OBJECTIVE: To develop a user-friendly visualization application for the automatic annotation of Human Phenotype Ontology (HPO) terms based on large language models and retrieval-augmented generation (RAG) technology, and to validate its performance in an authoritative case dataset. METHODS: By integrating the domestic open-source large language model DeepSeek-V3 with RAG technology, an interactive web application was deployed on the Streamlit cloud platform. Using only the latest official HPO dataset as the data source, the lightweight sentence-embedding model BAAI/bge-small-en-v1.5 was employed to construct a FAISS vector index. During the online phase, a four-step closed-loop process is automatically completed: multilingual translation, phenotype phrase extraction, RAG candidate retrieval, term mapping, and official database validation. 121 English case reports publicly released by BMJ Case Reports and Oxford Medical Case Reports (with a gold-standard HPO set of 1 794 terms) were selected for application validation. Precision, recall, and F1 score were calculated and compared horizontally with traditional dictionary tools, standalone large language models, and the similar application "RAG-HPO". Finally, replace the model with the more advanced ChatGPT-5 and evaluate its performance on the newly extracted dataset. RESULTS: An HPO term automatic annotation visualization application named PhenoRAG, based on large language models and RAG technology, was successfully developed. Users can access it directly via a web link. Across the 112 cases, a total of 2 150 HPO terms were generated; 2,064 (96.0%) were fully validated by the official database, with a hallucination rate of 1.3% and an HPO ID-name mismatch rate of 2.7%. After deduplication, 1,906 terms remained for testing. The overall precision was 63.65%, recall was 67.34%, and F1 was 65.44%, significantly outperforming traditional annotation tools (F1: 0.45-0.49, P < 0.001). Although PhenoRAG's F1 was lower than that of RAG-HPO (F1 = 0.78, P < 0.001), which relies on a manually constructed synonym database of 54 000 entries plus the HPO dataset, it requires no additional dictionary maintenance and can be used without any background in computer programming. Moreover, after switching to the GPT-5 model, PhenoRAG exhibited no hallucination rate on the new dataset, and its F1 score significantly increased (P = 0.038). CONCLUSION Without constructing a synonym database, the PhenoRAG achieved high-accuracy automatic mapping from clinical text to standard HPO terms. It features a low usage threshold, free access, and a Chinese-language interface, and can directly serve rare disease diagnosis, genetic counseling, and research scenarios in China and worldwide, warranting further clinical promotion and multicenter validation.
Humans ; Phenotype ; Biological Ontologies ; Language ; Software ; Large Language Models

Academic publishing is at a critical juncture. The challenges faced by the academics are mired in controversy. Among theseare three hotly debated concerns. First is the issue of whether technological innovations such as artificial intelligence (AI)improves research efficiency or if its use sacrifices research integrity.Another is the controversy between paywall publishingand open access. Lastly, adapting an appropriate business model for sustainability is a contentious issue and the choice betweena commercial or a university-based publishing platform is a difficult one.

Traditional models of scientific investigation relied on tedious intellectual calisthenics in all aspects of research —identifying research gaps, reviewing of published literature, devising valid methodology, collecting data, analysing results, and,finally, drawing conclusions. With the advent of powerful tools employing artificial intelligence, these heavy tasks are efficientlycarried out. The dilemma lies in determining which parts of the work can be attributed to the authors and which are ascribedto the output of large language models (LLMs) and other automated assistance employed.Despite requiring adequate vettingby experts of these AI-aided output, many in the scientific community still question these methods. Can research employingAI be considered honest work? Will full disclosure answer doubts as to the integrity of the scientific work?

Indeed, LLMs just gather information that is already out there, albeit more efficiently. After all, science progresses bystanding on the shoulder of giants. AI makes such work comprehensive and efficient. Standing on those proverbial shoulders,however, require access to prior work, hence our next challenge in academic publishing--open access versus paid access.Paywalls limit the benefits of valuable research to institutions and universities with the capacity to pay. Excluded from these arethose from low resourced countries, with nations from the global south being affected disproportionately. Additionally, whilenumerous authors appreciate the features of open access as it improves their impact and visibility, many feel unduly burdenedsince the cost of publishing in this format is passed on to them.

This brings us to our third issue: who bears the cost of academic publishing? Indeed, it is a lucrative industry, generatingan annual revenue of US$19 billion and an estimated 40 percent profit margin. Many, however, find fault in this businessmodel as concerns about the profit motives of the commercial publishers far overshadow their sustainability goals.

How do we navigate this landscape of controversies? We, at the ACTA, as part of the community of scholars, would needto clarify our mission. Our goals for this publication should be consistent with our values. These values, such as scientific rigor,integrity, and accountability, should be reflected in our policies. We should be cognizant of the role we play in national scientificdiscourse while we endeavor to make an impact in the global scene. We are accountable to our stakeholders — nurturingearly career scholars, supplying evidence to health policymakers, and being accountable to those who provide resources tosustain us. This stewardship is essential so that ACTA will stand shoulder to shoulder with the giants on which science buildsupon to benefit future generations.

Artificial Intelligence ; Commerce ; Costs And Cost Analysis ; Disclosure ; Drawing ; Efficiency ; Family Characteristics ; Forecasting ; Goals ; Gymnastics ; Health ; Health Resources ; Industry ; Intelligence ; Inventions ; Language ; Literature ; Methods ; Play And Playthings ; Policy ; Publications ; Publishing ; Research ; Residence Characteristics ; Role ; Science ; Shoulder ; Social Responsibility ; Universities ; Ursidae ; Volition ; Work ; World Health Organization

OBJECTIVE: To explore the clinical features and genetic etiology of a child with a SETD1B gene variant causing seizures and language delay. METHODS: A child with a SETD1B gene variant admitted to the Department of Pediatric Neurology at the Third Affiliated Hospital of Zhengzhou University in September 2022 was selected as the study subject. Clinical data of the child were collected, and peripheral blood samples from the child and her parents were obtained. Whole exome sequencing (WES) was performed for genetic testing, and Sanger sequencing was used for familial validation of the candidate variant. Using "SETD1B" and "epilepsy" as the Chinese and English keywords, relevant cases were retrieved from databases including CNKI, Wanfang Data, OMIM and PubMed, with the search period spanning from database inception to June 2024. RESULTS: The child was a 6-year-old female presenting with myoclonic seizures accompanied by global developmental delay. WES and Sanger sequencing revealed that the child has carried a de novo SETD1B gene variant, namely c.5582G>A (p.Cys1961Tyr). According to the American College of Medical Genetics and Genomics (ACMG) guidelines for sequence variant interpretation, this variant was classified as likely pathogenic (PS2+PM2_Supporting+PP2+PP3). The child was not controlled with effective doses of valproate, levetiracetam, or clonazepam but was successfully managed with low-dose lamotrigine. Follow-up electroencephalography showed normal results, and developmental progress gradually improved. A total of 37 epilepsy cases with SETD1B gene variants were reported across six studies. The predominant seizure types included absence seizures and myoclonic absence seizures, accompanied by delayed language development. The response to pharmacological treatment was generally poor, with no significant difference in incidence between males and females. CONCLUSION SETD1B gene variants may cause neurological disorders with drug-resistant epilepsy and severe clinical manifestations. Lamotrigine is effective in controlling the epileptic seizures.
Humans ; Female ; Child ; Epilepsy/genetics* ; Language Development Disorders/genetics* ; Histone-Lysine N-Methyltransferase/genetics* ; Exome Sequencing ; Male

Large language models (LLMs) such as ChatGPT, Claude, Llama, and Qwen are emerging as transformative technologies for the diagnosis and treatment of various diseases. With their exceptional long-context reasoning capabilities, LLMs are proficient in clinically relevant tasks, particularly in medical text analysis and interactive dialogue. They can enhance diagnostic accuracy by processing vast amounts of patient data and medical literature and have demonstrated their utility in diagnosing common diseases and facilitating the identification of rare diseases by recognizing subtle patterns in symptoms and test results. Building on their image-recognition abilities, multimodal LLMs (MLLMs) show promising potential for diagnosis based on radiography, chest computed tomography (CT), electrocardiography (ECG), and common pathological images. These models can also assist in treatment planning by suggesting evidence-based interventions and improving clinical decision support systems through integrated analysis of patient records. Despite these promising developments, significant challenges persist regarding the use of LLMs in medicine, including concerns regarding algorithmic bias, the potential for hallucinations, and the need for rigorous clinical validation. Ethical considerations also underscore the importance of maintaining the function of supervision in clinical practice. This paper highlights the rapid advancements in research on the diagnostic and therapeutic applications of LLMs across different medical disciplines and emphasizes the importance of policymaking, ethical supervision, and multidisciplinary collaboration in promoting more effective and safer clinical applications of LLMs. Future directions include the integration of proprietary clinical knowledge, the investigation of open-source and customized models, and the evaluation of real-time effects in clinical diagnosis and treatment practices.
Humans ; Large Language Models ; Tomography, X-Ray Computed

Humans ; Language Development Disorders/diagnosis* ; Early Diagnosis ; Child

Automatic classification of medical questions is of great significance in improving the quality and efficiency of online medical services, and belongs to the task of intent recognition. Joint entity recognition and intent recognition perform better than single task models. Currently, most publicly available medical text intent recognition datasets lack entity annotation, and manual annotation of these entities requires a lot of time and manpower. To solve this problem, this paper proposes a medical text classification model, bidirectional encoder representation based on transformer-recurrent convolutional neural network-entity-label-semantics (BRELS), which integrates medical entity label semantics. This model firstly utilizes an adaptive fusion mechanism to absorb prior knowledge of medical entity labels, achieving local feature enhancement. Then in global feature extraction, a lightweight recurrent convolutional neural network (LRCNN) is used to suppress parameter growth while preserving the original semantics of the text. The ablation and comparison experiments are conducted on three public medical text intent recognition datasets to validate the performance of the model. The results show that F1 score reaches 87.34%, 81.71%, and 77.74% on each dataset, respectively. The results show that the BRELS model can effectively identify and understand medical terminology, thereby effectively identifying users' intentions, which can improve the quality and efficiency of online medical services.
Semantics ; Neural Networks, Computer ; Humans ; Natural Language Processing

OBJECTIVES: To summarize the clinical manifestations and genetic characteristics of creatine deficiency syndrome (CDS) caused by GAMT gene mutations. METHODS: A retrospective analysis was conducted on the clinical and genetic data of two children diagnosed with GAMT deficiency-type CDS at the Children's Medical Center of Xiangya Hospital, Central South University, from December 2020 to December 2024. RESULTS: The two patients presented with symptoms in infancy, and both had compound heterozygous mutations in the GAMT gene. Case 1 exhibited seizures and intellectual disability, while Case 2 had intellectual disability and attention-deficit hyperactivity disorder. Magnetic resonance spectroscopy of cranial MRI in both patients indicated reduced creatine peaks. After creatine treatment, seizures in Case 1 were controlled, but both patients continued to experience intellectual disabilities and behavioral issues. As of December 2024, a total of 21 cases have been reported in China (including this study), and 115 cases have been reported abroad. All patients exhibited developmental delay or intellectual disabilities, with 66.9% (91/136) experiencing seizures, 33.8% (46/136) presenting with motor disorders, and 36.8% (50/136) having behavioral problems. Seventy-five percent (102/136) of patients received creatine treatment, leading to significant improvements in seizures and motor disorders, although cognitive improvement was not substantial. CONCLUSIONS GAMT deficiency-type CDS is rare and presents with nonspecific clinical features. Timely diagnosis facilitates targeted treatment, which can partially improve prognosis.
Child ; Female ; Humans ; Male ; Creatine/deficiency* ; Guanidinoacetate N-Methyltransferase/deficiency* ; Intellectual Disability/genetics* ; Mutation ; Retrospective Studies ; Rhabdomyolysis/genetics* ; Language Development Disorders ; Movement Disorders/congenital*

OBJECTIVES: Accurate identification of risk factors for precocious puberty is essential for clinical diagnosis and management, yet the performance of natural language processing methods applied to unstructured electronic medical record (EMR) data remains to be fully evaluated. This study aims to assess the performance of a prompt engineering method for extracting individual risk factors of precocious puberty from EMRs. METHODS: Based on the capacity and role-insight-statement-personality-experiment (CRISPE) prompt framework, both simple and optimized prompts were designed to guide the large language model GLM-4-9B in extracting 10 types of risk factors for precocious puberty from 653 EMRs. Accuracy, precision, recall, and F1-score were used as evaluation metrics for the information extraction task. RESULTS: Under simple and optimized prompt conditions, the overall accuracy, precision, recall, and F1-score of the model were 84.18%, 98.09%, 81.99%, and 89.32% versus 97.15%, 98.31%, 98.16%, and 98.23%, respectively. The optimized prompts achieved more stable performance across age (<9 years vs ≥9 years) and visit-time (<2023 vs ≥2023) subgroups compared with simple prompts. The accuracy range for extracting each risk factor was 60.03%-97.24%, while with optimized prompts, the range improved to 92.19%-99.85%. The largest performance improvement occurred for "beverage intake" (60.03% vs 92.19%), and the smallest for "maternal age of menarche" (97.24% vs 99.23%). In comparing distributions among simple prompts, optimized prompts, and ground truth, statistically significant differences were observed for snack intake, beverage intake, soy milk intake, honey intake, supplement use, tonic use, sleep quality, and sleeping with the light on (all P<0.001), while exercise (P=0.966) and maternal menarche age (P=0.952) showed no significant differences. CONCLUSIONS Compared with simple prompts, optimized prompts substantially improved the extraction performance of individual risk factors for precocious puberty from EMRs, underscoring the critical role of prompt engineering in enhancing large language model performance.
Humans ; Puberty, Precocious/epidemiology* ; Risk Factors ; Electronic Health Records ; Female ; Child ; Natural Language Processing

OBJECTIVE

To evaluate the reliability and validity of the Filipino version of the Vascular Access Quality of Life (VASQoL) measure.

METHODS

The Vascular Access Quality of Life (VASQoL) measure was translated to the Filipino language and, subsequently back translated to the English version by professors from the university. This translated questionnaire was face validated by a vascular surgeon, both adept in using both Filipino and English language. Pre-test was done on 10 subjects to assess cross cultural applicability of the questionnaire. Reliability was tested on 24 patients who were all diagnosed with Stage V Chronic Kidney Disease (CKD).

RESULTS

A total of 24 patients were recruited in the study with a mean age of 50.6, ranging from 15 to 77 years old. Slightly higher population of female versus male participants. For both test and retest, the internal consistency of the VASQoL was unacceptable with Cronbach’s alpha of 0.5209. Test-retest reliability showed almost perfect agreement except for one item (Q7) with substantial agreement.

CONCLUSION

The evaluation of the Filipino version of the VASQoL measure has revealed shortcomings in terms of reliability, suggesting that it may not be a dependable tool for assessing the experiences of patients with CKD undergoing hemodialysis (HD) in the Filipino population.

Human ; Quality Of Life ; Language ; Patients

OBJECTIVES

This study aimed to describe the demographic profile, intervention sessions, goal attainment scaling (GAS), and health-related quality of life (HRQOL) of autistic children receiving speech and language therapy (SLT) in a higher educational institution in the Philippines.

METHODS

Deidentified data from 18 autistic children aged 4–16 years (mean=8.2; SD=2.9) who received SLT for two months were analyzed. Their demographic profile, intervention sessions, GAS scores, and generic HRQOL scores were documented.

RESULTS

Most participants were school-age children (n=12; 66%) and were boys (n=14; 78%). After two months, the GAS scores of 11 participants (61%) increased by 1–2 points, whereas the scores of the remaining participants decreased (n=6; 33%) or did not change (n=1; 6%). Their mean generic HRQOL scores before and after SLT were 65.6 (SD=15.2) and 61.2 (SD=17.4), respectively.

CONCLUSIONS

While the GAS scores increased for most participants, their generic HRQOL scores did not show clinically significant changes after two months of SLT. This can be attributed to the few therapy sessions and short follow-up period. The findings highlight the need to provide long-term support to SLT services of autistic children in the Philippines to document more desirable quality of life outcomes.

Human ; Quality Of Life ; Autistic Disorder ; Child ; Language Therapy