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

Lara Janssens, Lara Janssens, Ella Roelant et al.

Over the last decades, the gonadotropin-releasing hormone (GnRH) antagonist protocol has become widely used for prevention of premature luteinizing hormone surge during ovarian stimulation with exogenous gonadotropins. Literature has shown its efficacy and safety, while maintaining similar live birth rates compared to agonist protocols. Clinicians occasionally notice a drop in serum estradiol levels after GnRH antagonist initiation. This study aimed to analyze the impact of a drop in serum estradiol levels after the administration of a GnRH antagonist on clinical outcomes in in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) cycles. The results showed that estradiol drop was related to lower ongoing pregnancy, live birth and cumulative live birth rates. Estradiol drop was less frequent in cycles using gonadotropins containing luteinizing hormone (LH) activity and cycles with a drop showed a larger decrease in serum LH after the first GnRH antagonist administration, resulting in lower serum values on the day of ovulation triggering. These findings suggest that future prospective research could focus on potential optimization of luteinizing hormone (LH) levels and this study might also add to the discussion that estradiol monitoring could be useful in detecting cycles with an increased risk of less favorable outcomes.

Mohammad Tahmid Noor, Shayan Abrar, Jannatul Adan Mahi et al.

Early and accurate classification of retinal diseases is critical to counter vision loss and for guiding clinical management of retinal diseases. In this study, we proposed a deep learning method for retinal disease classification utilizing optical coherence tomography (OCT) images from the Retinal OCT Image Classification - C8 dataset (comprising 24,000 labeled images spanning eight conditions). Images were resized to 224x224 px and tested on convolutional neural network (CNN) architectures: Xception and InceptionV3. Data augmentation techniques (CutMix, MixUp) were employed to enhance model generalization. Additionally, we applied GradCAM and LIME for interpretability evaluation. We implemented this in a real-world scenario via our web application named RetinaVision. This study found that Xception was the most accurate network (95.25%), followed closely by InceptionV3 (94.82%). These results suggest that deep learning methods allow effective OCT retinal disease classification and highlight the importance of implementing accuracy and interpretability for clinical applications.

Yuka Ono, Norio Wada, Shuhei Baba et al.

We report the case of a 41-year-old Japanese woman with visual field disturbances during late pregnancy. At 39 weeks of gestation, she was diagnosed with bitemporal hemianopsia at the ophthalmology department. An MRI revealed a symmetrical pituitary gland enlargement, compressing the optic chiasm. An emergency cesarean section was performed immediately, resulting in the delivery of a male infant weighing 3,112 grams. Laboratory tests indicated low serum free thyroxine (T4), thyroid-stimulating hormone (TSH), cortisol, luteinizing hormone, and follicle-stimulating hormone. The patient was clinically diagnosed with lymphocytic hypophysitis (LHy). Due to her visual field impairment, she was administered 60 mg of prednisolone daily. After 2 days, her visual field impairment improved rapidly, leading to a gradual tapering of the dose. Six months after treatment initiation, an MRI showed shrinkage of the pituitary gland. Her prednisolone dose was reduced to 5 mg daily, and she was switched to hydrocortisone at 15 mg daily. Twelve months after starting treatment, the patient developed thyrotoxicosis. Testing revealed a positive TSH receptor antibody, resulting in a diagnosis of Graves’ disease (GD). Treatment with thiamazole (15 mg daily) and potassium iodide (76 mg daily) was initiated, and her thyroid function normalized after 2 months. LHy is believed to have an autoimmune mechanism and is frequently associated with other autoimmune diseases; however, the development of GD is rare. Development of Graves’ disease should be considered in patients with LHy, particularly during the postpartum period and the glucocorticoid treatment process.

Jose Fuentes-López, Rubén Vidal-Espinoza, Ofelia Mamani-Luque et al.

ObjectiveThe evaluation of bone health during the growth stage is extremely important, as it is a key factor to prevent bone diseases in adulthood. The objectives of the study were: a) to verify if there are differences in bone health with other geographic regions, b) to develop bone health curves using quantitative ultrasonography (QUS) through the Broadband Ultrasonic Attenuation (BUA) parameter in children and adolescents residing in a high altitude region of Peru and c) to determine specific cut-off points for bone health assessment in this particular population.MethodsA cross-sectional study was carried out in schoolchildren in a high altitude region of Peru. The sample consisted of 1468 children and adolescents (724 males and 744 females). The age range was 6.0 to 17.9 years old. Weight and height were evaluated. Body Mass Index (BMI) was calculated. Bone quality was evaluated by quantitative ultrasonography (QUS) of the calcaneus. The parameters measured were Speed of Sound (SOS, m/s); Broadband Ultrasonic Attenuation (BUA, dB/MHz); and Bone Quality Index (BQI= αSOS+αBUA, αβ: temperature corrections).ResultsThere were small discrepancies in bone health (BUA) between studies from various geographic regions. Values differed across all age ranges from ~0.36 to ~10.86 in males and from ~0.26 to ~6.68 in females. At later ages during adolescence the values are relatively similar, reaching a plateau around 16 and 17 years of age. Percentiles were calculated for BUA by age and sex. Sensitivity and specificity values in females are slightly higher relative to males. However, the Youden Index reflects 0.84 for both sexes and the appropriate cut-off point for men is ≤67.8 and for women is ≤63.7.ConclusionThe study demonstrated that there are small discrepancies in bone health (BUA) among children between children and adolescents from different geographic regions. These findings support the creation of specific references and cut-off points for bone health in the pediatric population of a high altitude region of Peru. The results suggest the application of percentiles for the assessment of bone health in school and epidemiological contexts.

Jiawang Huang, Yucheng Jin, Zhigang Yi

The Leucine-rich repeat-containing G protein-coupled receptor 4 (LGR4) is a member of the G protein-coupled receptor family and plays an important role in bone metabolism and tumor bone metastasis. LGR4 affects bone metabolism by regulating the differentiation and activity of osteoblasts and osteoclasts, and is involved in the balance between bone resorption and bone formation. Deficiency of LGR4 leads to osteoporosis, whereas the up-regulation of LGR4 may help to alleviate the development of traumatic osteoarthritis. Furthermore, in breast cancer and multiple myeloma, LGR4 promotes tumor cell metastasis to bone tissue by activating related signaling pathways. Therefore, LGR4 may be a potential target for the treatment of bone metabolic diseases and tumor bone metastasis.

Weichen Si, Yihao Ou, Zhen Tian

In this study, we propose a machine learning-based method for noise reduction and disease-causing gene feature extraction in gene sequencing DeepSeqDenoise algorithm combines CNN and RNN to effectively remove the sequencing noise, and improves the signal-to-noise ratio by 9.4 dB. We screened 17 key features by feature engineering, and constructed an integrated learning model to predict disease-causing genes with 94.3% accuracy. We successfully identified 57 new candidate disease-causing genes in a cardiovascular disease cohort validation, and detected 3 missed variants in clinical applications. The method significantly outperforms existing tools and provides strong support for accurate diagnosis of genetic diseases.

Wenliang Li, Rui Yan, Xu Zhang et al.

Large language models (LLMs) have demonstrated notable potential in medical applications, yet they face substantial challenges in handling complex real-world clinical diagnoses using conventional prompting methods. Current prompt engineering and multi-agent approaches typically optimize isolated inferences, neglecting the accumulation of reusable clinical experience. To address this, this study proposes a novel Multi-Agent Clinical Diagnosis (MACD) framework, which allows LLMs to self-learn clinical knowledge via a multi-agent pipeline that summarizes, refines, and applies diagnostic insights. It mirrors how physicians develop expertise through experience, enabling more focused and accurate diagnosis on key disease-specific cues. We further extend it to a MACD-human collaborative workflow, where multiple LLM-based diagnostician agents engage in iterative consultations, supported by an evaluator agent and human oversight for cases where agreement is not reached. Evaluated on 4,390 real-world patient cases across seven diseases using diverse open-source LLMs (Llama-3.1 8B/70B, DeepSeek-R1-Distill-Llama 70B), MACD significantly improves primary diagnostic accuracy, outperforming established clinical guidelines with gains up to 22.3% (MACD). In direct comparison with physician-only diagnosis under the same evaluation protocol, MACD achieves comparable or superior performance, with improvements up to 16%. Furthermore, the MACD-human workflow yields an 18.6% improvement over physician-only diagnosis, demonstrating the synergistic potential of human-AI collaboration. Notably, the self-learned clinical knowledge exhibits strong cross-model stability, transferability across LLMs, and capacity for model-specific personalization.This work thus presents a scalable self-learning paradigm that bridges the gap between the intrinsic knowledge of LLMs.

Lang Qin, Zijian Gan, Xu Cao et al.

Computational drug repurposing for rare diseases is especially challenging when no prior associations exist between drugs and target diseases. Therefore, knowledge graph completion and message-passing GNNs have little reliable signal to learn and propagate, resulting in poor performance. We present RareAgent, a self-evolving multi-agent system that reframes this task from passive pattern recognition to active evidence-seeking reasoning. RareAgent organizes task-specific adversarial debates in which agents dynamically construct evidence graphs from diverse perspectives to support, refute, or entail hypotheses. The reasoning strategies are analyzed post hoc in a self-evolutionary loop, producing textual feedback that refines agent policies, while successful reasoning paths are distilled into transferable heuristics to accelerate future investigations. Comprehensive evaluations reveal that RareAgent improves the indication AUPRC by 18.1% over reasoning baselines and provides a transparent reasoning chain consistent with clinical evidence.

Stanley S. Schwartz, Mary E. Herman

Our understanding of type 2 diabetes (T2D) has evolved dramatically. Advances have upended entrenched dogmas pertaining to the onset and progression of T2D, beliefs that have prevailed from the early era of diabetes research—and continue to populate our medical textbooks and continuing medical education materials. This review article highlights key insights that lend new governing principles for gold standard management of T2D. From the historical context upon which old beliefs arose to new findings, this article outlines evidence and perspectives on beta cell function, the underlying defects in glucoregulation, the remediable nature of T2D, and, the rationale supporting the shift to complication-centric prescribing. Practical approaches translate this rectified understanding of T2D into strategies that fill gaps in current management practices of prediabetes through late type 2 diabetes.

Anna L. Trella, Kelly W. Zhang, Hinal Jajal et al.

Dental disease is a prevalent chronic condition associated with substantial financial burden, personal suffering, and increased risk of systemic diseases. Despite widespread recommendations for twice-daily tooth brushing, adherence to recommended oral self-care behaviors remains sub-optimal due to factors such as forgetfulness and disengagement. To address this, we developed Oralytics, a mHealth intervention system designed to complement clinician-delivered preventative care for marginalized individuals at risk for dental disease. Oralytics incorporates an online reinforcement learning algorithm to determine optimal times to deliver intervention prompts that encourage oral self-care behaviors. We have deployed Oralytics in a registered clinical trial. The deployment required careful design to manage challenges specific to the clinical trials setting in the U.S. In this paper, we (1) highlight key design decisions of the RL algorithm that address these challenges and (2) conduct a re-sampling analysis to evaluate algorithm design decisions. A second phase (randomized control trial) of Oralytics is planned to start in spring 2025.

Clément Christophe, Praveen K Kanithi, Tathagata Raha et al.

Med42-v2 introduces a suite of clinical large language models (LLMs) designed to address the limitations of generic models in healthcare settings. These models are built on Llama3 architecture and fine-tuned using specialized clinical data. They underwent multi-stage preference alignment to effectively respond to natural prompts. While generic models are often preference-aligned to avoid answering clinical queries as a precaution, Med42-v2 is specifically trained to overcome this limitation, enabling its use in clinical settings. Med42-v2 models demonstrate superior performance compared to the original Llama3 models in both 8B and 70B parameter configurations and GPT-4 across various medical benchmarks. These LLMs are developed to understand clinical queries, perform reasoning tasks, and provide valuable assistance in clinical environments. The models are now publicly available at \href{https://huggingface.co/m42-health}{https://huggingface.co/m42-health}.

Mehmet Yigit Avci, Emily Chan, Veronika Zimmer et al.

With the increasing incidence of neurodegenerative diseases such as Alzheimer's Disease (AD), there is a need for further research that enhances detection and monitoring of the diseases. We present MORPHADE (Morphological Autoencoders for Alzheimer's Disease Detection), a novel unsupervised learning approach which uses deformations to allow the analysis of 3D T1-weighted brain images. To the best of our knowledge, this is the first use of deformations with deep unsupervised learning to not only detect, but also localize and assess the severity of structural changes in the brain due to AD. We obtain markedly higher anomaly scores in clinically important areas of the brain in subjects with AD compared to healthy controls, showcasing that our method is able to effectively locate AD-related atrophy. We additionally observe a visual correlation between the severity of atrophy highlighted in our anomaly maps and medial temporal lobe atrophy scores evaluated by a clinical expert. Finally, our method achieves an AUROC of 0.80 in detecting AD, out-performing several supervised and unsupervised baselines. We believe our framework shows promise as a tool towards improved understanding, monitoring and detection of AD. To support further research and application, we have made our code publicly available at github.com/ci-ber/MORPHADE.

Shahran Rahman Alve, Muhammad Zawad Mahmud, Samiha Islam et al.

Artificial intelligence and deep learning are increasingly applied in the clinical domain, particularly for early and accurate disease detection using medical imaging and sound. Due to limited trained personnel, there is a growing demand for automated tools to support clinicians in managing rising patient loads. Respiratory diseases such as cancer and diabetes remain major global health concerns requiring timely diagnosis and intervention. Auscultation of lung sounds, combined with chest X-rays, is an established diagnostic method for respiratory illness. This study presents a Deep Convolutional Neural Network (CNN)-based approach for the analysis of respiratory sound data to detect Chronic Obstructive Pulmonary Disease (COPD). Acoustic features extracted with the Librosa library, including Mel-Frequency Cepstral Coefficients (MFCCs), Mel-Spectrogram, Chroma, Chroma (Constant Q), and Chroma CENS, were used in training. The system also classifies disease severity as mild, moderate, or severe. Evaluation on the ICBHI database achieved 96% accuracy using 10-fold cross-validation and 90% accuracy without cross-validation. The proposed network outperforms existing methods, demonstrating potential as a practical tool for clinical deployment.

Cécile Trottet, Manuel Schürch, Ahmed Allam et al.

We propose a deep generative approach using latent temporal processes for modeling and holistically analyzing complex disease trajectories, with a particular focus on Systemic Sclerosis (SSc). We aim to learn temporal latent representations of the underlying generative process that explain the observed patient disease trajectories in an interpretable and comprehensive way. To enhance the interpretability of these latent temporal processes, we develop a semi-supervised approach for disentangling the latent space using established medical knowledge. By combining the generative approach with medical definitions of different characteristics of SSc, we facilitate the discovery of new aspects of the disease. We show that the learned temporal latent processes can be utilized for further data analysis and clinical hypothesis testing, including finding similar patients and clustering SSc patient trajectories into novel sub-types. Moreover, our method enables personalized online monitoring and prediction of multivariate time series with uncertainty quantification.

Qiuhao Lu, Rui Li, Andrew Wen et al.

Large Language Models (LLMs) have revolutionized various sectors, including healthcare where they are employed in diverse applications. Their utility is particularly significant in the context of rare diseases, where data scarcity, complexity, and specificity pose considerable challenges. In the clinical domain, Named Entity Recognition (NER) stands out as an essential task and it plays a crucial role in extracting relevant information from clinical texts. Despite the promise of LLMs, current research mostly concentrates on document-level NER, identifying entities in a more general context across entire documents, without extracting their precise location. Additionally, efforts have been directed towards adapting ChatGPT for token-level NER. However, there is a significant research gap when it comes to employing token-level NER for clinical texts, especially with the use of local open-source LLMs. This study aims to bridge this gap by investigating the effectiveness of both proprietary and local LLMs in token-level clinical NER. Essentially, we delve into the capabilities of these models through a series of experiments involving zero-shot prompting, few-shot prompting, retrieval-augmented generation (RAG), and instruction-fine-tuning. Our exploration reveals the inherent challenges LLMs face in token-level NER, particularly in the context of rare diseases, and suggests possible improvements for their application in healthcare. This research contributes to narrowing a significant gap in healthcare informatics and offers insights that could lead to a more refined application of LLMs in the healthcare sector.

Mahmoud Sakr, Tarek Koraitim

Vin Tangpricha

Tarang K Vora, Rahul Lath, Meenakshi Swain et al.

Cong He, Dong Xie, Lin-feng Fu et al.

ObjectiveEarly identifying arteriosclerosis in newly diagnosed type 2 diabetes (T2D) patients could contribute to choosing proper subjects for early prevention. Here, we aimed to investigate whether radiomic intermuscular adipose tissue (IMAT) analysis could be used as a novel marker to indicate arteriosclerosis in newly diagnosed T2D patients.MethodsA total of 549 patients with newly diagnosed T2D were included in this study. The clinical information of the patients was recorded and the carotid plaque burden was used to indicate arteriosclerosis. Three models were constructed to evaluate the risk of arteriosclerosis: a clinical model, a radiomics model (a model based on IMAT analysis proceeded on chest CT images), and a clinical-radiomics combined model (a model that integrated clinical-radiological features). The performance of the three models were compared using the area under the curve (AUC) and DeLong test. Nomograms were constructed to indicate arteriosclerosis presence and severity. Calibration curves and decision curves were plotted to evaluate the clinical benefit of using the optimal model.ResultsThe AUC for indicating arteriosclerosis of the clinical-radiomics combined model was higher than that of the clinical model [0.934 (0.909, 0.959) vs. 0.687 (0.634, 0.730), P < 0.001 in the training set, 0.933 (0.898, 0.969) vs. 0.721 (0.642, 0.799), P < 0.001 in the validation set]. Similar indicative efficacies were found between the clinical-radiomics combined model and radiomics model (P = 0.5694). The AUC for indicating the severity of arteriosclerosis of the combined clinical-radiomics model was higher than that of both the clinical model and radiomics model [0.824 (0.765, 0.882) vs. 0.755 (0.683, 0.826) and 0.734 (0.663, 0.805), P < 0.001 in the training set, 0.717 (0.604, 0.830) vs. 0.620 (0.490, 0.750) and 0.698 (0.582, 0.814), P < 0.001 in the validation set, respectively]. The decision curve showed that the clinical-radiomics combined model and radiomics model indicated a better performance than the clinical model in indicating arteriosclerosis. However, in indicating severe arteriosclerosis, the clinical-radiomics combined model had higher efficacy than the other two models.ConclusionRadiomics IMAT analysis could be a novel marker for indicating arteriosclerosis in patients with newly diagnosed T2D. The constructed nomograms provide a quantitative and intuitive way to assess the risk of arteriosclerosis, which may help clinicians comprehensively analyse radiomics characteristics and clinical risk factors more confidently.

Qi-Yu Wang, Qi-Yu Wang, Zhe-Zhen Liao et al.