Citations

Abstract Alzheimer’s Disease (AD) 1 is a progressive neurodegenerative disorder with limited therapeutic options, driving interest in drug repurposing to accelerate treatment discovery. Drug repurposing has emerged as a promising strategy to accelerate therapeutic discovery by repositioning existing drugs for new clinical indications. Recent computational repurposing approaches, including knowledge graph reasoning, transcriptomic signature analysis, and integrative literature mining, have demonstrated strong predictive capabilities 2 . However, these methods often yield divergent drug rankings, which makes it difficult to decide which candidates to advance for experimental follow-up and results in substantial gaps between computational predictions and feasible in vivo validation 2 .To bridge this computational-experimental gap, we proposed an advanced prioritization framework leveraging large language models (LLMs). Our method systematically evaluated three state-of-the-art (SOTA) and representative computational methods (TxGNN 3 , Composition-based Graph Convolutional Network (CompGCN) 4 , and a regularized logistic regression (RLR) 5 , to analyze both their predictive performance and pharmaceutical class distributions. By integrating the strengths and divergences of these models, we generated a unified, streamlined list of 90 candidate drugs for further prioritization. We then utilized an LLM-based agent to perform evidence synthesis from biomedical literature abstracts for each candidate. This process mimics expert manual curation but significantly reduces human effort and time by efficiently distilling vast textual data into actionable insights. Applying consistent and transparent selection criteria, we obtained a refined and prioritized list of drug candidates suitable for subsequent in vivo experimental validation. The robustness and clinical relevance of our framework were validated using real-world data from Alzheimer’s patient cohorts, clinical trial registries, and expert pharmacological reviews. This comprehensive validation confirmed that our LLM-driven approach enhances efficiency, consistency, scalability, and generalizability. By integrating computational predictions with scalable evidence synthesis and multifaceted validation, our framework facilitated rapid and informed prioritization of repurposed drugs. Our framework can potentially accelerate the translational pathway toward viable AD therapeutics. Moreover, the versatility of our framework can also be applied to drug repurposing efforts for other diseases beyond AD.

Treatment-induced ovarian function loss is a significant concern for many young patients with breast cancer. Accurately predicting this risk is crucial for counselling young patients and informing their fertility-related decision-making. However, current risk prediction models for treatment-related ovarian function loss have limitations. To provide a broader representation of patient cohorts and improve feature selection, we combined retrospective data from six datasets within the FoRECAsT (Infertility after Cancer Predictor) databank, including 2679 pre-menopausal women diagnosed with breast cancer. This combined dataset presented notable missingness, prompting us to employ cross imputation using the k-nearest neighbours (KNN) machine learning (ML) algorithm. Employing Lasso regression, we developed an ML model to forecast the risk of treatment-related amenorrhea as a surrogate marker of ovarian function loss at 12 months after starting chemotherapy. Our model identified 20 variables significantly associated with risk of developing amenorrhea. Internal validation resulted in an area under the receiver operating characteristic curve (AUC) of 0.820 (95% CI: 0.817–0.823), while external validation with another dataset demonstrated an AUC of 0.743 (95% CI: 0.666–0.818). A cutoff of 0.20 was chosen to achieve higher sensitivity in validation, as false negatives—patients incorrectly classified as likely to regain menses—could miss timely opportunities for fertility preservation if desired. At this threshold, internal validation yielded sensitivity and precision rates of 91.3% and 61.7%, respectively, while external validation showed 92.9% and 60.0%. Leveraging ML methodologies, we not only devised a model for personalised risk prediction of amenorrhea, demonstrating substantial enhancements over existing models but also showcased a robust framework for maximally harnessing available data sources.

Abstract Transcription factors and their cofactors are major and selective nononcogene dependencies in multiple myeloma cells. By performing a gain-of-function perturbation screen in human multiple myeloma cell lines, we identified the inhibitor of DNA binding (ID) genes as putative suppressors of multiple myeloma cell fitness. Among them, ID2 was found to be downregulated in multiple myeloma patient cells and acted as a tumor suppressor by directly binding and repressing the basic helix-loop-helix factor TCF3, also known as E2A. Lower ID2 expression in multiple myeloma cells conferred a proliferative advantage by increasing TCF3 activity, leading to a dependency on this transcription factor. In contrast, ID2 overexpression reduced TCF3 binding to DNA, which resulted in cell-cycle arrest and a halt in multiple myeloma cell proliferation. The myeloma bone marrow milieu supported this process by further decreasing the expression of ID2 and enhancing TCF3 activity, partly via IL6, revealing a mechanism by which the tumor microenvironment affects multiple myeloma cell behavior. Significance: Multiple myeloma cells exploit the oncogenic and proliferative potential of TCF3 by downregulating the transcriptional regulator ID2, a process facilitated by the bone marrow microenvironment.

Abstract Objective This study reports development and validation of a deep learning seizure detection algorithm for two‐channel subscalp electroencephalographic (EEG) recordings. Ultra‐long‐term monitoring of people with epilepsy may produce new insights into the timing and pattern of their seizures and may pave the way for novel therapeutic options. Although EEG is the accepted standard for epilepsy monitoring, ultra‐long‐term EEG recordings generate a massive quantity of data and are not reasonably reviewable in full by human readers. Methods The convolutional neural network‐bidirectional long short‐term memory (CNN‐BiLSTM) hybrid algorithm uses nine layers, operating on channel spectrograms 5 min in length with 50% overlap. Retrospective subscalp EEG data from 16 patients recorded at three centers were available for algorithm development and testing. EEG was recorded for a median of 63 days (range = 9–508), and a median of 17 seizures (range = 0–96) were recorded. Training data were augmented with scalp EEG seizures, and performance was benchmarked against a conventional spectral power classifier algorithm. We also evaluated an intrapatient training–testing approach where the algorithm was trained on the first 45% of data. Results The CNN‐BiLSTM trained on scalp EEG and subscalp EEG achieved an area under the receiver operating characteristic curve (AUROC) of .98 and an area under the precision–recall curve (AUPRC) of .50, which correspond to 94% sensitivity with 1.11 false detections per day. The same model trained on iEEG achieved only AUROC = .94 and AUPRC = .36. The conventional spectral band power detector achieved AUROC = .93 and AUPRC = .38. The CNN‐BiLSTM detector trained on the earliest half of the subscalp EEG data achieved AUROC = .93 and AUPRC = .37, corresponding to 87% sensitivity and 5.9 false detections per day. Significance High sensitivity and specificity are possible in automated seizure detection in two‐channel subscalp EEG data using a CNN‐BiLSTM framework. Performance of the detector is superior using subcutaneous EEG data for training rather than intracranial EEG, but addition of scalp EEG seizures for training was helpful.

Abstract Background Acupuncture, a key modality in traditional Chinese medicine, is gaining global recognition as a complementary therapy and a subject of increasing scientific interest. However, fragmented and unstructured acupuncture knowledge spread across diverse sources poses challenges for semantic retrieval, reasoning, and in-depth analysis. To address this gap, we developed AcuKG, a comprehensive knowledge graph that systematically organizes acupuncture-related knowledge to support sharing, discovery, and artificial intelligence–driven innovation in the field. Methods AcuKG integrates data from multiple sources, including online resources, guidelines, PubMed literature, ClinicalTrials.gov, and multiple ontologies (SNOMED CT, UBERON, and MeSH). We employed entity recognition, relation extraction, and ontology mapping to establish AcuKG, with human-in-the-loop to ensure data quality. Two cases evaluated AcuKG’s usability: (1) how AcuKG advances acupuncture research for obesity and (2) how AcuKG enhances large language model (LLM) application on acupuncture question-answering. Results AcuKG comprises 1839 entities and 11 527 relations, mapped to 1836 standard concepts in 3 ontologies. Two use cases demonstrated AcuKG’s effectiveness and potential in advancing acupuncture research and supporting LLM applications. In the obesity use case, AcuKG identified highly relevant acupoints (eg, ST25, ST36) and uncovered novel research insights based on evidence from clinical trials and literature. When applied to LLMs in answering acupuncture-related questions, integrating AcuKG with GPT-4o and LLaMA 3 significantly improved accuracy (GPT-4o: 46% → 54%, P = .03; LLaMA 3: 17% → 28%, P = .01). Conclusion AcuKG is an open dataset that provides a structured and computational framework for acupuncture applications, bridging traditional practices with acupuncture research and cutting-edge LLM technologies.

Objective: To evaluate the feasibility and effectiveness of wrist-worn wearable devices combined with machine learning (ML) approaches for detecting a diverse array of seizure types beyond generalized tonic–clonic (GTC), including focal, generalized, and subclinical seizures. Materials and Methods: Twenty-eight patients undergoing inpatient video-EEG monitoring at Mayo Clinic were concurrently monitored using Empatica E4 wrist-worn devices. These devices captured accelerometry, blood volume pulse, electrodermal activity, skin temperature, and heart rate. Seizures were annotated by neurologists. The data were preprocessed to experiment with various segment lengths (10 s and 60 s) and multiple feature sets. Three ML strategies, XGBoost, deep learning models (LSTM, CNN, Transformer), and ROCKET, were evaluated using leave-one-patient-out cross-validation. Performance was assessed using area under the receiver operating characteristic curve (AUROC), seizure-wise recall (SW-Recall), and false alarms per hour (FA/h). Results: Detection performance varied by seizure type and model. GTC seizures were detected most reliably (AUROC = 0.86, SW-Recall = 0.81, FA/h = 3.03). Hyperkinetic and tonic seizures showed high SW-Recall but also high FA/h. Subclinical and aware-dyscognitive seizures exhibited the lowest SW-Recall and highest FA/h. MultiROCKET and XGBoost performed best overall, though no single model was optimal for all seizure types. Longer segments (60 s) generally reduced FA/h. Feature set effectiveness varied, with multi-biosignal sets improving performance across seizure types. Conclusions: Wrist-worn wearables combined with ML can extend seizure detection beyond GTC seizures, though performance remains limited for non-motor types. Optimizing model selection, feature sets, and segment lengths, and minimizing false alarms, are key to clinical utility and real-world adoption.

Background: Artificial Intelligence (AI)-driven Decision Support Systems (DSSs) promise improvements in diagnostic accuracy and clinical workflow efficiency, but their adoption is hindered by inadequate confidence calibration, limited transparency, and poor alignment with real-world decision processes, which limit clinician trust and lead to high override rates. Methods: We developed and validated a dynamic scoring framework to enhance trust in AI-generated diagnoses by integrating AI confidence scores, semantic similarity measures, and transparency weighting into the override decision process using 6689 cardiovascular cases from the MIMIC-III dataset. Override thresholds were calibrated and validated across varying transparency and confidence levels, with override rate as the primary acceptance measure. Results: The implementation of this framework reduced the override rate to 33.29%, with high-confidence predictions (90–99%) overridden at a rate of only 1.7%, and low-confidence predictions (70–79%) at a rate of 99.3%. Minimal transparency diagnoses had a 73.9% override rate compared to 49.3% for moderate transparency. Statistical analyses confirmed significant associations between confidence, transparency, and override rates (p < 0.001). Conclusions: These findings suggest that enhanced transparency and confidence calibration can substantially reduce override rates and promote clinician acceptance of AI diagnostics. Future work should focus on clinical validation to optimize patient safety, diagnostic accuracy, and efficiency.

Abstract With the increasing volume of biomedical experimental data, standardizing, sharing, and integrating heterogeneous experimental data across domains has become a major challenge. To address this challenge, we have developed an ontology-supported Study-Experiment-Assay (SEA) common data model (CDM), which includes 10 core and 3 auxiliary classes based on object-oriented modeling. SEA CDM uses interoperable ontologies for data standardization and knowledge inference. Building on the SEA CDM, we developed the Ontology-based SEA Network (OSEAN) relational database and knowledge graph, along with a set of ETL (Extract, Transform, Load) and query tools, and further applied them to represent 1,278 immune studies with over two million samples from three resources: VIGET, ImmPort, and CELLxGENE. Using simple, robust queries and analyses, our research identified multiple scientific insights into sex-specific immune responses, such as neutrophil degranulation and TNF binding to physiological receptors, following live attenuated and trivalent inactivated influenza vaccination. The novel SEA CDM system lays a foundation for establishing an integrative biodata ecosystem across biological and biomedical domains.

Retrieval-Augmented Generation (RAG) offers a promising strategy to harness large language models (LLMs) for delivering up-to-date, accurate clinical guidance while reducing physicians’ cognitive burden, yet its effectiveness hinges on query clarity and structure. We propose an adaptive Self-Query Retrieval (SQR) framework that integrates three refinement modules—PICOT (Population, Intervention, Comparison, Outcome, Time), SPICE (Setting, Population, Intervention, Comparison, Evaluation), and Iterative Query Refinement (IQR)—to automatically restructure and iteratively enhance clinical questions until they meet predefined retrieval-quality thresholds. Implemented on Gemini-1.0 Pro, we benchmarked SQR using thirty postoperative rhinoplasty queries, evaluating responses for accuracy and relevance on a three-point Likert scale and for retrieval quality via precision, recall, and F1 score; statistical significance was assessed by one-way ANOVA with Tukey post-hoc testing. The full SQR pipeline achieved 87% accuracy (Likert 2.4 ± 0.7) and 100% relevance (Likert 3.0 ± 0.0), significantly outperforming a non-refined RAG baseline (50% accuracy, 80% relevance; p < 0.01 and p = 0.03). Precision, recall, and F1 rose from 0.17, 0.39 and 0.24 to 0.53, 1.00, and 0.70, respectively, while PICOT-only and SPICE-only variants yielded intermediate improvements. These findings demonstrate that automated structuring and iterative enhancement of queries via SQR substantially elevate LLM-based clinical decision support, and its model-agnostic architecture enables rapid adaptation across specialties, data sources, and LLM platforms.

Abstract Background/ Objectives: Agentic AI represents a promising evolution of AI technology applied to healthcare, with systems increasingly capable of operating autonomously to achieve defined clinical goals. However, the literature lacks conceptual clarity between “AI agents” and “agentic AI”, and few studies have rigorously explored their clinical applications. Therefore, this study aims to conduct a novel systematic review addressing this gap by examining agentic AI systems in healthcare settings, characterizing their applications, features, outcomes, and limitations, and clarifying the conceptual distinctions between AI agents and agentic AI systems using predefined and objective criteria. Methods A comprehensive search was conducted across PubMed, Embase, Cochrane, Scopus, and Google Scholar on April 6th, 2025. Studies were included if they involved AI systems in healthcare settings that demonstrated the following agentic features: autonomous operation, goal-directed behavior, and initiating action. Data on the clinical tasks achieved by the agents, key findings, features, and limitations were collected from the included studies. Screening and extraction followed PRISMA guidelines, with Risk of Bias assessed using ROBINS-I and Cochrane's Risk of Bias tools. Results Of 984 retrieved records, seven studies met the inclusion criteria, spanning domains such as emergency medicine, oncology, radiology, and rehabilitation. Multi-agent architecture was frequently used to decompose and coordinate complex workflows. Among the included studies, the AIs showed high accuracy in diagnosing cancer patients, conducting treatment plans, sending alerts, coaching messages, analysing image data, and adapting to challenging experimental scenarios. While demonstrating potential for improved efficiency, task accuracy, and patient engagement, significant limitations were noted: narrow task scope, lack of physical agency, limited clinical validation, and barriers to integration into real-world healthcare systems. Only one system had been deployed in a patient-facing trial setting. Conclusion The current literature suggests an emerging role and application of Agentic AI, holding promise with the potential to revolutionize diagnostics, triage, treatment planning, and patient management. However, real-world implementation and evaluations in the literature are limited. Future research must address critical validation, regulation, ethics, and clinical integration challenges to realize their full potential. Clear operational definitions and frameworks for evaluating agency are essential to support safe and effective deployment of these systems.