Citations

Primary care is a language-intensive environment where clinicians frequently synthesize, document, and communicate large amounts of information. Given the amount of information in primary care, large language models (LLMs) have emerged as tools to support primary care clinician workflows, aiming to reduce administrative burden and augment patient care. LLMs are artificial intelligence tools that can generate text and assist with chart summarization, message drafting, and clinical decision support. However, because LLMs are probabilistic, identical inputs may yield inconsistent responses. The LLMs output can be significantly improved through prompt engineering. Prompt engineering is the intentional design of text entered into the LLM to help produce better LLM results. Prompting is best approached as an iterative process where a clinician defines the task, the LLM generates output, and the clinician critically reviews and refines the results. Low-quality prompts are often vague and produce generic or misleading responses. High-quality prompts are specific, context-rich, and well-structured. Using structured prompting frameworks can improve the LLMs response. Including core elements, such as the LLMs role, the clinical context, the requested task, and the desired output formatting, can help improve the LLMs response by providing situational background and framing response formatting. Additional frameworks, such as Ask-Context-Expectation and PICO+O, can be utilized for specific patient circumstances and cases. However, despite their promise, LLMs have important limitations. They may generate inaccurate or hallucinated information and produce inconsistent results. Given these limitations, LLMs should not be viewed as knowledge authorities but rather as cognitive assistants. Thoughtful prompt design can help reduce these limitations and improve response accuracy, always keeping in mind that LLMs should not replace human judgment or clinical reasoning. As LLMs continue to be integrated into primary care, prompt engineering will become an increasingly important clinical skill. Careful prompt design and critical appraisal of LLM responses will be key to maintaining excellent patient care in the AI era.

Abstract How glial cells sense and regulate neuronal hypoactivity remains poorly understood. Using in vivo imaging in anesthetized, sleeping, and behaving mice, we identified a glial purinergic signaling mechanism that converts neuronal hypoactivity into local circuit regulation. Cortical hypoactivity increased spatially confined ATP release through astrocytic pannexin-1 hemichannels, which were structurally enriched near perisomatic parvalbumin-positive boutons. Local ATP release directed microglial process movement and stabilized bulbous endings through microglial P2Y12 signaling, accompanied by localized Ca²⁺ activity in microglial processes. Disruption of astrocytic pannexin-1 or microglial P2Y12 impaired bulbous-ending stabilization and abolished rebound increases in neuronal activity during emergence from anesthesia. Similar ATP release and bulbous-ending formation were observed during chemogenetic neuronal silencing or natural sleep in freely behaving mice. These findings identify astrocyte-to-microglia purinergic signaling as a mechanism that converts neuronal hypoactivity into synapse-selective microglial regulation of circuit activity.

Background/Objectives: To develop and evaluate a deep learning-based quality control system for Lumbar Spinal Digital Radiographs (LSDR), designed to automate and improve their evaluation and reduce reliance on manual reviews. Methods: This retrospective study utilized a deep learning workflow comprising image segmentation, feature extraction, and a classification model. The dataset, including anteroposterior (AP) and lateral (LAT) X-ray images, was expanded through data augmentation techniques. Four U-Net-based models were assessed: standard U-Net, Swin-UNet, Attention U-Net, and Attention U-Net with the weight map, with the latter selected for its superior performance. Extracted features, such as brightness, contrast, and anatomical positioning, were used in an XGBoost classifier, which was evaluated using mean intersection over union (mIoU), accuracy, sensitivity, specificity, and AUC. Results: The Attention U-Net with weighted attention outperformed the other models, achieving high mIoU scores in both AP and LAT views. The XGBoost classifier achieved the best performance in classifying images as “qualified” or “unqualified,” with an AUC of approximately 0.9, high accuracy, and balanced sensitivity and specificity. This approach effectively addressed class imbalances and improved model accuracy compared to traditional machine learning models such as MLP and SVM. Conclusions: The developed automated quality control system demonstrated potential for enhancing image quality, enhancing diagnostic reliability, and optimizing clinical workflow efficiency.

Immune-mediated epilepsies represent an etiologically unique category defined by immune activation within the central nervous system that initiates a pathophysiological cascade resulting in seizures, encephalopathy, and sometimes epileptogenesis with long-term epilepsy. Once considered rare and largely confined to paraneoplastic or postinfectious syndromes, immune-mediated and inflammatory mechanisms are now recognized across a broad spectrum ranging from acute autoimmune encephalitis with reversible seizures, new-onset refractory status epilepticus, or febrile illness-related refractory status epilepticus, to chronic autoimmune-associated epilepsy and drug-refractory structural epilepsies. This review integrates historical perspectives, mechanistic insights from animal and human studies, and contemporary clinical frameworks. It synthesizes advances in neuroimmunology, clinical phenotyping, diagnostics, immunomodulatory and antiseizure therapies, neuromodulation, and patient and family centered outcomes, and outlines future directions focused on biomarker-driven precision medicine, disease-modifying strategies, and interdisciplinary care models.

Interest in using gene therapy to treat osteoarthritis (OA) is growing and a number of clinical trials have been initiated. This commentary identifies three intersecting areas that need to be addressed for the field to move forward expeditiously. The first relates to lowering the cost of manufacturing clinical grade viral vectors, addressing various aspects of their deployment, and overcoming immune barriers to dosing and re-dosing. The second area requires an improved understanding of the pathophysiology of OA, including its stratification by endotype and phenotype. Coupled to the development of reliable biomarkers, this will enable the creation of personalized gene therapies, facilitate patient selection, and aid the identification of additional molecular targets. Moreover, progress in the early diagnosis of OA will enable administration of gene therapeutics at a stage when they are most likely to be successful. Finally, important issues with regard to financing and regulation are discussed. Top line data from two pivotal Phase III clinical trials are expected to be released this year. The findings from these trials will exercise considerable influence on the future development of the field.

Community engagement and trust are vital to medical education and improving health. The experience of national level, extramural seed grant programs to support faculty in both education and health care delivery with the focus of building trust has not been previously reported. The authors reviewed a total of four cohorts of the “Building Trust: Advancing Health Equity” grant recipients supported by the ABIM Foundation and other funding partners. Over $1.5 million was awarded from 2021 to 2024, which represented 92 grants ranging from $2,500–$40,000 each. The geographic regions receiving the most grants and funding were the middle and south Atlantic, and the Pacific regions in the United States. University programs received more grants and funding than affiliated or community programs. The majority of project types primarily focused on developing educational curricula for internal medicine residents (68%) and engaging with local communities (19%). The approach to implementation often included service-learning and community experts (29%), quality improvement and patient care delivery methods (22%), and traditional curriculum development (21%). The most common education topic themes were health equity and social determinants of health (43%) and addressing bias in care delivery and learning environment (20%). The Building Trust grant program reflects the unique needs of local communities and supported innovative practices and educational techniques. Implementation support was provided to grantees through shared experiences and mentorship. Evaluating outcomes upon completion of grants will be an imperative next step. This serves as a model that can inform institutions and other professional organizations approaching implementation of seed grant funding.