Leveraging Medical Knowledge Graphs Into Large Language Models for Diagnosis Prediction: Design and Application Study.

Background: Electronic health records (EHRs) and routine documentation practices play a vital role in patients' daily care, providing a holistic record of health, diagnoses, and treatment. However, complex and verbose EHR narratives can overwhelm health care providers, increasing the risk of diagnostic inaccuracies. While large language models (LLMs) have showcased their potential in diverse language tasks, their application in health care must prioritize the minimization of diagnostic errors and the prevention of patient harm.

Alleviating effect of intranasal zinc on symptoms of allergic rhinitis.

Background: There is no information about the clinical implications and kinetics of zinc (Zn) in the nasal cavity, a center of allergic inflammation, and serum in subjects with allergic rhinitis (AR).

Objective: Effects of intranasal Zn on symptoms before and after allergen provocation were investigated in humans and mice with or without AR.

Methods: The first clinical follow-up study for Zn levels in nasal epithelial lining fluid (ELF) and serum was conducted in 57 control subjects and 44 patients with Japanese cedar pollinosis (JCP), a representative seasonal AR, from preseason to season.

Stabilization of RUNX1 Induced by O-GlcNAcylation Promotes PDGF-BB-Mediated Resistance to CDK4/6 Inhibitors in Breast Cancer.

Cyclin-dependent kinases 4 and 6 (CDK4/6) are crucial in regulating cell cycle progression and cancer development. Targeting CDK4/6 has shown considerable promise in treating various cancers, including breast cancer. Despite significant therapeutic efficacy, resistance to CDK4/6 inhibitors (CDK4/6i), such as palbociclib, remains a substantial hurdle in clinical practice.

Uncertainty estimation in diagnosis generation from large language models: next-word probability is not pre-test probability.

Objective: To evaluate large language models (LLMs) for pre-test diagnostic probability estimation and compare their uncertainty estimation performance with a traditional machine learning classifier.

Materials And Methods: We assessed 2 instruction-tuned LLMs, Mistral-7B-Instruct and Llama3-70B-chat-hf, on predicting binary outcomes for Sepsis, Arrhythmia, and Congestive Heart Failure (CHF) using electronic health record (EHR) data from 660 patients. Three uncertainty estimation methods-Verbalized Confidence, Token Logits, and LLM Embedding+XGB-were compared against an eXtreme Gradient Boosting (XGB) classifier trained on raw EHR data.

Lessons learned on information retrieval in electronic health records: a comparison of embedding models and pooling strategies.

Objectives: Applying large language models (LLMs) to the clinical domain is challenging due to the context-heavy nature of processing medical records. Retrieval-augmented generation (RAG) offers a solution by facilitating reasoning over large text sources. However, there are many parameters to optimize in just the retrieval system alone.

Quantitative evaluation of pancreatic neuroendocrine tumors utilizing dual-source CT perfusion imaging.

Objective: We aimed to quantitatively analyze the perfusion characteristics of pancreatic neuroendocrine tumors (pNETs) utilizing dual-source CT imaging.

Methods: Dual-source CT perfusion scans were obtained from patients with pNETs confirmed by surgical or biopsy pathology. Perfusion parameters, including blood flow (BF), blood volume (BV), capillary permeability surface (PS), mean transit time (MTT), contrast transit time to the start (TTS), and contrast transit time to the peak (TTP), were statistically analyzed and compared with nearby healthy tissue.

LCD benchmark: long clinical document benchmark on mortality prediction for language models.

Objectives: The application of natural language processing (NLP) in the clinical domain is important due to the rich unstructured information in clinical documents, which often remains inaccessible in structured data. When applying NLP methods to a certain domain, the role of benchmark datasets is crucial as benchmark datasets not only guide the selection of best-performing models but also enable the assessment of the reliability of the generated outputs. Despite the recent availability of language models capable of longer context, benchmark datasets targeting long clinical document classification tasks are absent.

Association Between Blood-Brain Barrier Disruption and Stroke-Associated Pneumonia in Acute Ischemic Stroke Patients After Endovascular Therapy: A Retrospective Cohort Study.

Background: Stroke, particularly due to large vessel occlusion (LVO), is a major cause of mortality and disability globally. Endovascular therapy (ET) significantly improves outcomes for acute ischemic stroke (AIS) patients, but complications such as stroke-associated pneumonia (SAP) increase mortality and healthcare costs. This study investigates the association between blood-brain barrier (BBB) disruption and the increased risk of SAP and explores the relationship between BBB disruption and medium-term functional outcomes.

Prompt engineering with a large language model to assist providers in responding to patient inquiries: a real-time implementation in the electronic health record.

Background: Large language models (LLMs) can assist providers in drafting responses to patient inquiries. We examined a prompt engineering strategy to draft responses for providers in the electronic health record. The aim was to evaluate the change in usability after prompt engineering.

On the role of the UMLS in supporting diagnosis generation proposed by Large Language Models.

Objective: Traditional knowledge-based and machine learning diagnostic decision support systems have benefited from integrating the medical domain knowledge encoded in the Unified Medical Language System (UMLS). The emergence of Large Language Models (LLMs) to supplant traditional systems poses questions of the quality and extent of the medical knowledge in the models' internal knowledge representations and the need for external knowledge sources. The objective of this study is three-fold: to probe the diagnosis-related medical knowledge of popular LLMs, to examine the benefit of providing the UMLS knowledge to LLMs (grounding the diagnosis predictions), and to evaluate the correlations between human judgments and the UMLS-based metrics for generations by LLMs.

Sequential severe immune-related adverse events induced by PD-1 inhibitor: a case report and literature review.

In a variety of cancers, immune checkpoint inhibitors (ICIs) have demonstrated substantial survival advantages. Nevertheless, the widespread use of ICIs in the clinic has resulted in a growing interest in immune-related adverse events (irAEs) and their treatment methods. This paper reports a case in which a patient with three sequential severe irAEs was successfully treated.

The Enormous Potential of Sodium/Potassium-Ion Batteries as the Mainstream Energy Storage Technology for Large-Scale Commercial Applications.

Cost-effectiveness plays a decisive role in sustainable operating of rechargeable batteries. As such, the low cost-consumption of sodium-ion batteries (SIBs) and potassium-ion batteries (PIBs) provides a promising direction for "how do SIBs/PIBs replace Li-ion batteries (LIBs) counterparts" based on their resource abundance and advanced electrochemical performance. To rationalize the SIBs/PIBs technologies as alternatives to LIBs from the unit energy cost perspective, this review gives the specific criteria for their energy density at possible electrode-price grades and various battery-longevity levels.

[Therapeutic effect of calcaneal beak-like fracture secondary to calcaneal osteomyelitis caused by diabetic foot].

Objective: To explore clinical effect of vancomycin calcium sulfate combined with internal fixation on calcaneal beak-like fracture secondary to calcaneal osteomyelitis caused by diabetic foot.

Methods: From April 2018 to October 2021, a retrospective analysis was performed on 5 patients with calcaneal bone osteomyelitis secondary to diabetic foot, including 2 males and 3 females, aged from 48 to 60 years old;diabetes course ranged from 5 to 13 years;the courses of diabetic foot disease ranged from 18 to 52 days;5 patients were grade Ⅲ according to Wagner classification. All patients were treated with debridement, vancomycin bone cement implantation, negative pressure aspiration at stageⅠ, vancomycin calcium sulfate and internal fixation at stageⅡfor calcaneal beak-like fracture.

LCD Benchmark: Long Clinical Document Benchmark on Mortality Prediction for Language Models.

Objective: The application of Natural Language Processing (NLP) in the clinical domain is important due to the rich unstructured information in clinical documents, which often remains inaccessible in structured data. When applying NLP methods to a certain domain, the role of benchmark datasets is crucial as benchmark datasets not only guide the selection of best-performing models but also enable the assessment of the reliability of the generated outputs. Despite the recent availability of language models (LMs) capable of longer context, benchmark datasets targeting long clinical document classification tasks are absent.

Development of a Human Evaluation Framework and Correlation with Automated Metrics for Natural Language Generation of Medical Diagnoses.

In the evolving landscape of clinical Natural Language Generation (NLG), assessing abstractive text quality remains challenging, as existing methods often overlook generative task complexities. This work aimed to examine the current state of automated evaluation metrics in NLG in healthcare. To have a robust and well-validated baseline with which to examine the alignment of these metrics, we created a comprehensive human evaluation framework.

A T-Cell Inspired Sonoporation System Enhances Low-Dose X-Ray-Mediated Pyroptosis and Radioimmunotherapy Efficacy by Restoring Gasdermin-E Expression.

Genome editing has the potential to improve the unsatisfactory therapeutic effect of antitumor immunotherapy. However, the cell plasma membrane prevents the entry of almost all free genome-manipulation agents. Therefore, a system can be spatiotemporally controlled and can instantly open the cellular membrane to allow the entry of genome-editing agents into target cells is needed.

Transfer-Free Analog and Digital Flexible Memristors Based on Boron Nitride Films.

The traditional von Neumann architecture of computers, constrained by the inherent separation of processing and memory units, faces challenges, for instance, memory wall issue. Neuromorphic computing and in-memory computing offer promising paradigms to overcome the limitations of additional data movement and to enhance computational efficiency. In this work, transfer-free flexible memristors based on hexagonal boron nitride films were proposed for analog neuromorphic and digital memcomputing.

Predicting 3-month Functional Outcome After Endovascular Thrombectomy in Patients with Anterior Circulation Occlusion with an Arterial Transit Artifact Grading System.

Purpose: The objective of this study was to evaluate the relationship between arterial transit artifact (ATA), arterial spin labeling (ASL) perfusion imaging, and the outcome of patients with acute ischemic stroke (AIS) due to occlusion of large vessels in anterior circulation after endovascular thrombectomy (EVT).

Methods: Patients with anterior circulation occlusion treated with EVT between October 2017 and December 2021 were enrolled in this retrospective study, and ATA was quantified by a 4-point scale. A favorable outcome was defined by modified Rankin Scale (mRS) scores of 0-2 at 3 months.

AHSA1 Regulates Hepatocellular Carcinoma Progression via the TGF-β/Akt-Cyclin D1/CDK6 Pathway.

Background: Activator of heat shock protein 90 (HSP90) ATPase Activity 1 (AHSA1) regulates proliferation, apoptosis, migration, and invasion of osteosarcoma and hepatocellular carcinoma (HCC). However, the novel mechanism of AHSA1 in the tumor biology of hepatocellular carcinoma (HCC) remains unclear.

Methods: We analyzed AHSA1 expression in 85 pairs of clinical samples of HCC and the Cancer Genome Atlas database.

Improving the Transferability of Clinical Note Section Classification Models with BERT and Large Language Model Ensembles.

Text in electronic health records is organized into sections, and classifying those sections into section categories is useful for downstream tasks. In this work, we attempt to improve the transferability of section classification models by combining the dataset-specific knowledge in supervised learning models with the world knowledge inside large language models (LLMs). Surprisingly, we find that zero-shot LLMs out-perform supervised BERT-based models applied to out-of-domain data.

Improving model transferability for clinical note section classification models using continued pretraining.

Objective: The classification of clinical note sections is a critical step before doing more fine-grained natural language processing tasks such as social determinants of health extraction and temporal information extraction. Often, clinical note section classification models that achieve high accuracy for 1 institution experience a large drop of accuracy when transferred to another institution. The objective of this study is to develop methods that classify clinical note sections under the SOAP ("Subjective," "Object," "Assessment," and "Plan") framework with improved transferability.

Factors Affecting English Language Teachers' Behavioral Intentions to Teach Online under the Pandemic Normalization of COVID-19 in China.

The unexpected outbreak of COVID-19 pandemic forced most teachers across the globe to switch their teaching from traditional face-to-face to online without having received adequate preparation and knowledge related to online teaching. To better comprehend teachers' willingness to conduct emergency remote teaching during the worldwide crisis, the current study was designed to examine teachers' intentions and, in particular, the factors affecting their behavioral intentions to teach online, by contextualizing the research in the English language teaching settings in China. The research model was developed based on an extended technology acceptance model (TAM) by adding subjective norm, self-efficacy, technological complexity, and facilitating conditions into the original TAM.

Overview of the Problem List Summarization (ProbSum) 2023 Shared Task on Summarizing Patients' Active Diagnoses and Problems from Electronic Health Record Progress Notes.

The BioNLP Workshop 2023 initiated the launch of a shared task on Problem List Summarization (ProbSum) in January 2023. The aim of this shared task is to attract future research efforts in building NLP models for real-world diagnostic decision support applications, where a system generating relevant and accurate diagnoses will augment the healthcare providers' decision-making process and improve the quality of care for patients. The goal for participants is to develop models that generated a list of diagnoses and problems using input from the daily care notes collected from the hospitalization of critically ill patients.