and integrate them with other monitoring systems, thereby limiting their use. OBJECTIVE: We propose an interpretable deep learning model that forecasts BIS values 25 s in advance using 30 s electroencephalogram (EEG) data. MATERIAL AND METHODS: The proposed model utilized EEG data as a predictor, which is then decomposed into amplitude and phase components using fast Fourier Transform. An attention mechanism was applied to interpret the importance of these components in predicting BIS. The predictability of the model was evaluated on both regression and binary classification tasks, where the former involved predicting a continuous BIS value, and the latter involved classifying a dichotomous status at a BIS value of 60. To evaluate the interpretability of the model, we analyzed the attention values expressed in the amplitude and phase components according to five ranges of BIS values. The proposed model was trained and evaluated using datasets collected from two separate medical institutions. RESULTS AND CONCLUSION: The proposed model achieved excellent performance on both the internal and external validation datasets. The model achieved a root-mean-square error of 6.614 for the regression task, and an area under the receiver operating characteristic curve of 0.937 for the binary classification task. Interpretability analysis provided insight into the relationship between EEG frequency components and BIS values. Specifically, the attention mechanism revealed that higher BIS values were associated with increased amplitude attention values in high-frequency bands and increased phase attention values in various frequency bands. This finding is expected to facilitate a more profound understanding of the BIS prediction mechanism, thereby contributing to the advancement of anesthesia technologies.

Osteoporosis is a bone-related disease characterized by decreased bone density and mass, leading to brittle fractures. Osteoporosis assessment from radiographs using a deep learning algorithm has proven a low-cost alternative to the golden standard DXA. Due to the considerable noise and low contrast, automated diagnosis of osteoporosis in X-ray images still poses a significant challenge for traditional diagnostic methods. In this paper, an end-to-end transformer-style network was proposed, termed FCoTNet, to overcome the shortcoming of insufficient fusion of texture information and local features in the traditional CoTNet. To extract complementary geometric representations at each scale of the transformer module, we integrated parallel multi-scale feature extraction architectures in each unit layer of FCoTNet to utilize convolution to aggregate features from different receptive fields. Moreover, in order to extract small-scale texture features which were more critical to the diagnosis of osteoporosis in radiographs, larger fusion weights were assigned to the feature maps with small-size receptive fields. Afterward, the multi-scale global modeling was conducted by self-attention mechanism. The proposed model was first investigated on a private lumbar spine X-ray dataset with the 5-fold cross-validation strategy, obtaining an average accuracy of 78.29 ± 0.93 %, an average sensitivity of 69.72 ± 2.35 %, and an average specificity of 88.92 ± 0.67 % for the multi-classification of normal, osteopenia, and osteoporosis categories. We then conducted a controlled trial with five orthopedic clinicians to evaluate the clinical value of the model. The average clinician's accuracy improved from 61.50 ± 10.79 % unaided to 80.00 ± 5.92 % aided (18.50 % improvement), sensitivity improved from 64.38 ± 8.07 % unaided to 83.31 ± 5.43 % aided (18.93 % improvement), and specificity improved from 80.11 ± 4.72 % unaided to 89.94 ± 3.82 % aided (9.83 % improvement). Meanwhile, the prediction consistency among clinicians significantly improved with the assistance of FCoTNet. Furthermore, the proposed model showed good robustness on an external test dataset. These investigations indicate that the proposed deep learning model achieves state-of-the-art performance for osteoporosis prediction, which substantially improves osteoporosis screening and reduced osteoporosis fractures.

Recent advances in causal inference techniques, more specifically, in the theory of structural causal models, provide the framework for identifying causal effects from observational data in cases where the causal graph is identifiable, i.e., the data generation mechanism can be recovered from the joint distribution. However, no such studies have been performed to demonstrate this concept with a clinical example. We present a complete framework to estimate the causal effects from observational data by augmenting expert knowledge in the model development phase and with a practical clinical application. Our clinical application entails a timely and essential research question, the effect of oxygen therapy intervention in the intensive care unit (ICU). The result of this project is helpful in a variety of disease conditions, including severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) patients in the ICU. We used data from the MIMIC-III database, a widely used health care database in the machine learning community with 58,976 admissions from an ICU in Boston, MA, to estimate the oxygen therapy effect on morality. We also identified the model's covariate-specific effect on oxygen therapy for more personalized intervention.

Malignant Mesothelioma is a difficult to diagnose and highly lethal cancer usually associated with asbestos exposure. It can be broadly classified into three subtypes: Epithelioid, Sarcomatoid, and a hybrid Biphasic subtype in which significant components of both of the previous subtypes are present. Early diagnosis and identification of the subtype informs treatment and can help improve patient outcome. However, the subtyping of malignant mesothelioma, and specifically the recognition of transitional features from routine histology slides has a high level of inter-observer variability. In this work, we propose an end-to-end multiple instance learning (MIL) approach for malignant mesothelioma subtyping. This uses an adaptive instance-based sampling scheme for training deep convolutional neural networks on bags of image patches that allows learning on a wider range of relevant instances compared to max or top-N based MIL approaches. We also investigate augmenting the instance representation to include aggregate cellular morphology features from cell segmentation. The proposed MIL approach enables identification of malignant mesothelial subtypes of specific tissue regions. From this a continuous characterisation of a sample according to predominance of sarcomatoid vs epithelioid regions is possible, thus avoiding the arbitrary and highly subjective categorisation by currently used subtypes. Instance scoring also enables studying tumor heterogeneity and identifying patterns associated with different subtypes. We have evaluated the proposed method on a dataset of 234 tissue micro-array cores with an AUROC of 0.89±0.05 for this task. The dataset and developed methodology is available for the community at: https://github.com/measty/PINS.

Certain life-threatening abnormalities, such as cholangiocarcinoma, in the human biliary tract are curable if detected at an early stage, and ultrasonography has been proven to be an effective tool for identifying them. However, the diagnosis often requires a second opinion from experienced radiologists, who are usually overwhelmed by many cases. Therefore, we propose a deep convolutional neural network model, named biliary tract network (BiTNet), developed to solve problems in the current screening system and to avoid overconfidence issues of traditional deep convolutional neural networks. Additionally, we present an ultrasound image dataset for the human biliary tract and demonstrate two artificial intelligence (AI) applications: auto-prescreening and assisting tools. The proposed model is the first AI model to automatically screen and diagnose upper-abdominal abnormalities from ultrasound images in real-world healthcare scenarios. Our experiments suggest that prediction probability has an impact on both applications, and our modifications to EfficientNet solve the overconfidence problem, thereby improving the performance of both applications and of healthcare professionals. The proposed BiTNet can reduce the workload of radiologists by 35% while keeping the false negatives to as low as 1 out of every 455 images. Our experiments involving 11 healthcare professionals with four different levels of experience reveal that BiTNet improves the diagnostic performance of participants of all levels. The mean accuracy and precision of the participants with BiTNet as an assisting tool (0.74 and 0.61, respectively) are statistically higher than those of participants without the assisting tool (0.50 and 0.46, respectively (p<0.001)). These experimental results demonstrate the high potential of BiTNet for use in clinical settings.

Low-dose CT techniques attempt to minimize the radiation exposure of patients by estimating the high-resolution normal-dose CT images to reduce the risk of radiation-induced cancer. In recent years, many deep learning methods have been proposed to solve this problem by building a mapping function between low-dose CT images and their high-dose counterparts. However, most of these methods ignore the effect of different radiation doses on the final CT images, which results in large differences in the intensity of the noise observable in CT images. What'more, the noise intensity of low-dose CT images exists significantly differences under different medical devices manufacturers. In this paper, we propose a multi-level noise-aware network (MLNAN) implemented with constrained cycle Wasserstein generative adversarial networks to recovery the low-dose CT images under uncertain noise levels. Particularly, the noise-level classification is predicted and reused as a prior pattern in generator networks. Moreover, the discriminator network introduces noise-level determination. Under two dose-reduction strategies, experiments to evaluate the performance of proposed method are conducted on two datasets, including the simulated clinical AAPM challenge datasets and commercial CT datasets from United Imaging Healthcare (UIH). The experimental results illustrate the effectiveness of our proposed method in terms of noise suppression and structural detail preservation compared with several other deep-learning based methods. Ablation studies validate the effectiveness of the individual components regarding the afforded performance improvement. Further research for practical clinical applications and other medical modalities is required in future works.

Over the past decade, machine learning (ML) and artificial intelligence (AI) have become increasingly prevalent in the medical field. In the United States, the Food and Drug Administration (FDA) is responsible for regulating AI algorithms as "medical devices" to ensure patient safety. However, recent work has shown that the FDA approval process may be deficient. In this study, we evaluate the evidence supporting FDA-approved neuroalgorithms, the subset of machine learning algorithms with applications in the central nervous system (CNS), through a systematic review of the primary literature. Articles covering the 53 FDA-approved algorithms with applications in the CNS published in PubMed, EMBASE, Google Scholar and Scopus between database inception and January 25, 2022 were queried. Initial searches identified 1505 studies, of which 92 articles met the criteria for extraction and inclusion. Studies were identified for 26 of the 53 neuroalgorithms, of which 10 algorithms had only a single peer-reviewed publication. Performance metrics were available for 15 algorithms, external validation studies were available for 24 algorithms, and studies exploring the use of algorithms in clinical practice were available for 7 algorithms. Papers studying the clinical utility of these algorithms focused on three domains: workflow efficiency, cost savings, and clinical outcomes. Our analysis suggests that there is a meaningful gap between the FDA approval of machine learning algorithms and their clinical utilization. There appears to be room for process improvement by implementation of the following recommendations: the provision of compelling evidence that algorithms perform as intended, mandating minimum sample sizes, reporting of a predefined set of performance metrics for all algorithms and clinical application of algorithms prior to widespread use. This work will serve as a baseline for future research into the ideal regulatory framework for AI applications worldwide.

Significant difficulties in medical image segmentation include the high variability of images caused by their origin (multi-center), the acquisition protocols (multi-parametric), the variability of human anatomy, illness severity, the effect of age and gender, and notable other factors. This work addresses problems associated with the automatic semantic segmentation of lumbar spine magnetic resonance images using convolutional neural networks. We aimed to assign a class label to each pixel of an image, with classes defined by radiologists corresponding to structural elements such as vertebrae, intervertebral discs, nerves, blood vessels, and other tissues. The proposed network topologies represent variants of the U-Net architecture, and we used several complementary blocks to define the variants: three types of convolutional blocks, spatial attention models, deep supervision, and multilevel feature extractor. Here, we describe the topologies and analyze the results of the neural network designs that obtained the most accurate segmentation. Several proposed designs outperform the standard U-Net used as a baseline, primarily when used in ensembles, where the outputs of multiple neural networks are combined according to different strategies.

Decision tree (DT) models provide a transparent approach to prediction of patient's outcomes within a probabilistic framework. Averaging over DT models under certain conditions can deliver reliable estimates of predictive posterior probability distributions, which is of critical importance in the case of predicting an individual patient's outcome. Reliable estimations of the distribution can be achieved within the Bayesian framework using Markov chain Monte Carlo (MCMC) and its Reversible Jump extension enabling DT models to grow to a reasonable size. Existing MCMC strategies however have limited ability to control DT structures and tend to sample overgrown DT models, making unreasonably small partitions, thus deteriorating the uncertainty calibration. This happens because the MCMC explores a DT model parameter space within a limited knowledge of the distribution of data partitions. We propose a new adaptive strategy which overcomes this limitation, and show that in the case of predicting trauma outcomes the number of data partitions can be significantly reduced, so that the unnecessary uncertainty of estimating the predictive posterior density is avoided. The proposed and existing strategies are compared in terms of entropy which, being calculated for predicted posterior distributions, represents the uncertainty in decisions. In this framework, the proposed method has outperformed the existing sampling strategies, so that the unnecessary uncertainty in decisions is efficiently avoided.

BACKGROUND OF THE STUDY: Breast cancer is the most fatal disease that widely affects women. When the cancerous lumps grow from the cells of the breast, it causes breast cancer. Self-analysis and regular medical check-ups help for detecting the disease earlier and enhance the survival rate. Hence, an automated breast cancer detection system in mammograms can assist clinicians in the patient's treatment. In medical techniques, the categorization of breast cancer becomes challenging for investigators and researchers. The advancement in deep learning approaches has established more attention to their advantages to medical imaging issues, especially for breast cancer detection. AIM: The research work plans to develop a novel hybrid model for breast cancer diagnosis with the support of optimized deep-learning architecture. METHODS: The required images are gathered from the benchmark datasets. These collected datasets are used in three pre-processing approaches like "Median Filtering, Histogram Equalization, and morphological operation", which helps to remove unwanted regions from the images. Then, the pre-processed images are applied to the Optimized U-net-based tumor segmentation phase for obtaining accurate segmented results along with the optimization of certain parameters in U-Net by employing "Adapted-Black Widow Optimization (A-BWO)". Further, the detection is performed in two different ways that is given as model 1 and model 2. In model 1, the segmented tumors are used to extract the significant patterns with the help of the "Gray-Level Co-occurrence Matrix (GLCM) and Local Gradient pattern (LGP)". Further, these extracted patterns are utilized in the "Dual Model accessed Optimized Long Short-Term Memory (DM-OLSTM)" for performing breast cancer detection and the detected score 1 is obtained. In model 2, the same segmented tumors are given into the different variants of CNN, such as "VGG19, Resnet150, and Inception". The extracted deep features from three CNN-based approaches are fused to form a single set of deep features. These fused deep features are inserted into the developed DM-OLSTM for getting the detected score 2 for breast cancer diagnosis. In the final phase of the hybrid model, the score 1 and score 2 obtained from model 1 and model 2 are averaged to get the final detection output. RESULTS: The accuracy and F1-score of the offered DM-OLSTM model are achieved at 96 % and 95 %. CONCLUSION: Experimental analysis proves that the recommended methodology achieves better performance by analyzing with the benchmark dataset. Hence, the designed model is helpful for detecting breast cancer in real-time applications.

OBJECTIVE: The pathophysiological concepts of diseases are encapsulated in patients' medical histories. Whether information on the pathophysiology or anatomy of "infarction" can be preserved and objectively expressed in the distributed representation obtained from a corpus of scientific Japanese medical texts in the "infarction" domain is currently unknown. Word2Vec was used to obtain distributed representations, meanings, and word analogies of word vectors, and this process was verified mathematically. MATERIALS & METHODS: The texts were abstracts that were obtained by searching for "infarction," "abstract," and "case report" in the Japan Medical Journal Association's Ichushi Data Base. The abstracted text was morphologically analyzed to produce word sequences converted into their standard form. MeCab was used for morphological analysis and mecab-ipadic-NEologd and ComeJisyo were used as dictionaries. The accuracy of the known tasks for medical terms was evaluated using a word analogy task specific to the "infarction" domain. RESULTS: Only 33 % of the word analogy tasks for medical terminology were correct. However, 52 % of the new original tasks, which were specific to the "infarction" domain, were correct, especially those regarding anatomical differences. DISCUSSION: Documents related to "infarction" were collected from a corpus of Japanese medical documents and word-embedded expressions were obtained using Word2Vec. Terminology that had similar meanings to "infarction" included words such as "cavity" and "ischemia," which suggest the pathology of an infarction. CONCLUSION: The pathophysiological and anatomical features of an "infarction" may be retained in a distributed representation.

Traditional Chinese medicine (TCM) has gradually played an indispensable role in people's health maintenance, especially in the treatment of chronic diseases. However, there is always uncertainty and hesitation in the judgment and understanding of diseases by doctors, which affects the status recognition and optimal diagnosis and treatment decision-making of patients. In order to overcome the above problems, we lead into probabilistic double hierarchy linguistic term set (PDHLTS) to accurately describe language information in traditional Chinese medicine and make decisions. In this paper, a multi-criteria group decision making (MCGDM) model is constructed based on the MSM-MCBAC (Maclaurin symmetric mean-MultiCriteria Border Approximation area Comparison) method in the PDHL environment. Firstly, a PDHL weighted Maclaurin symmetric mean (PDHLWMSM) operator is proposed to aggregate the evaluation matrices of multiple experts. Then, combined with the BWM and maximizing deviation method, a comprehensive weight determination method is put forward to calculate the weights of criteria. Furthermore, we propose PDHL MSM-MCBAC method based on the Multi-Attributive Border Approximation area Comparison (MABAC) method and the PDHLWMSM operator. Finally, an example of a selection of TCM prescriptions is used and some comparative analyses are made to verify the effectiveness and superiority of this paper.

The paradigm of evidence-based medicine requires that medical decisions are made on the basis of the best available knowledge published in the literature. Existing evidence is often summarized in the form of systematic reviews and/or meta-reviews and is rarely available in a structured form. Manual compilation and aggregation is costly, and conducting a systematic review represents a high effort. The need to aggregate evidence arises not only in the context of clinical trials, but is also important in the context of pre-clinical animal studies. In this context, evidence extraction is important to support translation of the most promising pre-clinical therapies into clinical trials or to optimize clinical trial design. Aiming at developing methods that facilitate the task of aggregating evidence published in pre-clinical studies, in this paper a new system is presented that automatically extracts structured knowledge from such publications and stores it in a so-called domain knowledge graph. The approach follows the paradigm of model-complete text comprehension by relying on guidance from a domain ontology creating a deep relational data-structure that reflects the main concepts, protocol, and key findings of studies. Focusing on the domain of spinal cord injuries, a single outcome of a pre-clinical study is described by up to 103 outcome parameters. Since the problem of extracting all these variables together is intractable, we propose a hierarchical architecture that incrementally predicts semantic sub-structures according to a given data model in a bottom-up fashion. At the heart of our approach is a statistical inference method that relies on conditional random fields to infer the most likely instance of the domain model given the text of a scientific publication as input. This approach allows modeling dependencies between the different variables describing a study in a semi-joint fashion. We present a comprehensive evaluation of our system to understand the extent to which our system can capture a study in the depth required to enable the generation of new knowledge. We conclude the article with a brief description of some applications of the populated knowledge graph and show the potential implications of our work for supporting evidence-based medicine.

BACKGROUND: Medical use cases for machine learning (ML) are growing exponentially. The first hospitals are already using ML systems as decision support systems in their daily routine. At the same time, most ML systems are still opaque and it is not clear how these systems arrive at their predictions. METHODS: In this paper, we provide a brief overview of the taxonomy of explainability methods and review popular methods. In addition, we conduct a systematic literature search on PubMed to investigate which explainable artificial intelligence (XAI) methods are used in 450 specific medical supervised ML use cases, how the use of XAI methods has emerged recently, and how the precision of describing ML pipelines has evolved over the past 20 years. RESULTS: A large fraction of publications with ML use cases do not use XAI methods at all to explain ML predictions. However, when XAI methods are used, open-source and model-agnostic explanation methods are more commonly used, with SHapley Additive exPlanations (SHAP) and Gradient Class Activation Mapping (Grad-CAM) for tabular and image data leading the way. ML pipelines have been described in increasing detail and uniformity in recent years. However, the willingness to share data and code has stagnated at about one-quarter. CONCLUSIONS: XAI methods are mainly used when their application requires little effort. The homogenization of reports in ML use cases facilitates the comparability of work and should be advanced in the coming years. Experts who can mediate between the worlds of informatics and medicine will become more and more in demand when using ML systems due to the high complexity of the domain.

Digital mammography is currently the most common imaging tool for breast cancer screening. Although the benefits of using digital mammography for cancer screening outweigh the risks associated with the x-ray exposure, the radiation dose must be kept as low as possible while maintaining the diagnostic utility of the generated images, thus minimizing patient risks. Many studies investigated the feasibility of dose reduction by restoring low-dose images using deep neural networks. In these cases, choosing the appropriate training database and loss function is crucial and impacts the quality of the results. In this work, we used a standard residual network (ResNet) to restore low-dose digital mammography images and evaluated the performance of several loss functions. For training purposes, we extracted 256,000 image patches from a dataset of 400 images of retrospective clinical mammography exams, where dose reduction factors of 75% and 50% were simulated to generate low and standard-dose pairs. We validated the network in a real scenario by using a physical anthropomorphic breast phantom to acquire real low-dose and standard full-dose images in a commercially available mammography system, which were then processed through our trained model. We benchmarked our results against an analytical restoration model for low-dose digital mammography. Objective assessment was performed through the signal-to-noise ratio (SNR) and the mean normalized squared error (MNSE), decomposed into residual noise and bias. Statistical tests revealed that the use of the perceptual loss (PL4) resulted in statistically significant differences when compared to all other loss functions. Additionally, images restored using the PL4 achieved the closest residual noise to the standard dose. On the other hand, perceptual loss PL3, structural similarity index (SSIM) and one of the adversarial losses achieved the lowest bias for both dose reduction factors. The source code of our deep neural network is available at https://github.com/WANG-AXIS/LdDMDenoising.

Adoption of artificial intelligence (AI) by the medical community has long been anticipated, endorsed by a stream of machine learning literature showcasing AI systems that yield extraordinary performance. However, many of these systems are likely over-promising and will under-deliver in practice. One key reason is the community's failure to acknowledge and address the presence of inflationary effects in the data. These simultaneously inflate evaluation performance and prevent a model from learning the underlying task, thus severely misrepresenting how that model would perform in the real world. This paper investigated the impact of these inflationary effects on healthcare tasks, as well as how these effects can be addressed. Specifically, we defined three inflationary effects that occur in medical data sets and allow models to easily reach small training losses and prevent skillful learning. We investigated two data sets of sustained vowel phonation from participants with and without Parkinson's disease, and revealed that published models which have achieved high classification performances on these were artificially enhanced due to the inflationary effects. Our experiments showed that removing each inflationary effect corresponded with a decrease in classification accuracy, and that removing all inflationary effects reduced the evaluated performance by up to 30%. Additionally, the performance on a more realistic test set increased, suggesting that the removal of these inflationary effects enabled the model to better learn the underlying task and generalize. Source code is available at https://github.com/Wenbo-G/pd-phonation-analysis under the MIT license.

BACKGROUND: Deep learning has been successfully applied to ECG data to aid in the accurate and more rapid diagnosis of acutely decompensated heart failure (ADHF). Previous applications focused primarily on classifying known ECG patterns in well-controlled clinical settings. However, this approach does not fully capitalize on the potential of deep learning, which directly learns important features without relying on a priori knowledge. In addition, deep learning applications to ECG data obtained from wearable devices have not been well studied, especially in the field of ADHF prediction. METHODS: We used ECG and transthoracic bioimpedance data from the SENTINEL-HF study, which enrolled patients (≥21 years) who were hospitalized with a primary diagnosis of heart failure or with ADHF symptoms. To build an ECG-based prediction model of ADHF, we developed a deep cross-modal feature learning pipeline, termed ECGX-Net, that utilizes raw ECG time series and transthoracic bioimpedance data from wearable devices. To extract rich features from ECG time series data, we first adopted a transfer learning approach in which ECG time series were transformed into 2D images, followed by feature extraction using ImageNet-pretrained DenseNet121/VGG19 models. After data filtering, we applied cross-modal feature learning in which a regressor was trained with ECG and transthoracic bioimpedance. Then, we concatenated the DenseNet121/VGG19 features with the regression features and used them to train a support vector machine (SVM) without bioimpedance information. RESULTS: The high-precision classifier using ECGX-Net predicted ADHF with a precision of 94 %, a recall of 79 %, and an F1-score of 0.85. The high-recall classifier with only DenseNet121 had a precision of 80 %, a recall of 98 %, and an F1-score of 0.88. We found that ECGX-Net was effective for high-precision classification, while DenseNet121 was effective for high-recall classification. CONCLUSION: We show the potential for predicting ADHF from single-channel ECG recordings obtained from outpatients, enabling timely warning signs of heart failure. Our cross-modal feature learning pipeline is expected to improve ECG-based heart failure prediction by handling the unique requirements of medical scenarios and resource limitations.

Alzheimer's disease and related dementias (ADRD) present a looming public health crisis, affecting roughly 5 million people and 11 % of older adults in the United States. Despite nationwide efforts for timely diagnosis of patients with ADRD, >50 % of them are not diagnosed and unaware of their disease. To address this challenge, we developed ADscreen, an innovative speech-processing based ADRD screening algorithm for the protective identification of patients with ADRD. ADscreen consists of five major components: (i) noise reduction for reducing background noises from the audio-recorded patient speech, (ii) modeling the patient's ability in phonetic motor planning using acoustic parameters of the patient's voice, (iii) modeling the patient's ability in semantic and syntactic levels of language organization using linguistic parameters of the patient speech, (iv) extracting vocal and semantic psycholinguistic cues from the patient speech, and (v) building and evaluating the screening algorithm. To identify important speech parameters (features) associated with ADRD, we used the Joint Mutual Information Maximization (JMIM), an effective feature selection method for high dimensional, small sample size datasets. Modeling the relationship between speech parameters and the outcome variable (presence/absence of ADRD) was conducted using three different machine learning (ML) architectures with the capability of joining informative acoustic and linguistic with contextual word embedding vectors obtained from the DistilBERT (Bidirectional Encoder Representations from Transformers). We evaluated the performance of the ADscreen on an audio-recorded patients' speech (verbal description) for the Cookie-Theft picture description task, which is publicly available in the dementia databank. The joint fusion of acoustic and linguistic parameters with contextual word embedding vectors of DistilBERT achieved F1-score = 84.64 (standard deviation [std] = ±3.58) and AUC-ROC = 92.53 (std = ±3.34) for training dataset, and F1-score = 89.55 and AUC-ROC = 93.89 for the test dataset. In summary, ADscreen has a strong potential to be integrated with clinical workflow to address the need for an ADRD screening tool so that patients with cognitive impairment can receive appropriate and timely care.

Coronavirus (COVID-19) is a newly discovered viral disease from the SARS-CoV-2 family. This has caused a moral panic resulting in the spread of informative and uninformative information about COVID-19 and its effects. Twitter is a popular social media platform used extensively during the current outbreak. This paper aims to predict informative tweets related to COVID-19 on Twitter using a novel set of fuzzy rules involving deep learning techniques. This study focuses on identifying informative tweets during the pandemic to provide the public with trustworthy information and forecast how quickly diseases could spread. In this case, we have implemented RoBERTa and CT-BERT models using the fuzzy methodology to identify COVID-19 patient tweets. The proposed architecture combines deep learning transformer models RoBERTa and CT-BERT with the fuzzy technique to categorize posts as INFORMATIVE or UNINFORMATIVE. We performed a comparative analysis of our method with machine learning models and deep learning approaches. The results show that our proposed model can classify informative and uninformative tweets with an accuracy of 91.40% and an F1-score of 91.94% using the COVID-19 English tweet dataset. The proposed model is accurate and ready for real-world application.

With an estimated five million fatal cases each year, lung cancer is one of the significant causes of death worldwide. Lung diseases can be diagnosed with a Computed Tomography (CT) scan. The scarcity and trustworthiness of human eyes is the fundamental issue in diagnosing lung cancer patients. The main goal of this study is to detect malignant lung nodules in a CT scan of the lungs and categorize lung cancer according to severity. In this work, cutting-edge Deep Learning (DL) algorithms were used to detect the location of cancerous nodules. Also, the real-life issue is sharing data with hospitals around the world while bearing in mind the organizations' privacy issues. Besides, the main problems for training a global DL model are creating a collaborative model and maintaining privacy. This study presented an approach that takes a modest amount of data from multiple hospitals and uses blockchain-based Federated Learning (FL) to train a global DL model. The data were authenticated using blockchain technology, and FL trained the model internationally while maintaining the organization's anonymity. First, we presented a data normalization approach that addresses the variability of data obtained from various institutions using various CT scanners. Furthermore, using a CapsNets method, we classified lung cancer patients in local mode. Finally, we devised a way to train a global model cooperatively utilizing blockchain technology and FL while maintaining anonymity. We also gathered data from real-life lung cancer patients for testing purposes. The suggested method was trained and tested on the Cancer Imaging Archive (CIA) dataset, Kaggle Data Science Bowl (KDSB), LUNA 16, and the local dataset. Finally, we performed extensive experiments with Python and its well-known libraries, such as Scikit-Learn and TensorFlow, to evaluate the suggested method. The findings showed that the method effectively detects lung cancer patients. The technique delivered 99.69 % accuracy with the smallest possible categorization error.

Attention Deficit/Hyperactivity Disorder (ADHD) is a prevalent neurodevelopmental disorder in childhood that often persists into adulthood. Objectively diagnosing ADHD can be challenging due to the reliance on subjective questionnaires in clinical assessment. Fortunately, recent advancements in artificial intelligence (AI) have shown promise in providing objective diagnoses through the analysis of medical images or activity recordings. These AI-based techniques have demonstrated accurate ADHD diagnosis; however, the growing complexity of deep learning models has introduced a lack of interpretability. These models often function as black boxes, unable to offer meaningful insights into the data patterns that characterize ADHD. OBJECTIVE: This paper proposes a methodology to interpret the output of an AI-based diagnosis system for combined ADHD in age and gender-stratified populations. METHODS: Our system is based on the analysis of 24 hour-long activity records using Convolutional Neural Networks (CNNs) to classify spectrograms of activity windows. These windows are interpreted using occlusion maps to highlight the time-frequency patterns explaining ADHD activity. RESULTS: Significant differences in the frequency patterns between ADHD and controls both in diurnal and nocturnal activity were found for all the populations. Temporal dispersion also presented differences in the male population. CONCLUSION: The proposed interpretation techniques for CNNs highlighted gender- and age-related differences between ADHD patients and controls. Leveraging these differences could potentially lead to improved diagnostic accuracy, especially if a larger and more balanced dataset is utilized. SIGNIFICANCE: Our findings pave the way for the development of an AI-based diagnosis system for ADHD that offers interpretability, thereby providing valuable insights into the underlying etiology of the disease.

Automatic segmentation of the cardiac left ventricle with scars remains a challenging and clinically significant task, as it is essential for patient diagnosis and treatment pathways. This study aimed to develop a novel framework and cost function to achieve optimal automatic segmentation of the left ventricle with scars using LGE-MRI images. To ensure the generalization of the framework, an unbiased validation protocol was established using out-of-distribution (OOD) internal and external validation cohorts, and intra-observation and inter-observer variability ground truths. The framework employs a combination of traditional computer vision techniques and deep learning, to achieve optimal segmentation results. The traditional approach uses multi-atlas techniques, active contours, and k-means methods, while the deep learning approach utilizes various deep learning techniques and networks. The study found that the traditional computer vision technique delivered more accurate results than deep learning, except in cases where there was breath misalignment error. The optimal solution of the framework achieved robust and generalized results with Dice scores of 82.8 ± 6.4% and 72.1 ± 4.6% in the internal and external OOD cohorts, respectively. The developed framework offers a high-performance solution for automatic segmentation of the left ventricle with scars using LGE-MRI. Unlike existing state-of-the-art approaches, it achieves unbiased results across different hospitals and vendors without the need for training or tuning in hospital cohorts. This framework offers a valuable tool for experts to accomplish the task of fully automatic segmentation of the left ventricle with scars based on a single-modality cardiac scan.

Several recent studies indicate that atypical changes in driving behaviors appear to be early signs of mild cognitive impairment (MCI) and dementia. These studies, however, are limited by small sample sizes and short follow-up duration. This study aims to develop an interaction-based classification method building on a statistic named Influence Score (i.e., I-score) for prediction of MCI and dementia using naturalistic driving data collected from the Longitudinal Research on Aging Drivers (LongROAD) project. Naturalistic driving trajectories were collected through in-vehicle recording devices for up to 44 months from 2977 participants who were cognitively intact at the time of enrollment. These data were further processed and aggregated to generate 31 time-series driving variables. Because of high dimensional time-series features for driving variables, we used I-score for variable selection. I-score is a measure to evaluate variables' ability to predict and is proven to be effective in differentiating between noisy and predictive variables in big data. It is introduced here to select influential variable modules or groups that account for compound interactions among explanatory variables. It is explainable regarding to what extent variables and their interactions contribute to the predictiveness of a classifier. In addition, I-score boosts the performance of classifiers over imbalanced datasets due to its association with the F1 score. Using predictive variables selected by I-score, interaction-based residual blocks are constructed over top I-score modules to generate predictors and ensemble learning aggregates these predictors to boost the prediction of the overall classifier. Experiments using naturalistic driving data show that our proposed classification method achieves the best accuracy (96%) for predicting MCI and dementia, followed by random forest (93%) and logistic regression (88%). In terms of F1 score and AUC, our proposed classifier achieves 98% and 87%, respectively, followed by random forest (with an F1 score of 96% and an AUC of 79%) and logistic regression (with an F1 score of 92% and an AUC of 77%). The results indicate that incorporating I-score into machine learning algorithms could considerably improve the model performance for predicting MCI and dementia in older drivers. We also performed the feature importance analysis and found that the right to left turn ratio and the number of hard braking events are the most important driving variables to predict MCI and dementia.

PURPOSE: Automated diagnosis and prognosis of Alzheimer's Disease remain a challenging problem that machine learning (ML) techniques have attempted to resolve in the last decade. This study introduces a first-of-its-kind color-coded visualization mechanism driven by an integrated ML model to predict disease trajectory in a 2-year longitudinal study. The main aim of this study is to help capture visually in 2D and 3D renderings the diagnosis and prognosis of AD, therefore augmenting our understanding of the processes of multiclass classification and regression analysis. METHOD: The proposed method, Machine Learning for Visualizing AD (ML4VisAD), is designed to predict disease progression through a visual output. This newly developed model takes baseline measurements as input to generate a color-coded visual image that reflects disease progression at different time points. The architecture of the network relies on convolutional neural networks. With 1123 subjects selected from the ADNI QT-PAD dataset, we use a 10-fold cross-validation process to evaluate the method. Multimodal inputs* include neuroimaging data (MRI, PET), neuropsychological test scores (excluding MMSE, CDR-SB, and ADAS to avoid bias), cerebrospinal fluid (CSF) biomarkers with measures of amyloid beta (ABETA), phosphorylated tau protein (PTAU), total tau protein (TAU), and risk factors that include age, gender, years of education, and ApoE4 gene. FINDINGS/RESULTS: Based on subjective scores reached by three raters, the results showed an accuracy of 0.82 ± 0.03 for a 3-way classification and 0.68 ± 0.05 for a 5-way classification. The visual renderings were generated in 0.08 msec for a 23 × 23 output image and in 0.17 ms for a 45 × 45 output image. Through visualization, this study (1) demonstrates that the ML visual output augments the prospects for a more accurate diagnosis and (2) highlights why multiclass classification and regression analysis are incredibly challenging. An online survey was conducted to gauge this visualization platform's merits and obtain valuable feedback from users. All implementation codes are shared online on GitHub. CONCLUSION: This approach makes it possible to visualize the many nuances that lead to a specific classification or prediction in the disease trajectory, all in context to multimodal measurements taken at baseline. This ML model can serve as a multiclass classification and prediction model while reinforcing the diagnosis and prognosis capabilities by including a visualization platform.

Machine learning (ML) for diagnosis of thyroid nodules on ultrasound is an active area of research. However, ML tools require large, well-labeled datasets, the curation of which is time-consuming and labor-intensive. The purpose of our study was to develop and test a deep-learning-based tool to facilitate and automate the data annotation process for thyroid nodules; we named our tool Multistep Automated Data Labelling Procedure (MADLaP). MADLaP was designed to take multiple inputs including pathology reports, ultrasound images, and radiology reports. Using multiple step-wise 'modules' including rule-based natural language processing, deep-learning-based imaging segmentation, and optical character recognition, MADLaP automatically identified images of a specific thyroid nodule and correctly assigned a pathology label. The model was developed using a training set of 378 patients across our health system and tested on a separate set of 93 patients. Ground truths for both sets were selected by an experienced radiologist. Performance metrics including yield (how many labeled images the model produced) and accuracy (percentage correct) were measured using the test set. MADLaP achieved a yield of 63 % and an accuracy of 83 %. The yield progressively increased as the input data moved through each module, while accuracy peaked part way through. Error analysis showed that inputs from certain examination sites had lower accuracy (40 %) than the other sites (90 %, 100 %). MADLaP successfully created curated datasets of labeled ultrasound images of thyroid nodules. While accurate, the relatively suboptimal yield of MADLaP exposed some challenges when trying to automatically label radiology images from heterogeneous sources. The complex task of image curation and annotation could be automated, allowing for enrichment of larger datasets for use in machine learning development.

Difference of Gaussians (DoG) convolutional filters are one of the earliest image processing methods employed for detecting microcalcifications on mammogram images before machine and deep learning methods became widespread. DoG is a blob enhancement filter that consists in subtracting one Gaussian-smoothed version of an image from another less Gaussian-smoothed version of the same image. Smoothing with a Gaussian kernel suppresses high-frequency spatial information, thus DoG can be regarded as a band-pass filter. However, due to their small size and overimposed breast tissue, microcalcifications vary greatly in contrast-to-noise ratio and sharpness. This makes it difficult to find a single DoG configuration that enhances all microcalcifications. In this work, we propose a convolutional network, named DoG-MCNet, where the first layer automatically learns a bank of DoG filters parameterized by their associated standard deviations. We experimentally show that when employed for microcalcification detection, our DoG layer acts as a learnable bank of band-pass preprocessing filters and improves detection performance by 4.86% AUFROC over baseline MCNet and 1.53% AUFROC over state-of-the-art multicontext ensemble of CNNs.

BACKGROUND/INTRODUCTION: Manual detection and localization of the brain's epileptogenic areas using electroencephalogram (EEG) signals is time-intensive and error-prone. An automated detection system is, thus, highly desirable for support in clinical diagnosis. A set of relevant and significant non-linear features plays a major role in developing a reliable, automated focal detection system. METHODS: A new feature extraction method is designed to classify focal EEG signals using eleven non-linear geometrical attributes derived from the Fourier-Bessel series expansion-based empirical wavelet transform (FBSE-EWT) segmented rhythm's second-order difference plot (SODP). A total of 132 features (2 channels × 6 rhythms × 11 geometrical attributes) were computed. However, some of the obtained features might be non-significant and redundant features. Hence, to acquire an optimal set of relevant non-linear features, a new hybridization of 'Kruskal-Wallis statistical test (KWS)' with 'VlseKriterijuska Optimizacija I Komoromisno Resenje' termed as the KWS-VIKOR approach was adopted. The KWS-VIKOR has a two-fold operational feature. First, the significant features are selected using the KWS test with a p-value lesser than 0.05. Next, the multi-attribute decision-making (MADM) based VIKOR method ranks the selected features. Several classification methods further validate the efficacy of the features of the selected top n%. RESULTS: The proposed framework has been evaluated using the Bern-Barcelona dataset. The highest classification accuracy of 98.7% was achieved using the top 35% ranked features in classifying the focal and non-focal EEG signals with the least-squares support vector machine (LS-SVM) classifier. CONCLUSIONS: The achieved results exceeded those reported through other methods. Hence, the proposed framework will more effectively assist the clinician in localizing the epileptogenic areas.

The overexpression of the human epidermal growth factor receptor 2 (HER2) is a predictive biomarker in therapeutic effects for metastatic breast cancer. Accurate HER2 testing is critical for determining the most suitable treatment for patients. Fluorescent in situ hybridization (FISH) and dual in situ hybridization (DISH) have been recognized as FDA-approved methods to determine HER2 overexpression. However, analysis of HER2 overexpression is challenging. Firstly, the boundaries of cells are often unclear and blurry, with large variations in cell shapes and signals, making it challenging to identify the precise areas of HER2-related cells. Secondly, the use of sparsely labeled data, where some unlabeled HER2-related cells are classified as background, can significantly confuse fully supervised AI learning and result in unsatisfactory model outcomes. In this study, we present a weakly supervised Cascade R-CNN (W-CRCNN) model to automatically detect HER2 overexpression in HER2 DISH and FISH images acquired from clinical breast cancer samples. The experimental results demonstrate that the proposed W-CRCNN achieves excellent results in identification of HER2 amplification in three datasets, including two DISH datasets and a FISH dataset. For the FISH dataset, the proposed W-CRCNN achieves an accuracy of 0.970±0.022, precision of 0.974±0.028, recall of 0.917±0.065, F1-score of 0.943±0.042 and Jaccard Index of 0.899±0.073. For DISH datasets, the proposed W-CRCNN achieves an accuracy of 0.971±0.024, precision of 0.969±0.015, recall of 0.925±0.020, F1-score of 0.947±0.036 and Jaccard Index of 0.884±0.103 for dataset 1, and an accuracy of 0.978±0.011, precision of 0.975±0.011, recall of 0.918±0.038, F1-score of 0.946±0.030 and Jaccard Index of 0.884±0.052 for dataset 2, respectively. In comparison with the benchmark methods, the proposed W-CRCNN significantly outperforms all the benchmark approaches in identification of HER2 overexpression in FISH and DISH datasets (p<0.05). With the high degree of accuracy, precision and recall , the results show that the proposed method in DISH analysis for assessment of HER2 overexpression in breast cancer patients has significant potential to assist precision medicine.

We present a method for training neural networks with synthetic electrocardiograms that mimic signals produced by a wearable single lead electrocardiogram monitor. We use domain randomization where the synthetic signal properties such as the waveform shape, RR-intervals and noise are varied for every training example. Models trained with synthetic data are compared to their counterparts trained with real data. Detection of r-waves in electrocardiograms recorded during different physical activities and in atrial fibrillation is used to assess the performance. By allowing the randomization of the synthetic signals to increase beyond what is typically observed in the real-world data the performance is on par or superseding the performance of networks trained with real data. Experiments show robust model performance using different seeds and on different unseen test sets that were fully separated from the training phase. The ability of the model to generalize well to hidden test sets without any specific tuning provides a simple and explainable alternative to more complex adversarial domain adaptation methods for model generalization. This method opens up the possibility of extending the use of synthetic data towards domain insensitive cardiac disease classification when disease specific a priori information is used in the electrocardiogram generation. Additionally, the method provides training with free-to-collect data with accurate labels, control of the data distribution eliminating class imbalances that are typically observed in health-related data, and the generated data is inherently private.

This article proposes a virtual reality (VR) system for diagnosing and rehabilitating lower limb amputees. A virtual environment and an intelligent space are the basis of the proposed solution. The target audiences are physiotherapists and doctors, and the aim is to provide a VR-based system to allow visualization and analysis of gait parameters and conformity. The multi-camera system from the intelligent space acquires images from patients during gait. This way, it is possible to generate tridimensional information for the VR-based system. Among the provided functionalities, the user can explore the virtual environment and manage several features, such as gait reproduction and parameters displayed, using a head-mounted display and hand controllers. Besides, the system presents an automatic classifier that can assist physiotherapists and doctors in assessing abnormalities from conventional human gait. We evaluate the system through two quantitative experiments. The first one addresses the performance evaluation of the automatic classifier. The second analysis is through a Likert scale questionnaire submitted to a group of physiotherapists. In this case, the specialists evaluate the existing features of the proposed framework. The results from the questionnaire showed that the virtual environment is suitable for helping track patients' rehabilitation. Also, the neural network-based classifier results are promising, averaging higher than 91% for all evaluation metrics. Finally, a comparison with related works in the literature highlights the contributions of the proposed solution to the field.

Civil registration and vital statistics systems capture birth and death events to compile vital statistics and to provide legal rights to citizens. Vital statistics are a key factor in promoting public health policies and the health of the population. Medical certification of cause of death is the preferred source of cause of death information. However, two thirds of all deaths worldwide are not captured in routine mortality information systems and their cause of death is unknown. Verbal autopsy is an interim solution for estimating the cause of death distribution at the population level in the absence of medical certification. A Verbal Autopsy (VA) consists of an interview with the relative or the caregiver of the deceased. The VA includes both Closed Questions (CQs) with structured answer options, and an Open Response (OR) consisting of a free narrative of the events expressed in natural language and without any pre-determined structure. There are a number of automated systems to analyze the CQs to obtain cause specific mortality fractions with limited performance. We hypothesize that the incorporation of the text provided by the OR might convey relevant information to discern the CoD. The experimental layout compares existing Computer Coding Verbal Autopsy methods such as Tariff 2.0 with other approaches well suited to the processing of structured inputs as is the case of the CQs. Next, alternative approaches based on language models are employed to analyze the OR. Finally, we propose a new method with a bi-modal input that combines the CQs and the OR. Empirical results corroborated that the CoD prediction capability of the Tariff 2.0 algorithm is outperformed by our method taking into account the valuable information conveyed by the OR. As an added value, with this work we made available the software to enable the reproducibility of the results attained with a version implemented in R to make the comparison with Tariff 2.0 evident.

Machine learning (ML) has demonstrated its ability to exploit important relationships within data collection, which can be used in the diagnosis, treatment, and prediction of outcomes in a variety of clinical contexts. Anxiety mental disorder analysis is one of the pending difficulties that ML can help with. A thorough study is demanded to gain a better understanding of this illness. Since the anxiety data is generally multidimensional, which complicates processing and as a result of technology improvements, medical data from several perspectives, known as multiview data (MVD), is being collected. Each view has its own data type and feature values, so there is a lot of diversity. This work introduces a novel preprocessing feature selection (FS) approach, multiview harris hawk optimization (MHHO), which has the potential to reduce the dimensionality of anxiety data, hence reducing analytical effort. The uniqueness of MHHO originates from combining a multiview linking methodology with the power of the harris hawk optimization (HHO) method. The HHO is used to identify the lowest optimal MVD feature subset, while multiview linking is utilized to find a promising fitness function to direct the HHO FS while accounting for all data views' heterogeneity. The complexity of MHHO is O(THL2), where T is the number of iterations, H is the number of involved harris hawks, and L is the number of objects. Using two publicly available anxiety MVDs, MHHO is validated against ten recent rivals in its category. The experimental findings show that MHHO has a considerable advantage in terms of convergence speed (converging in less than ten iterations), subset size (removing 75% of the views; reducing feature size by 66%), and classification accuracy (approaching 100%). Furthermore, statistical analyses reveal that MHHO is statistically different from its competitors, bolstering its applicability. Finally, feature importance is evaluated, shedding light on the most anxiety-inducing characteristics. The likelihood of developing additional disorders (such as depression or stress) is also investigated.

Predicting the mode of child birth is still remains one of the most complex and challenging tasks in ancient times. Also, there is no such strong methodologies are developed in the conventional works for birth mode prediction. Therefore, the proposed work objects to develop a novel and distinct optimization based machine learning technique for creating the child birth mode prediction system. This framework includes the modules of data imputation, feature selection, classification, and prediction. Initially, the data imputation process is performed to improve the quality of dataset by normalizing the attributes and filling the missed fields. Then, the Multivariate Intensified Mine Blast Optimization (MIMBO) technique is implemented to choose the best set of features by estimating the optimal function. After that, an integrated Naïve Bayes - Random Forest (NBRF) technique is developed by incorporating the functions of conventional NB and RF techniques. The novel contribution of this technique, a Bird Mating (BM) optimization technique is used in NBRF classifier for estimating the likelihood parameter to generate the Bayesian rules. The main idea of this paper is to develop a simple as well as efficient automated system with the use of hybrid machine learning model for predicting the mode of child birth. For this purpose, advanced algorithms such as MIMBO based feature selection, and NBRF based classification are implemented in this work. Due to the inclusion of MIMBO and BM optimization techniques, the performance of classifier is greatly improved with low computational burden and increased prediction accuracy. Moreover, the combination of proposed MIMBO-NBRF technique outperforms the existing child birth prediction methods with superior results in terms of average accuracy up to 99 %. In addition, some other parameters are also estimated and compared with the existing techniques for proving the overall superiority of the proposed framework.