Rowzatul Zannat, Abdullah Al Shafi, Abdul Muntakim
Increased access to reliable health information is essential for non-English-speaking populations, yet resources in Bangla for disease prediction remain limited. This study addresses this gap by developing a comprehensive Bangla symptoms-disease dataset containing 758 unique symptom-disease relationships spanning 85 diseases. To ensure transparency and reproducibility, we also make our dataset publicly available. The dataset enables the prediction of diseases based on Bangla symptom inputs, supporting healthcare accessibility for Bengali-speaking populations. Using this dataset, we evaluated multiple machine learning models to predict diseases based on symptoms provided in Bangla and analyzed their performance on our dataset. Both soft and hard voting ensemble approaches combining top-performing models achieved 98\% accuracy, demonstrating superior robustness and generalization. Our work establishes a foundational resource for disease prediction in Bangla, paving the way for future advancements in localized health informatics and diagnostic tools. This contribution aims to enhance equitable access to health information for Bangla-speaking communities, particularly for early disease detection and healthcare interventions.
Raquel Behar-Lagares, Raquel Behar-Lagares, Ana Virseda-Berdices
et al.
IntroductionA significant proportion of COVID-19 Intensive Care Unit (ICU) survivors develop long-term respiratory complications, including pulmonary fibrosis. Telomere attrition, a marker of cellular senescence, has emerged as a potential biomarker for post-COVID-19 sequelae. This study investigated the association between peripheral blood relative telomere length (RTL) and long-term pulmonary outcomes in COVID-19 ICU survivors, with a specific focus on gender-specific differences.MethodsICU-admitted COVID-19 patients were followed for at least one year post-discharge. RTL was quantified from peripheral blood using monochromatic multiplex quantitative PCR (MMqPCR) at hospital admission and one-year post-discharge. Primary outcomes were respiratory symptoms and diffuse parenchymal lung disease (DPLD), assessed via imaging. Data were analyzed using gender-stratified generalized linear models, adjusted for clinical covariates.ResultsAt one year, 43.8% of patients reported respiratory symptoms and 23.9% developed DPLD. A total of 73 ICU survivors were included, with 51 men and 22 women. At one year, 43.8% of patients reported respiratory symptoms and 23.9% developed DPLD. Longitudinal analysis showed significant RTL shortening in both men and women who underwent IMV (p=0.011 and p=0.016, respectively), and in men who needed pronation during their ICU stay (p=0.037). Regarding one-year symptoms, in women, repeated-measures analysis showed an association with persistent respiratory symptoms, particularly in those who needed pronation during their ICU stay [adjusted arithmetic mean ratio (aAMR)=0.73) (95%CI=0.60-0.90); p=0.003]. At follow-up, women who had undergone pronation and had shorter RTL continued to show a higher prevalence of symptoms [aAMR= 0.66 (0.58-0.76); p< 0.001]. In contrast, men with shorter RTL at one-year post-discharge had an association with the presence of DPLD [aAMR = 0.64 (0.50-0.81); p = 0.001]. This association in men was significant particularly among those who required IMV [aAMR= 0.61 (0.49-0.76); p< 0.001] or prone positioning [aAMR= 0.56 (0.44-0.71); p= 0.016]. DiscussionThese findings underscore the role of telomere attrition as a sex-specific biomarker of aging-associated pulmonary vulnerability in the aftermath of critical COVID-19 illness. RTL may serve as a prognostic marker for long-term respiratory sequelae, potentially guiding risk stratification and individualized follow-up strategies in post-ICU COVID-19 survivors.
Generalized pustular psoriasis (GPP) is a rare, potentially life-threatening inflammatory disease characterized by neutrophilic pustules and systemic inflammation. We report a case of severe GPP triggered by SARS-CoV-2 infection in a 46-year-old woman with a long history of psoriasis. Eleven days after recovery from COVID-19 pneumonia, she developed widespread pustules and fever. Histopathology revealed subcorneal spongiform pustules and dermal neutrophilic infiltration consistent with GPP. Systemic corticosteroids followed by etretinate and deucravacitinib achieved complete remission. A literature review identified 11 infection- and 10 vaccine-related GPP cases. Compared with vaccine-associated cases, infection-related flares showed longer latency and higher corticosteroid use. Mechanistically, both SARS-CoV-2 infection and vaccination may be associated with IL-36 axis activation, potentially via spike protein–driven, Toll-like receptor–mediated innate immune signaling. This case highlights that distinct immune kinetics may underlie infection- and vaccine-related GPP, while supporting a putative role of IL-36–driven inflammation in COVID-19–associated disease exacerbation.
Shaheer Ahmad Khan, Muhammad Usamah Shahid, Ahmad Abdullah
et al.
This study addresses a critical gap in the healthcare system by developing a clinically meaningful, practical, and explainable disease surveillance system for multiple chronic diseases, utilizing routine EHR data from multiple U.S. practices integrated with CureMD's EMR/EHR system. Unlike traditional systems--using AI models that rely on features from patients' labs--our approach focuses on routinely available data, such as medical history, vitals, diagnoses, and medications, to preemptively assess the risks of chronic diseases in the next year. We trained three distinct models for each chronic disease: prediction models that forecast the risk of a disease 3, 6, and 12 months before a potential diagnosis. We developed Random Forest models, which were internally validated using F1 scores and AUROC as performance metrics and further evaluated by a panel of expert physicians for clinical relevance based on inferences grounded in medical knowledge. Additionally, we discuss our implementation of integrating these models into a practical EMR system. Beyond using Shapley attributes and surrogate models for explainability, we also introduce a new rule-engineering framework to enhance the intrinsic explainability of Random Forests.
India, as a predominantly agrarian economy, faces significant challenges in agriculture, including substantial crop losses caused by diseases, pests, and environmental stress. Early detection and accurate identification of diseases across different crops are critical for improving yield and ensuring food security. This paper proposes a deep learning based solution for detecting multiple diseases in multiple crops, aimed to cover India's diverse agricultural landscape. We first create a unified dataset encompassing images of 17 different crops and 34 different diseases from various available repositories. Proposed deep learning model is trained on this dataset and outperforms the state-of-the-art in terms of accuracy and the number of crops, diseases covered. We achieve a significant detection accuracy, i.e., 99 percent for our unified dataset which is 7 percent more when compared to state-of-the-art handling 14 crops and 26 different diseases only. By improving the number of crops and types of diseases that can be detected, proposed solution aims to provide a better product for Indian farmers.
Accurate disease detection is of paramount importance for effective medical treatment and patient care. However, the process of disease detection is often associated with extensive medical testing and considerable costs, making it impractical to perform all possible medical tests on a patient to diagnose or predict hundreds or thousands of diseases. In this work, we propose Collaborative Learning for Disease Detection (CLDD), a novel graph-based deep learning model that formulates disease detection as a collaborative learning task by exploiting associations among diseases and similarities among patients adaptively. CLDD integrates patient-disease interactions and demographic features from electronic health records to detect hundreds or thousands of diseases for every patient, with little to no reliance on the corresponding medical tests. Extensive experiments on a processed version of the MIMIC-IV dataset comprising 61,191 patients and 2,000 diseases demonstrate that CLDD consistently outperforms representative baselines across multiple metrics, achieving a 6.33\% improvement in recall and 7.63\% improvement in precision. Furthermore, case studies on individual patients illustrate that CLDD can successfully recover masked diseases within its top-ranked predictions, demonstrating both interpretability and reliability in disease prediction. By reducing diagnostic costs and improving accessibility, CLDD holds promise for large-scale disease screening and social health security.
Nathalia A. Loureiro, Camilo R. Neto, Jack Sutton
et al.
Inter-city interactions are critical for the transmission of infectious diseases, yet their effects on the scaling of disease cases remain largely underexplored. Here, we use the commuting network as a proxy for inter-city interactions, integrating it with a general scaling framework to describe the incidence of seven infectious diseases across Brazilian cities as a function of population size and the number of commuters. Our models significantly outperform traditional urban scaling approaches, revealing that the relationship between disease cases and a combination of population and commuters varies across diseases and is influenced by both factors. Although most cities exhibit a less-than-proportional increase in disease cases with changes in population and commuters, more-than-proportional responses are also observed across all diseases. Notably, in some small and isolated cities, proportional rises in population and commuters correlate with a reduction in disease cases. These findings suggest that such towns may experience improved health outcomes and socioeconomic conditions as they grow and become more connected. However, as growth and connectivity continue, these gains diminish, eventually giving way to challenges typical of larger urban areas - such as socioeconomic inequality and overcrowding - that facilitate the spread of infectious diseases. Our study underscores the interconnected roles of population size and commuter dynamics in disease incidence while highlighting that changes in population size exert a greater influence on disease cases than variations in the number of commuters.
Sri Varsha Mulakala, G. Neeharika, P. Vinay Kumar
et al.
The recent increase in morbidity is primarily due to chronic diseases including Diabetes, Heart disease, Lung cancer, and brain tumours. The results for patients can be improved, and the financial burden on the healthcare system can be lessened, through the early detection and prevention of certain disorders. In this study, we built a machine-learning model for predicting the existence of numerous diseases utilising datasets from various sources, including Kaggle, Dataworld, and the UCI repository, that are relevant to each of the diseases we intended to predict. Following the acquisition of the datasets, we used feature engineering to extract pertinent features from the information, after which the model was trained on a training set and improved using a validation set. A test set was then used to assess the correctness of the final model. We provide an easy-to-use interface where users may enter the parameters for the selected ailment. Once the right model has been run, it will indicate whether the user has a certain ailment and offer suggestions for how to treat or prevent it.
Background: We aimed to assess knowledge of emergency department (ED) physicians regarding the classification and treatment of angioedema and to evaluate the impact of a training program on this knowledge base. Methods: A total of 11 questions about angioedema and its types were posed to ED physicians from various hospitals, either in person or via e-mail, before the implementation of an educational module on the subject (pre-test). Following a brief training period, the ED physicians were presented with the same set of questions once again (post-test). The reliability between the repeated tests were estimated using intraclass correlation coefficients (ICC). Item difficulty was calculated separately for each question in both the pre-test and post-test. Results: A total of 541 ED physicians participated in the pre-test survey, and 162 of them declined to participate in the post-test survey. The remaining 379 participated in the post-test survey as well. The ICC between the repeated tests indicated a moderate level of reliability (mean ICC = 0.5; 0.42–0.57; lower and upper 95% confidence intervals). The mean item difficulty was 0.36 in the pre-test, indicating that the items had an appropriate level of difficulty. In the post-test, the mean item difficulty increased to 0.57, suggesting that the items were generally easier, likely reflecting improved knowledge or skills following the intervention. The level of knowledge regarding the clinical features of different types of angioedema was found to be inadequate. Following the training period, there was a notable increase in the number of correct answers, with a statistically significant difference (p = 0.002). Similarly, a remarkable increase was observed in the number of respondents who indicated that bradykinin-mediated-angioedema should be considered in cases of unresponsiveness to antihistamine and-corticosteroid treatment (p < 0.001). Regarding queries about hereditary angioedema (HAE), the majority of ED physicians had no prior experience in treating a patient with HAE, and only a small number were familiar with the symptoms of HAE. Following the training ED physicians demonstrated enhanced knowledge of HAE symptoms and diagnostic criteria (p < 0.001). Similarly, a notable enhancement in familiarity with HAE attack treatments was observed when the inquiries related to these treatments were compared between the pre-test and post-test phases (p < 0.001). Conclusion: In light of the potential lethality of attacks mediated by bradykinin, a training program should include the recognition of rare types of angioedema, with a particular emphasis on HAE disease.
Aimee D. Potter, Vonetta L. Edwards, Adonis D’Mello
et al.
Neisseria gonorrhoeae (the gonococcus, Gc) causes the sexually transmitted infection gonorrhea. Gc is a prominent threat to human health by causing severe lifelong sequelae, including infertility and chronic pelvic pain, which is amplified by the emergence of “superbug” strains resistant to all current antibiotics. Gc is highly adapted to colonize human mucosal surfaces, where it survives despite initiating a robust inflammatory response and influx of polymorphonuclear leukocytes (PMNs, neutrophils) that typically clear bacteria. Here, dual-species RNA-sequencing was used to define Gc and PMN transcriptional profiles alone and after infection. Core host and bacterial responses were assessed for two strains of Gc and three human donors’ PMNs. Comparative analysis of Gc transcripts revealed overlap between Gc responses to PMNs, iron, and hydrogen peroxide; 98 transcripts were differentially expressed across both Gc strains in response to PMN co-culture, including iron-responsive and oxidative stress response genes. We experimentally determined that the iron-dependent TbpB is suppressed by PMN co-culture, and iron-limited Gc have a survival advantage when cultured with PMNs. Analysis of PMN transcripts modulated by Gc infection revealed differential expression of genes driving cell adhesion, migration, inflammatory responses, and inflammation resolution pathways. Production of pro-inflammatory cytokines, including IL1B and IL8, the adhesion factor ICAM1, and prostaglandin PGE2 were induced in PMNs in response to Gc. Together, this study represents a comprehensive and experimentally validated dual-species transcriptomic analysis of two isolates of Gc and primary human PMNs that gives insight into how this bacterium survives innate immune onslaught to cause disease.
Generalist Large Language Models (LLMs), such as GPT-4, have shown considerable promise in various domains, including medical diagnosis. Rare diseases, affecting approximately 300 million people worldwide, often have unsatisfactory clinical diagnosis rates primarily due to a lack of experienced physicians and the complexity of differentiating among many rare diseases. In this context, recent news such as "ChatGPT correctly diagnosed a 4-year-old's rare disease after 17 doctors failed" underscore LLMs' potential, yet underexplored, role in clinically diagnosing rare diseases. To bridge this research gap, we introduce RareBench, a pioneering benchmark designed to systematically evaluate the capabilities of LLMs on 4 critical dimensions within the realm of rare diseases. Meanwhile, we have compiled the largest open-source dataset on rare disease patients, establishing a benchmark for future studies in this domain. To facilitate differential diagnosis of rare diseases, we develop a dynamic few-shot prompt methodology, leveraging a comprehensive rare disease knowledge graph synthesized from multiple knowledge bases, significantly enhancing LLMs' diagnostic performance. Moreover, we present an exhaustive comparative study of GPT-4's diagnostic capabilities against those of specialist physicians. Our experimental findings underscore the promising potential of integrating LLMs into the clinical diagnostic process for rare diseases. This paves the way for exciting possibilities in future advancements in this field.
Existing plant disease classification models have achieved remarkable performance in recognizing in-laboratory diseased images. However, their performance often significantly degrades in classifying in-the-wild images. Furthermore, we observed that in-the-wild plant images may exhibit similar appearances across various diseases (i.e., small inter-class discrepancy) while the same diseases may look quite different (i.e., large intra-class variance). Motivated by this observation, we propose an in-the-wild multimodal plant disease recognition dataset that contains the largest number of disease classes but also text-based descriptions for each disease. Particularly, the newly provided text descriptions are introduced to provide rich information in textual modality and facilitate in-the-wild disease classification with small inter-class discrepancy and large intra-class variance issues. Therefore, our proposed dataset can be regarded as an ideal testbed for evaluating disease recognition methods in the real world. In addition, we further present a strong yet versatile baseline that models text descriptions and visual data through multiple prototypes for a given class. By fusing the contributions of multimodal prototypes in classification, our baseline can effectively address the small inter-class discrepancy and large intra-class variance issues. Remarkably, our baseline model can not only classify diseases but also recognize diseases in few-shot or training-free scenarios. Extensive benchmarking results demonstrate that our proposed in-the-wild multimodal dataset sets many new challenges to the plant disease recognition task and there is a large space to improve for future works.
Carlo Perricone, Andrea Castellucci, Giacomo Cafaro
et al.
The treatment of patients with rheumatoid arthritis (RA) has dramatically changed in the past 30 years. Currently, numerous conventional, biologic, and targeted synthetic DMARDs have been licensed and used following recommendations provided by international and national scientific societies. However, the availability of biosimilars and the increasing necessity of savings impacted on the local/national prescription of these drugs. The information provided by data sheet of every single drug is a decisive factor on the choice of a certain treatment merged with the patient’s profile. Thus, our purpose was to construct a rational algorithm for the treatment strategy in RA according to costs and the product leaflet of the biologic and targeted-synthetic DMARDs currently licensed in Italy. We used the most recent available recommendations and then we performed a review of the literature considering all the factors that are known to influence drug safety/effectiveness. All these factors were considered in the context of the data sheets of currently available originators and biosimilars.
Yannis Papanikolaou, Francesco Tuveri, Misa Ogura
et al.
In this work we present a deep learning approach to conduct hypothesis-free, transcriptomics-based matching of drugs for diseases. Our proposed neural network architecture is trained on approved drug-disease indications, taking as input the relevant disease and drug differential gene expression profiles, and learns to identify novel indications. We assemble an evaluation dataset of disease-drug indications spanning 68 diseases and evaluate in silico our approach against the most widely used transcriptomics-based matching baselines, CMap and the Characteristic Direction. Our results show a more than 200% improvement over both baselines in terms of standard retrieval metrics. We further showcase our model's ability to capture different genes' expressions interactions among drugs and diseases. We provide our trained models, data and code to predict with them at https://github.com/healx/dgem-nn-public.
Cardiovascular diseases are widespread among patients with chronic noncommunicable diseases and are one of the leading causes of death, including in the working age. The article presents the relevance of the development and application of patient-oriented systems, in which machine learning (ML) is a promising technology that allows predicting cardiovascular diseases. Automated machine learning (AutoML) makes it possible to simplify and speed up the process of developing AI/ML applications, which is key in the development of patient-oriented systems by application users, in particular medical specialists. The authors propose a framework for the application of automatic machine learning and three scenarios that allowed for data combining five data sets of cardiovascular disease indicators from the UCI Machine Learning Repository to investigate the effectiveness in detecting this class of diseases. The study investigated one AutoML model that used and optimized the hyperparameters of thirteen basic ML models (KNeighborsUnif, KNeighborsDist, LightGBMXT, LightGBM, RandomForestGini, RandomForestEntr, CatBoost, ExtraTreesGini, ExtraTreesEntr, NeuralNetFastA, XGBoost, NeuralNetTorch, LightGBMLarge) and included the most accurate models in the weighted ensemble. The results of the study showed that the structure of the AutoML model for detecting cardiovascular diseases depends not only on the efficiency and accuracy of the basic models used, but also on the scenarios for preprocessing the initial data, in particular, on the technique of data normalization. The comparative analysis showed that the accuracy of the AutoML model in detecting cardiovascular disease varied in the range from 87.41% to 92.3%, and the maximum accuracy was obtained when normalizing the source data into binary values, and the minimum was obtained when using the built-in AutoML technique.
Peter A. Tauber, Bernhard Kratzer, Philipp Schatzlmaier
et al.
BackgroundTreg cells have been shown to be an important part of immune-homeostasis and IL-2 which is produced upon T cell receptor (TCR)-dependent activation of T lymphocytes has been demonstrated to critically participate in Treg development.ObjectiveTo evaluate small molecule inhibitors (SMI) for the identification of novel IL-2/Treg enhancing compounds.Materials and methodsWe used TCR-dependent and allergen-specific cytokine secretion of human and mouse T cells, next generation messenger ribonucleic acid sequencing (RNA-Seq) and two different models of allergic airway inflammation to examine lead SMI-compounds.ResultsWe show here that the reported 3-phosphoinositide dependent kinase-1 (PDK1) SMI BX-795 increased IL-2 in culture supernatants of Jurkat E6-1 T cells, human peripheral blood mononuclear cells (hPBMC) and allergen-specific mouse T cells upon TCR-dependent and allergen-specific stimulation while concomitantly inhibiting Th2 cytokine secretion. RNA-Seq revealed that the presence of BX-795 during allergen-specific activation of T cells induces a bona fide Treg cell type highly similar to iTreg but lacking Foxp3 expression. When applied in mugwort pollen and house dust mite extract-based models of airway inflammation, BX-795 significantly inhibited Th2 inflammation including expression of Th2 signature transcription factors and cytokines and influx into the lungs of type 2-associated inflammatory cells such as eosinophils.ConclusionsBX-795 potently uncouples IL-2 production from Th2 inflammation and induces Th-IL-2 cells, which highly resemble induced (i)Tregs. Thus, BX-795 may be a useful new compound for the treatment of allergic diseases.
Human cytomegalovirus (HCMV) infection is globally distributed, and the liver is one of the major targeting organs. So far, the mechanisms for cell and organ damage have not fully been elucidated and the treatments for the infection are mainly at symptoms. IL-37 has shown a protective role in certain inflammatory diseases. In the present study, potential protective effect of exogenous IL-37 on murine cytomegalovirus- (MCMV-) infected hepatitis was evaluated through analyses of serum transaminases, the liver histopathology and cytokine expression, and functional state of dendritic cells (DCs) and regulatory T cells (Tregs). These analyses showed a significant decrease in serum transaminase levels and a lower Ishak histopathologic score at the early stage of MCMV-infected mice with exogenous IL-37 pretreatment. The frequencies of MHC-Ⅱ, CD40, CD80, and CD86 positive DCs in the liver and spleen were decreased significantly at 7 days postinfection (dpi) in MCMV-infected mice with IL-37 pretreatment when compared with those without the pretreatment, while the total number of DCs in the liver was reduced in IL-37-pretreated mice. The induction of Tregs in the spleen was enhanced at dpi 3 with IL-37 pretreatment in MCMV-infected mice. The mRNA expression levels of cytokines in the liver were decreased significantly (IL-1β, IL-6, IL-10, IL-4) or to some extent (TGF-β and TNF-α). The present study suggested that exogenous IL-37 can alleviate MCMV-infected hepatitis, likely through reduced DCs and induced Tregs with a weaker cytokine storm, demonstrating its potential value in clinical management for HCMV-infected hepatitis.
Veronica Ghini, Gaia Meoni, Lorenzo Pelagatti
et al.
Metabolomics and lipidomics have been used in several studies to define the biochemical alterations induced by COVID-19 in comparison with healthy controls. Those studies highlighted the presence of a strong signature, attributable to both metabolites and lipoproteins/lipids. Here, 1H NMR spectra were acquired on EDTA-plasma from three groups of subjects: i) hospitalized COVID-19 positive patients (≤21 days from the first positive nasopharyngeal swab); ii) hospitalized COVID-19 positive patients (>21 days from the first positive nasopharyngeal swab); iii) subjects after 2-6 months from SARS-CoV-2 eradication. A Random Forest model built using the EDTA-plasma spectra of COVID-19 patients ≤21 days and Post COVID-19 subjects, provided a high discrimination accuracy (93.6%), indicating both the presence of a strong fingerprint of the acute infection and the substantial metabolic healing of Post COVID-19 subjects. The differences originate from significant alterations in the concentrations of 16 metabolites and 74 lipoprotein components. The model was then used to predict the spectra of COVID-19>21 days subjects. In this group, the metabolite levels are closer to those of the Post COVID-19 subjects than to those of the COVID-19≤21 days; the opposite occurs for the lipoproteins. Within the acute phase patients, characteristic trends in metabolite levels are observed as a function of the disease severity. The metabolites found altered in COVID-19≤21 days patients with respect to Post COVID-19 individuals overlap with acute infection biomarkers identified previously in comparison with healthy subjects. Along the trajectory towards healing, the metabolome reverts back to the "healthy" state faster than the lipoproteome.