51886 research outputs found
Sort by
Parechovirus meningitis in infants: report of the first cluster of cases in Saudi Arabia
No full textParechoviruses (PeVs) are RNA viruses that belong to the Picornaviridae family. PeV-A is known to cause human disease, particularly in neonates and young infants. The symptoms of infection can range from mild febrile and respiratory illnesses to severe sepsis-like conditions and central nervous system (CNS) infections. In 2022, clusters of PeV infections were reported in various parts of the world, prompting public health agencies to issue a health advisory to alert physicians about circulating PeVs causing CNS infections in children. Reports of PeV infections in Gulf Council countries are scarce. We present here the clinical presentation and outcome of PeV CNS infections in a neonate and a young infant
Therapeutic Targeting of BET Proteins in Sarcoma
No full textBromodomain and extraterminal (BET) domain protein family are epigenetic readers that regulate gene transcription, cell-cycle progression, and DNA damage response (DDR), making them attractive therapeutic targets for sarcomas, which are epigenetically dysregulated and genomically unstable. Sarcomas are molecularly heterogeneous with a high propensity for metastasis, resulting in poor clinical outcomes. BET inhibitors (BETi) hold promise for the treatment of sarcomas, for they block interaction of BETs with acetylated lysines, modify gene expression, and create an imbalance in transcription and replication kinetics. BETis also disrupt transcriptional programs driven by oncogenic fusion proteins found in some sarcomas. Preclinical studies demonstrate efficacy of BETis in inducing apoptosis, disrupting DDR, and reducing tumor growth, either as monotherapy or in combination with chemotherapy or other targeted agents, such as PI3K, histone deacetylases, and CHK1 inhibitors. Favorable results have been observed in clinical trials, but more studies are required to fully assess safety and efficacy as well as identify biomarkers of response and resistance. Ongoing research is focused on optimizing BETi safety and selectivity and exploring combination therapies, such as BETis with DDR inhibitors. This review summarizes the preclinical studies on BET inhibition and discusses clinical trial activity, providing insights into the potential of BETis in sarcoma therapy
Four-dimensional left atrial strain in a murine model of reversible pressure overload
No full textIntroduction: Left atrial (LA) strain is emerging as a valuable metric for assessing heart failure progression and treatment efficacy. However, the complex geometry and contractile patterns of the LA complicate the evaluation of global and regional function specifically during pressure overload and recovery.
Methods: We use advanced four-dimensional ultrasound (4DUS) imaging (Vevo 3100, VisualSonics) and develop novel analytical techniques to evaluate the progression of LA function during pressure overload and recovery. A cohort of mice (n=8) underwent 4DUS imaging to assess LA and left ventricular (LV) function at baseline, 3 weeks post-transverse aortic constriction (TAC), and 1 and 4 weeks post-minimally invasive aortic debanding (deTAC) (Figure 1).
Results: LV ejection fraction (EF) significantly decreased 3-weeks post-TAC (p=0.0008) before gradually returning to baseline levels 4-weeks post-deTAC. LA max volume was 11.8 ± 0.8μL at baseline and increased to 30.4 ± 9.4μL 3-weeks post-TAC (p=0.0024) before returning to baseline levels 1-week post-deTAC. Global circumferential strain dropped from 27.7 ± 2.9% at baseline to 8.3 ± 1.0% at 3 weeks post-TAC and improved to 20.7 ± 1.2% by 4 weeks post-deTAC. Significant differences were observed at each time point (Figure 2 A-C). Regional strain analysis revealed superior LA strain differences at baseline, 1 week, and 4 weeks post-deTAC (Figure 2 D-E). Global longitudinal strain decreased (from 24.8 ± 3.6% at baseline to 10.2 ± 3.7% post-TAC, p<0.0001) and partially recovered to 18.9 ± 3.5% at 4 weeks post-deTAC (p=0.0073 vs. baseline). Regional differences were evident in posterior LA strain post-TAC and 4 weeks post-deTAC.
Conclusion: Advanced 4DUS imaging and strain analysis techniques enable quantification of global LA function and provide detailed segmental mapping to track the dynamics of LA reverse remodeling. This approach offers a valuable tool for monitoring therapeutic efficacy and optimizing patient management strategies in heart failure or aortic stenosis
Evaluating the Reasoning Capabilities of Large Language Models for Medical Coding and Hospital Readmission Risk Stratification: Zero-Shot Prompting Approach
No full textBackground: Large language models (LLMs) such as ChatGPT-4, LLaMA-3.1, Gemini-1.5, DeepSeek-R1, and OpenAI-O3 have shown promising potential in health care, particularly for clinical reasoning and decision support. However, their reliability across critical tasks like diagnosis, medical coding, and risk prediction has received mixed reviews, especially in real-world settings without task-specific training.
Objective: This study aims to evaluate and compare the zero-shot performance of reasoning and nonreasoning LLMs in three essential clinical tasks: (1) primary diagnosis generation, (2) ICD-9 (International Classification of Diseases, Ninth Revision) medical code prediction, and (3) hospital readmission risk stratification. The goal is to assess whether these models can serve as general-purpose clinical decision support tools and to identify gaps in current capabilities.
Methods: Using the Medical Information Mart for Intensive Care-IV dataset, we selected a random cohort of 300 hospital discharge summaries. Prompts were engineered to include structured clinical content from 5 note sections: chief complaints, past medical history, surgical history, laboratories, and imaging. Prompts were standardized and zero-shot, with no model fine-tuning or repetition across runs. All model interactions were conducted through publicly available web user interfaces, without using application programming interfaces, to simulate real-world accessibility for nontechnical users. We incorporated rationale elicitation into prompts to evaluate model transparency, especially in reasoning models. Ground-truth labels were derived from the primary diagnosis documented in clinical notes, structured ICD-9 codes from diagnosis, and hospital-recorded readmission frequencies for risk stratification. Performance was measured using F1-scores and correctness percentages, and comparative performance was analyzed statistically.
Results: Among nonreasoning models, LLaMA-3.1 achieved the highest primary diagnosis accuracy (n=255, 85%), followed by ChatGPT-4 (n=254, 84.7%) and Gemini-1.5 (n=237, 79%). For ICD-9 prediction, correctness dropped significantly across all models: LLaMA-3.1 (n=128, 42.6%), ChatGPT-4 (n=122, 40.6%), and Gemini-1.5 (n=44, 14.6%). Hospital readmission risk prediction showed low performance in nonreasoning models: LLaMA-3.1 (n=124, 41.3%), Gemini-1.5 (n=122, 40.7%), and ChatGPT-4 (n=99, 33%). Among reasoning models, OpenAI-O3 outperformed in diagnosis (n=270, 90%) and ICD-9 coding (n=136, 45.3%), while DeepSeek-R1 performed slightly better in the readmission risk prediction (n=218, 72.6% vs O3's n=212, 70.6%). Despite improved explainability, reasoning models generated verbose responses. None of the models met clinical standards across all tasks, and performance in medical coding remained the weakest area across all models.
Conclusions: Current LLMs exhibit moderate success in zero-shot diagnosis and risk prediction but underperform in ICD-9 code generation, reinforcing findings from prior studies. Reasoning models offer marginally better performance and increased interpretability, with limited reliability. Overall, statistical analysis between the models revealed that OpenAI-O3 outperformed the other models. These results highlight the need for task-specific fine-tuning and need human-in-the-loop checking. Future work will explore fine-tuning, stability through repeated trials, and evaluation on a different subset of deidentified real-world data with a larger sample size
A digital workflow for full-arch implant fixed prosthesis: Clinical application of a 3D-printed polychromatic flangeless trial denture, anatomic bone reduction guide, and intraoral scanning for definitive prosthesis fabrication
No full textThis clinical report presents a comprehensive digital workflow for rehabilitating a patient with maxillary terminal dentition using a full-arch, implant-supported fixed dental prosthesis (FP-1). It highlights the integration of a 3D-printed polychromatic flangeless trial denture and a customized anatomic bone reduction template, enabling prosthetically driven implant planning and optimal bone architecture modification. The workflow incorporated fully guided implant surgery using sequential templates and immediate loading with a closed-mouth pickup system. Intraoral scanning protocols were employed for definitive prosthesis fabrication. This approach addressed high esthetic risks while achieving surgical precision, efficient treatment execution, and favorable clinical outcomes
Optimization and comparison of [18F]FET synthesis on two distinct automated radiochemistry systems
No full textO-(2-[18F]Fluoroethyl)- L -tyrosine ([18F]FET) is a promising amino acid PET tracer for assessment of malignant brain tumors. Herein, we report optimized production of [18F]FET for clinical use on two commercial radiochemistry systems: Sofie ELIXYS and GE FASTlab 2. While the Sofie ELIXYS procedure requires high performance liquid chromatography (HPLC) purification, the GE FASTlab 2 method uses solid-phase extraction (SPE) purification. In both cases, [18F]FET met release specifications for clinical investigation laid out in the United States Pharmacopeia (USP) and/or European Pharmacopeia (Ph. Eur.). The radiochemical yield of [18F]FET was 35-55% and 30-55% decay corrected to start of synthesis (SOS) for Sofie ELIXYS and GE FASTlab 2, respectively. The overall synthesis time was 75-85 and 70-80 min from SOS for Sofie ELIXYS and GE FASTlab 2, respectively. The radiochemical purity was > 99%, and the molar activity (Am) was 340-464 GBq/µmol at end of synthesis (EOS)
Decision-Level Multi-Sensor Fusion to Improve Limitations of Single-Camera-Based CNN Classification in Precision Farming: Application in Weed Detection
No full textThe United States leads in corn production and consumption in the world with an estimated USD 50 billion per year. There is a pressing need for the development of novel and efficient techniques aimed at enhancing the identification and eradication of weeds in a manner that is both environmentally sustainable and economically advantageous. Weed classification for autonomous agricultural robots is a challenging task for a single-camera-based system due to noise, vibration, and occlusion. To address this issue, we present a multi-camera-based system with decision-level sensor fusion to improve the limitations of a single-camera-based system in this paper. This study involves the utilization of a convolutional neural network (CNN) that was pre-trained on the ImageNet dataset. The CNN subsequently underwent re-training using a limited weed dataset to facilitate the classification of three distinct weed species: Xanthium strumarium (Common Cocklebur), Amaranthus retroflexus (Redroot Pigweed), and Ambrosia trifida (Giant Ragweed). These weed species are frequently encountered within corn fields. The test results showed that the re-trained VGG16 with a transfer-learning-based classifier exhibited acceptable accuracy (99% training, 97% validation, 94% testing accuracy) and inference time for weed classification from the video feed was suitable for real-time implementation. But the accuracy of CNN-based classification from video feed from a single camera was found to deteriorate due to noise, vibration, and partial occlusion of weeds. Test results from a single-camera video feed show that weed classification accuracy is not always accurate for the spray system of an agricultural robot (AgBot). To improve the accuracy of the weed classification system and to overcome the shortcomings of single-sensor-based classification from CNN, an improved Dempster–Shafer (DS)-based decision-level multi-sensor fusion algorithm was developed and implemented. The proposed algorithm offers improvement on the CNN-based weed classification when the weed is partially occluded. This algorithm can also detect if a sensor is faulty within an array of sensors and improves the overall classification accuracy by penalizing the evidence from a faulty sensor. Overall, the proposed fusion algorithm showed robust results in challenging scenarios, overcoming the limitations of a single-sensor-based system
Robot-Assisted Laparoscopic Repair of a Paraesophageal Hernia During the 3rd Trimester of Pregnancy
No full textBackground: New hiatal hernia diagnosis in pregnancy is rare. Type IV hernias comprise less than 5% of those diagnosed and have been historically managed non-operatively or by laparoscopic exploration. Previously described outcomes include significant maternal and perinatal morbidity and mortality. Minimal literature exists on the outcomes of robot-assisted laparoscopic surgery during pregnancy for any indication due to the rarity of these cases. We present a case of a type IV hernia diagnosed in the third trimester repaired via robot-assisted laparoscopy with minimal peri/postoperative maternal or fetal complications.
Case Description: A 29-year-old primigravida with prior sleeve gastrectomy presented at 27+6 weeks gestation with acute left upper quadrant abdominal pain, nausea, and vomiting. Computed tomography scan showed a type IV paraesophageal hernia containing stomach, small bowel, and colon; surgical repair was advised. She was counseled on the fetal risks of non-obstetric surgery, and the maternal-fetal medicine team recommended antenatal corticosteroids, pre/postoperative non-stress test (NST), left tilt of the hips intraoperatively, indomethacin for tocolysis, and mechanical and pharmacologic deep vein thrombosis prophylaxis. A robot-assisted laparoscopic hernia repair with mesh was performed with subcostal margin port placement to avoid the gravid uterus. Surgery was complicated by a left capnothorax requiring decompression. Postoperative NST was reassuring, postoperative milestones were met, and discharge occurred on postoperative day 2. Subsequent pregnancy course was uneventful. Labor was induced at 37+0 weeks gestation due to gestational hypertension, and uncomplicated vaginal delivery occurred the next day with discharge on postpartum day 1.
Clinical Significance: This case contributes to the limited reported outcomes on this rare diagnosis and uncommon surgical approach in the third trimester.
Conclusion: Robot-assisted surgical repair of paraesophageal hernias may be successfully performed in the third trimester
CalTrig: A GUI-Based Machine Learning Approach for Decoding Neuronal Calcium Transients in Freely Moving Rodents
No full textAdvances in in vivo Ca2+ imaging using miniature microscopes have enabled researchers to study single-neuron activity in freely moving animals. Tools such as Minian and CalmAn have been developed to convert Ca2+ visual signals to numerical data, collectively referred to as CalV2N. However, substantial challenges remain in analyzing the large datasets generated by CalV2N, particularly in integrating data streams, evaluating CalV2N output quality, and reliably and efficiently identifying Ca2+ transients. In this study, we introduce CalTrig, an open-source graphical user interface (GUI) tool designed to address these challenges at the post-CalV2N stage of data processing collected from C57BL/6J mice. CalTrig integrates multiple data streams, including Ca2+ imaging, neuronal footprints, Ca2+ traces, and behavioral tracking, and offers capabilities for evaluating the quality of CalV2N outputs. It enables synchronized visualization and efficient Ca2+ transient identification. We evaluated four machine learning models (i.e., GRU, LSTM, Transformer, and Local Transformer) for Ca2+ transient detection. Our results indicate that the GRU model offers the highest predictability and computational efficiency, achieving stable performance across training sessions, different animals, and even among different brain regions. The integration of manual, parameter-based, and machine learning-based detection methods in CalTrig provides flexibility and accuracy for various research applications. The user-friendly interface and low computing demands of CalTrig make it accessible to neuroscientists without programming expertise. We further conclude that CalTrig enables deeper exploration of brain function, supports hypothesis generation about neuronal mechanisms, and opens new avenues for understanding neurological disorders and developing treatments
Discovery of Novel and Selective GPR17 Antagonists as Pharmacological Tools for Developing New Therapeutic Strategies in Diabetes and Obesity
No full textG protein coupled receptors (GPCRs) are promising targets for diabetes and obesity therapy due to their roles in metabolism and excellent potential for pharmacological manipulation. We previously reported that Gpr17 ablation in the brain-gut axis leads to improved metabolic homeostasis, suggesting GPR17 antagonism could be developed for diabetes and obesity treatment. Here, we performed high throughput screening (HTS) and identified two new GPR17 antagonists (compound 978 and 527). Both compounds antagonized downstream Gαi/o, Gαq and β-arrestin signaling with high selectivity for GPR17, but not the closely related purinergic and cysteinyl leukotriene receptors. The molecular mechanisms of antagonism were revealed through Schild analysis, structure-activity relationship (SAR) studies and homology modeling. Compound 978, a competitive antagonist against the surrogate small molecule agonist MDL29,951 (MDL), and its analog (793) attenuated GPR17 signaling and promoted glucagon-like peptide-1 (GLP-1) secretion in enteroendocrine cells. In summary, we identified selective GPR17 antagonists through HTS, which represent promising pharmacological tools for developing new therapeutic strategies in diabetes and obesity