Hasil untuk "Diseases of the endocrine glands. Clinical endocrinology"

Sandeep Kumar Parvathareddy, Abdul K. Siraj, Zeeshan Qadri et al.

BackgroundAge cut-off of 55 years has been included in the eighth edition of the American Joint Committee on Cancer (AJCC) TNM staging, since it led to better prediction of disease-specific survival (DSS) of patients with differentiated thyroid cancer (DTC). However, optimal age cut-off in DTC patients from Middle Eastern ethnicity has not been fully explored.MethodsWe retrospectively analyzed a large cohort of 1721 adult DTC patients. The optimal age cut-off value was determined using several age cut-offs (between 20 and 85 years) to assess DSS. Harrel’s C-Index, Akaike information criterion (AIC) and Bayesian Information Criterion (BIC) were used to assess statistical model performance of the TNM staging system (eighth edition), with different age cut-offs for prediction of DSS.ResultsThe median age of patients at diagnosis was 39.9 years (inter-quartile range 31.0 – 51.7 years) and 75.5% (1299/1721) were female. Median follow up was 9.3 years and 10 years DSS was 97.1%. For DTC overall, an age cut-off of 50 years had the best statistical model performance. On receiver operating characteristic curve analysis, the optimal age cut-off for prediction of DSS was 50.5 years (area under the curve = 0.872, p < 0.0001).ConclusionIn this large cohort of Middle Eastern DTC patients, an age cut-off of 50 years was more appropriate for TNM staging to achieve better predictability for DSS. Therefore, implementation of different age cut-off for DTC in Middle Eastern patients could improve the predictive value for TNM staging system, allowing for better therapeutic and surveillance approach for these patients.

Vivek Yadav, Anugrah Jain

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.

Wuyang Lan, Wenzheng Wang, Changwei Ji et al.

Recent advances in reasoning with large language models (LLMs)has shown remarkable reasoning capabilities in domains such as mathematics and coding, yet their application to clinical diagnosis remains underexplored. Here, we introduce ClinicalGPT-R1, a reasoning enhanced generalist large language model for disease diagnosis. Trained on a dataset of 20,000 real-world clinical records, ClinicalGPT-R1 leverages diverse training strategies to enhance diagnostic reasoning. To benchmark performance, we curated MedBench-Hard, a challenging dataset spanning seven major medical specialties and representative diseases. Experimental results demonstrate that ClinicalGPT-R1 outperforms GPT-4o in Chinese diagnostic tasks and achieves comparable performance to GPT-4 in English settings. This comparative study effectively validates the superior performance of ClinicalGPT-R1 in disease diagnosis tasks. Resources are available at https://github.com/medfound/medfound.

Jindong Wang, Yutong Mao, Xiao Liu et al.

Alzheimer's disease (AD) is a complex, multifactorial neurodegenerative disorder with substantial heterogeneity in progression and treatment response. Despite recent therapeutic advances, predictive models capable of accurately forecasting individualized disease trajectories remain limited. Here, we present a machine learning-based operator learning framework for personalized modeling of AD progression, integrating longitudinal multimodal imaging, biomarker, and clinical data. Unlike conventional models with prespecified dynamics, our approach directly learns patient-specific disease operators governing the spatiotemporal evolution of amyloid, tau, and neurodegeneration biomarkers. Using Laplacian eigenfunction bases, we construct geometry-aware neural operators capable of capturing complex brain dynamics. Embedded within a digital twin paradigm, the framework enables individualized predictions, simulation of therapeutic interventions, and in silico clinical trials. Applied to AD clinical data, our method achieves high prediction accuracy exceeding 90% across multiple biomarkers, substantially outperforming existing approaches. This work offers a scalable, interpretable platform for precision modeling and personalized therapeutic optimization in neurodegenerative diseases.

Haokun Zhao, Yingzhe Bai, Qingyang Xu et al.

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.

Alina Ermilova, Dmitrii Kornilov, Sofia Samoilova et al.

Identifying disease interconnections through manual analysis of large-scale clinical data is labor-intensive, subjective, and prone to expert disagreement. While machine learning (ML) shows promise, three critical challenges remain: (1) selecting optimal methods from the vast ML landscape, (2) determining whether real-world clinical data (e.g., electronic health records, EHRs) or structured disease descriptions yield more reliable insights, (3) the lack of "ground truth," as some disease interconnections remain unexplored in medicine. Large language models (LLMs) demonstrate broad utility, yet they often lack specialized medical knowledge. To address these gaps, we conduct a systematic evaluation of seven approaches for uncovering disease relationships based on two data sources: (i) sequences of ICD-10 codes from MIMIC-IV EHRs and (ii) the full set of ICD-10 codes, both with and without textual descriptions. Our framework integrates the following: (i) a statistical co-occurrence analysis and a masked language modeling (MLM) approach using real clinical data; (ii) domain-specific BERT variants (Med-BERT and BioClinicalBERT); (iii) a general-purpose BERT and document retrieval; and (iv) four LLMs (Mistral, DeepSeek, Qwen, and YandexGPT). Our graph-based comparison of the obtained interconnection matrices shows that the LLM-based approach produces interconnections with the lowest diversity of ICD code connections to different diseases compared to other methods, including text-based and domain-based approaches. This suggests an important implication: LLMs have limited potential for discovering new interconnections. In the absence of ground truth databases for medical interconnections between ICD codes, our results constitute a valuable medical disease ontology that can serve as a foundational resource for future clinical research and artificial intelligence applications in healthcare.

Ricardo Correa

Megan Li Yuen Yeoh, Mahmoud Abouibrahim, Maya Venu

Jiska Embaye, Maartje de Wit, Frank Snoek

BackgroundMyDiaMate is a web-based intervention specifically designed for adults with type 1 diabetes (T1D) that aims to help them improve and maintain their mental health. Prior pilot-testing of MyDiaMate verified its acceptability, feasibility, and usability. ObjectiveThis study aimed to investigate the real-world uptake and usage of MyDiaMate in the Netherlands. MethodsBetween March 2021 and December 2022, MyDiaMate was made freely available to Dutch adults with T1D. Usage (participation and completion rates of the modules) was tracked using log data. Users could volunteer to participate in the user profile study, which required filling out a set of baseline questionnaires. The usage of study participants was examined separately for participants scoring above and below the cutoffs of the “Problem Areas in Diabetes” (PAID-11) questionnaire (diabetes distress), the “World Health Organization Well-being Index” (WHO-5) questionnaire (emotional well-being), and the fatigue severity subscale of the “Checklist Individual Strength” (CIS) questionnaire (fatigue). Two months after creating an account, study participants received an evaluation questionnaire to provide us with feedback. ResultsIn total, 1008 adults created a MyDiaMate account, of whom 343 (34%) participated in the user profile study. The mean age was 43 (SD 14.9; 18-76) years. Most participants were female (n=217, 63.3%) and higher educated (n=198, 57.6%). The majority had been living with T1D for over 5 years (n=241, 73.5%). Of the study participants, 59.1% (n=199) of them reported low emotional well-being (WHO-5 score≤50), 70.9% (n=239) of them reported elevated diabetes distress (PAID-11 score≥18), and 52.4% (n=178) of them reported severe fatigue (CIS score≥35). Participation rates varied between 9.5% (n=19) for social environment to 100% (n=726) for diabetes in balance, which opened by default. Completion rates ranged from 4.3% (n=1) for energy, an extensive cognitive behavioral therapy module, to 68.6% (n=24) for the shorter module on hypos. There were no differences in terms of participation and completion rates of the modules between study participants with a more severe profile, that is, lower emotional well-being, greater diabetes distress, or more fatigue symptoms, and those with a less severe profile. Further, no technical problems were reported, and various suggestions were made by study participants to improve the application, suggesting a need for more personalization. ConclusionsData from this naturalistic study demonstrated the potential of MyDiaMate as a self-help tool for adults with T1D, supplementary to ongoing diabetes care, to improve healthy coping with diabetes and mental health. Future research is needed to explore engagement strategies and test the efficacy of MyDiaMate in a randomized controlled trial.

Jose Miguel Dora, Rafael Selbach Scheffel

Rui Deng, Kaibo Mei, Tiangang Song et al.

BackgroundThe benefit of first-line use of sodium-dependent glucose transport 2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP-1RAs) in type 2 diabetes mellitus (T2DM) with low risk of cardiovascular diseases are not clear.MethodsPubMed, EMBASE and Cochrane Library databases were searched to identify eligible randomized controlled trials. We used the odds ratio (OR) and mean difference (MD) and the corresponding 95% confidence interval (CI) to assess the dichotomous and continuous variable, respectively.ResultsThirteen studies involving 2,885 T2DM at low risk of cardiovascular diseases were included. Compared to placebo, first line use of SGLT2i significantly reduced glycosylated hemoglobin type A1C (HbA1c) (MD: -0.72), weight (MD: -1.32) and fasting plasma glucose (FPG) (MD: -27.05) levels. Compared with metformin, SGLT2i reduced body weight (MD: -1.50) and FPG (MD: -10.13) more effectively, with similar reduction for HbA1c (MD: -0.05). No significant increased safety adverse was found for SGLT2i, including nasopharyngitis (OR: 1.07), urinary tract infection (OR: 2.31), diarrhea (OR: 1.18) and hypoglycemia (OR: 1.06). GLP-1RAs significantly reduced HbA1c (MD: -1.13), weight (MD: -2.12) and FPG (MD: -31.44) levels as first-line therapy compared to placebo. GLP-1RAs significantly increased occurrence of diarrhea (OR: 2.18), hypoglycemia (OR: 3.10), vomiting (OR: 8.22), and nausea (OR: 4.41).ConclusionFirst line use of SGLT2i and GLP-1RAs is effective in reducing HbA1c, weight, and FPG levels in T2DM patients at low risk for cardiovascular disease. SGLT2i may be superior to metformin in controlling body weight and FPG. GLP-1RAs may increase the occurrence of diarrhea, hypoglycemia, vomiting, and nausea.Systematic review registrationPROSPERO (International Prospective Register of Systematic Reviews. https://www.york.ac.uk/inst/crd, CRD42022347233).

Luis Grosembacher

Desde la aprobación del primer sistema infusor automatizado de insulina (AIDs, automated insulin delivery system) en 2016 como tratamiento para personas con diabetes mellitus tipo 1 (DM1), se ha incrementado significativamente el número de usuarios y consorcios de AIDs, mejorando los diseños y prestaciones de sus componentes (infusor, sensor y algoritmo). Los infusores que administran insulina a través de catéteres o sin ellos “bombas parches” son de menores dimensiones y con mayores variantes de operabilidad, desde teléfonos móviles, del propio infusor o mediante dispositivos administradores o bien operables en forma múltiple. Los actuales desafíos para desarrolladores de algoritmos son: a) controlar las variaciones de glucemia en el entorno prandial, con bolos calculados precomidas y bolos correctores automatizados prandiales, o bien anunciando comida, pero sin bolos calculados previos (AIDs semi o totalmente automáticos); b) ajustar la infusión de insulina durante la actividad física y ante cambios agudos de glucemia por estrés físico o emocional. Los numerosos trabajos randomizados y controlados (RCT) en DM1 con AIDs concluyen que los mismos son efectivos para mejorar los niveles de HbA1c, reducir tiempos en rango, y bajar riesgos y tiempos en hipoglucemias. También coinciden en cuanto a beneficios psicosociales, especialmente atenuando el miedo individual y familiar a las hipoglucemias. Los estudios del mundo real con AIDs confirman los beneficios metabólicos y de calidad de vida de los RCT. Lo anterior explica el uso creciente de los AIDs en todas las edades de DM1, siendo mayor su preferencia en adultos mayores y niños. En el primer consenso sobre recomendaciones de los AIDs se sugirió que la falla del tratamiento con múltiples dosis de insulina para alcanzar objetivos metabólicos, elevada variabilidad glucémica, hipoglucemias problemáticas e hiperglucemia nocturna (fenómeno del alba) en DM1 serían las indicaciones principales de los AIDs, entre otras. Algunos pendientes a resolver con los AIDs son: lograr una equitativa accesibilidad a todos los estratos socioeconómicos, adecuar la tecnológica en DM1 con capacidades diferentes (p. ej., ceguera), adaptar algoritmos en situaciones especiales (p. ej., embarazos u hospitalizaciones) y realizar RCT prospectivos y comparativos entre diferentes AIDs. Podemos concluir que los AIDs constituyen un cambio paradigmático en el tratamiento de la DM1, con un horizonte tecnológico sin límites, dirigido a alcanzar objetivos metabólicos, a atenuar la carga de tareas diarias y distrés por la enfermedad para finalmente elevar su calidad de vida.

Eren Er, Semih Aşıkovalı, Hatice Özışık et al.

ABSTRACT Objective: The aim of this study is to investigate the molecular genetic causes of non-syndromic primary ovarian insufficiency (POI) cases with the gene panel based on next generation sequencing analysis and to establish the relationship between genotype and phenotype. Subjects and methods: Twenty three cases aged 14-40 years followed up with POI were included. Patients with a karyotype of 46, XX, primary or secondary amenorrhea before the age of 40, with elevated FSH (>40 IU/mL) and low AMH levels (<0.03 ng/mL) were included in the study. Molecular genetic analyzes were performed by the next generation sequencing analysis method targeted with the TruSightTM Exome panel. Results: Median age of the cases was 17.8 (14.0-24.3) years, and 12 (52%) cases admitted before the age of 18. Fifteen (65%) patients had consanguineous parents. In 2 (8.6%) cases, variants detected were in genes that have been previously proven to cause POI. One was homozygous variant in FIGLA gene and the other was homozygous variant in PSMC3IP gene. Heterozygous variants were detected in PROK2, WDR11 and CHD7 associated with hypogonadotropic hypogonadism, but these variants are insufficient to contribute to the POI phenotype. Conclusion: Genetic panels based on next generation sequencing analysis technologies can be used to determine the molecular genetic diagnosis of POI, which has a highly heterogeneous genetic basis.

Jennifer M. Barker

Fabienne A U Fox, Lennart Koch, Monique M B Breteler et al.

Objective: Maintaining muscle function throughout life is critical for healthy ageing. Although in vitro studies consistently indicate beneficial effects of 25-hydroxyvitamin D (25-OHD) on muscle function, findings from population-based studies remain inconclusive. We therefore aimed to examine the association between 25-OHD concentration and handgrip strength across a wide age range and assess potential modifying effects of age, sex and season. Methods: We analysed cross-sectional baseline data of 2576 eligible participants out of the first 3000 participants (recruited from March 2016 to March 2019) of the Rhineland Study, a community-based cohort study in Bonn, Germany. Multivariate linear regression models were used to assess the relation between 25-OHD levels and grip strength while adjusting for age, sex, education, smoking, season, body mass index, physical activity levels, osteoporosis and vitamin D supplementation. Results: Compared to participants with deficient 25-OHD levels (<30 nmol/L), grip strength was higher in those with inadequate (30 to <50 nmol/L) and adequate (≥50 to ≤125 nmol/L) levels (ßinadequate = 1.222, 95% CI: 0.377; 2.067, P = 0.005; ßadequate = 1.228, 95% CI: 0.437; 2.019, P = 0.002). Modelling on a continuous scale revealed grip strength to increase with higher 25-OHD levels up to ~100 nmol/L, after which the direction reversed (ßlinear = 0.505, 95% CI: 0.179; 0.830, P = 0.002; ßquadratic = –0.153, 95% CI: –0.269; -0.038, P = 0.009). Older adults showed weaker effects of 25-OHD levels on grip strength than younger adults (ß25OHDxAge = –0.309, 95% CI: –0.594; –0.024, P = 0.033). Conclusions: Our findings highlight the importance of sufficient 25-OHD levels for optimal muscle function across the adult life span. However, vitamin D supplementation should be closely monitored to avoid detrimental effects.

Camilla P. Dias-Rocha, Julia C. B. Costa, Yamara S. Oliveira et al.

IntroductionMaternal high-fat (HF) diet during gestation and lactation programs obesity in rat offspring associated with sex-dependent and tissue-specific changes of the endocannabinoid system (ECS). The ECS activation induces food intake and preference for fat as well as lipogenesis. We hypothesized that maternal HF diet would increase the lipid endocannabinoid levels in breast milk programming cannabinoid and dopamine signaling and food preference in rat offspring.MethodsFemale Wistar rats were assigned into two experimental groups: control group (C), which received a standard diet (10% fat), or HF group, which received a high-fat diet (29% fat) for 8 weeks before mating and during gestation and lactation. Milk samples were collected to measure endocannabinoids and fatty acids by mass spectrometry. Cannabinoid and dopamine signaling were evaluated in the nucleus accumbens (NAc) of male and female weanling offspring. C and HF offspring received C diet after weaning and food preference was assessed in adolescence.ResultsMaternal HF diet reduced the milk content of anandamide (AEA) (p&lt;0.05) and 2-arachidonoylglycerol (2-AG) (p&lt;0.05). In parallel, maternal HF diet increased adiposity in male (p&lt;0.05) and female offspring (p&lt;0.05) at weaning. Maternal HF diet increased cannabinoid and dopamine signaling in the NAc only in male offspring (p&lt;0.05), which was associated with higher preference for fat in adolescence (p&lt;0.05).ConclusionContrary to our hypothesis, maternal HF diet reduced AEA and 2-AG in breast milk. We speculate that decreased endocannabinoid exposure during lactation may induce sex-dependent adaptive changes of the cannabinoid-dopamine crosstalk signaling in the developing NAc, contributing to alterations in neurodevelopment and programming of preference for fat in adolescent male offspring.

Tahmina Sultana Priya, Fan Leng, Anthony C. Luehrs et al.

Non-alcoholic fatty liver disease (NAFLD) is a prevalent chronic liver disorder characterized by the excessive accumulation of fat in the liver in individuals who do not consume significant amounts of alcohol, including risk factors like obesity, insulin resistance, type 2 diabetes, etc. We aim to identify subgroups of NAFLD patients based on demographic, clinical, and genetic characteristics for precision medicine. The genomic and phenotypic data (3,408 cases and 4,739 controls) for this study were gathered from participants in Mayo Clinic Tapestry Study (IRB#19-000001) and their electric health records, including their demographic, clinical, and comorbidity data, and the genotype information through whole exome sequencing performed at Helix using the Exome+$^\circledR$ Assay according to standard procedure (www$.$helix$.$com). Factors highly relevant to NAFLD were determined by the chi-square test and stepwise backward-forward regression model. Latent class analysis (LCA) was performed on NAFLD cases using significant indicator variables to identify subgroups. The optimal clustering revealed 5 latent subgroups from 2,013 NAFLD patients (mean age 60.6 years and 62.1% women), while a polygenic risk score based on 6 single-nucleotide polymorphism (SNP) variants and disease outcomes were used to analyze the subgroups. The groups are characterized by metabolic syndrome, obesity, different comorbidities, psychoneurological factors, and genetic factors. Odds ratios were utilized to compare the risk of complex diseases, such as fibrosis, cirrhosis, and hepatocellular carcinoma (HCC), as well as liver failure between the clusters. Cluster 2 has a significantly higher complex disease outcome compared to other clusters. Keywords: Fatty liver disease; Polygenic risk score; Precision medicine; Deep phenotyping; NAFLD comorbidities; Latent class analysis.

Mohammad Junayed Hasan, Suhra Noor, Mohammad Ashrafuzzaman Khan

Clinical texts, such as admission notes, discharge summaries, and progress notes, contain rich and valuable information that can be used for clinical decision making. However, a severe bottleneck in using transformer encoders for processing clinical texts comes from the input length limit of these models: transformer-based encoders use fixed-length inputs. Therefore, these models discard part of the inputs while processing medical text. There is a risk of losing vital knowledge from clinical text if only part of it is processed. This paper proposes a novel method to preserve the knowledge of long clinical texts in the models using aggregated ensembles of transformer encoders. Previous studies used either ensemble or aggregation, but we studied the effects of fusing these methods. We trained several pre-trained BERT-like transformer encoders on two clinical outcome tasks: mortality prediction and length of stay prediction. Our method achieved better results than all baseline models for prediction tasks on long clinical notes. We conducted extensive experiments on the MIMIC-III clinical database's admission notes by combining multiple unstructured and high-dimensional datasets, demonstrating our method's effectiveness and superiority over existing approaches. This study shows that fusing ensemble and aggregation improves the model performance for clinical prediction tasks, particularly the mortality and the length of hospital stay.

Hong-Yu Zhou, Yizhou Yu, Chengdi Wang et al.

During the diagnostic process, clinicians leverage multimodal information, such as chief complaints, medical images, and laboratory-test results. Deep-learning models for aiding diagnosis have yet to meet this requirement. Here we report a Transformer-based representation-learning model as a clinical diagnostic aid that processes multimodal input in a unified manner. Rather than learning modality-specific features, the model uses embedding layers to convert images and unstructured and structured text into visual tokens and text tokens, and bidirectional blocks with intramodal and intermodal attention to learn a holistic representation of radiographs, the unstructured chief complaint and clinical history, structured clinical information such as laboratory-test results and patient demographic information. The unified model outperformed an image-only model and non-unified multimodal diagnosis models in the identification of pulmonary diseases (by 12% and 9%, respectively) and in the prediction of adverse clinical outcomes in patients with COVID-19 (by 29% and 7%, respectively). Leveraging unified multimodal Transformer-based models may help streamline triage of patients and facilitate the clinical decision process.

Ting Liu, Xi Guo, Ying Liao et al.

In the past decade, the incidence of recurrent pregnancy loss (RPL) has increased significantly, and immunological disorders have been considered as one of the possible causes contributing to RPL. The presence of antinuclear antibodies (ANAs) is regarded as a typical antibody of autoimmunity. However, the relationship between the presence of ANAs and RPL, the underlying mechanism, and the possible role of immunotherapy is still controversial. The aim of this mini review is to assess the association between ANAs and RPL and the effects of immunotherapy on pregnancy outcomes in women with positive ANAs and a history of RPL from the available data and to provide a relevant reference basis for clinical application in this group of women.