Driven by the converging forces of metaverse medicine and the new quality productive forces in healthcare, respiratory medicine education must evolve from the traditional model of ‘knowledge delivery and guideline recitation’ toward a forward-looking ‘AI+ Respiratory New-Quality Productive Center’- a pioneering hub that integrates pedagogical prototyping with the incubation of interdisciplinary talent. Contemporary respiratory physicians, beyond mastering core clinical competencies, must also cultivate digital health literacy: the ability to critically understand, evaluate, and appropriately apply emerging technologies such as artificial intelligence (AI), the Internet of Things (IoT), digital twins, and extended reality (XR). Within human - AI collaborative multidisciplinary teams (MDTs) and frameworks augmented by ‘digital expert avatars,’ they should be empowered to make safe clinical decisions, conduct risk stratification, and simulate therapeutic strategies. International evidence-based medical education research indicates that AI has already been piloted in areas such as radiological interpretation, clinical skills assessment, adaptive learning, and real-time feedback. However, these efforts largely remain fragmented ‘point innovations’, lacking an integrated curriculum that embeds cross-disciplinary integration, structured design, and proactive governance. Respiratory medicine - by virtue of its inherently multimodal nature and richly visualizable longitudinal data (including imaging, pulmonary function tests, blood gas analysis, dynamic symptom trajectories, wearable-derived metrics, ventilator parameters, and sleep monitoring data) - is uniquely positioned to leverage digital twin and immersive XR scenarios to systematically address the core limitations of conventional classrooms: phenomena that are ‘invisible, difficult to simulate, and hard to reason through.’ This proposal is grounded in international consensus on digital health and AI competencies, systematically aligning with authoritative frameworks such as the Digital Education Competency Outcomes for Doctors in Europe (DECODE) and the Best Evidence Medical Education (BEME) Collaboration. It establishes a four-dimensional instructional framework encompassing Knowledge - Tools - Clinical Reasoning - Ethical Governance. At its core is an ‘AI + Case Chain’ modular curriculum: Pre-class: diagnostic pre-assessments and learner profiling to identify cognitive baselines; In-class: multimodal data visualization integrated with conversational standardized patients or interactive intelligent cases to enable immersive, context-rich clinical reasoning training; Post-class: learning analytics-driven personalized feedback loops that feed back into scholarly inquiry, creating a seamless teaching - learning- research cycle. Assessment and safeguard mechanisms center on a multi-modal evaluation system based on Objective Structured Clinical Examinations (OSCEs), adhering strictly to the principle of ‘AI-supported, never AI-substituted’ assessment. Ethical compliance and trustworthy AI governance -including data privacy protection, algorithmic bias mitigation, clear delineation of human -AI responsibility boundaries, and academic integrity - are embedded as non-negotiable safeguards. Ultimately, by leveraging interdisciplinary innovation centers and reusable digital infrastructures, this initiative transforms the traditional classroom into a next-generation smart learning environment characterized by visualizability, deep interactivity, and reasoning-driven pedagogy, systematically cultivating a vanguard cohort of talent ready to lead ‘smart respiratory care centers’ and ‘metaverse medicine’ practices by 2035.

Key Words: metaverse medicine; digital health competency; digital expert avatars; digital twin lung; multimodal learning and visualized reasoning; trustworthy AI governance and ethics

This study aims to compare the clinical value of a general large language model (DeepSeek GPT) and a disease-specific optimized model (BAIMGPT) in pulmonary nodule consultation and management. Through a multicenter real-world study, 1,000 patients with pulmonary nodules from 12 hospitals will be recruited for a randomized self-controlled trial evaluating the performance of the two models across eight dimensions: convenience, friendliness, sense of security, accuracy in question comprehension, response professionalism, voice interaction, visual empowerment, and demand level. The core methodologies include dual user evaluation, third-party blinded review, and endpoint assessment. The study was expected to validate BAIMGPT’s advantages in improving screening awareness, personalized management, and physician-patient trust, providing evidence-based support for AIpowered early lung cancer screening. The protocol has passed ethical review, with anonymized data processing ensuring both innovation and safety.

Key Words: plmonary nodule; BAIMGPT; DeepSeek GPT

Obstructive sleep apnea (OSA), a common sleep-related disorder with a high prevalence and significant burden, has attracted widespread attention. Currently, OSA management faces several challenges: lack of professional diagnostic equipment, insufficient expertise among primary care physicians, uneven distribution of medical resources, and low awareness of the disease. To further address these challenges, we need to adopt new quality productive forces and establish a virtual OSA platform. This platform will integrate advanced medical technology and big data analysis to overcome limitations in professional knowledge and availability of specific equipment, thereby providing patients with more accurate and personalized diagnosis and treatment plans.

Key Words: obstructive sleep apnea; internet of things; metaverse; virtual reality; augmented reality

Metaverse, through virtual reality, augmented reality, blockchain, and other technologies, offers immersive experiences in a virtual world. In the field of mental health care, metaverse technology enhances patients' cognitive functions and social capabilities, thereby facilitating early diagnosis and assessment of diseases. This paper provides an overview of the concept and steps involved in the application of metaverse medicine, as well as an analysis of the efficacy of metaverse technology in treating mental health disorders such as autism spectrum disorder (ASD), Alzheimer’s disease (AD), and anxiety.

Key Words: metaverse; mental health

This article reviews the development history and core technologies used in artificial intelligence （AI）, reviews the development history of large language models, summarizes the limitations and deficiencies of large language models, and identifies prospects for the development of general AI. This paper summarizes the performance strength and typical application scenarios of AI in the current medical field, analyzes the shortcomings of its application, and proposes the concept and classification of medical AI. On this basis, from the perspectives of medicine serving human beings and medicine’s own development, this commeutary defines the development goals of medical AI, and gives two different construction methods and paths for medical AI. 

Key Words: artificial intelligence; large model; metaverse in medicine; medical artificial intelligence

Sepsis remains challenging to teach in the intensive care unit (ICU), not due to a lack of guidelines, but because clinical priorities must be continually reassessed as patient physiology evolves. Although trainees are familiar with guideline recommendations, they often hesitate at the bedside when multiple time-sensitive decisions compete for attention. To address this gap, we implemented a time-axis-based framework to structure sepsis case discussions during routine ICU teaching, including bedside rounds and formal case reviews. This approach organizes management into four sequential yet overlapping decision phases, helping maintain focus on immediate priorities while anticipating subsequent changes. In our experience, trainees demonstrated improved ability to articulate decision prioritization and escalation of care. Rather than introducing new guideline content, the framework enhances the visibility of temporal prioritization and clinical reasoning during discussion. By centering on timing and prioritization, it complements guideline-based teaching and provides a clearer structure for time-critical decision-making in ICU training.

In the context of medical education transformation in the 21st century, neurology and neurosurgery education face multiple structural challenges, including complex neuroanatomy, scarcity of cadaveric specimens, and stringent safety requirements. Through systematic literature review and theoretical modeling, this study explores the current application status, technical infrastructure, and educational efficacy of Metaverse technology in this field. A four-layer analytical framework comprising the "Infrastructure Layer, Cognitive Layer, Clinical Decision-making Layer, and Ethical Governance Layer" was constructed to systematically analyze key technologies such as Head-Mounted Displays (HMD), haptic feedback, and digital twins. Empirical evidence demonstrates that Metaverse technology effectively reconstructs spatial cognition, showing significant advantages in enhancing spatial reasoning in neuroanatomy, clinical thinking training via Virtual Standardized Patients (VSP), and stroke team collaboration simulations. However, challenges such as cybersickness, insufficient haptic fidelity, and data privacy ethics remain bottlenecks for large-scale adoption. The Metaverse is best characterized as a "progressive enhancement tool" for traditional neurological education rather than a complete replacement. Future research should focus on multi-center prospective studies to verify its long-term impact on real-world clinical performance.

Key Words: Metaverse; neurological education; virtual reality; digital twin; educational efficacy; theoretical framework

Objective: To propose a method for predicting immunotherapy outcome in non-small cell lung cancer based on computed tomography images before and after treatment. Methods: An improved U-net model incorporating Efficient Channel Attention was used to segment lesions. Radiomic features of lesions were extracted using PyRadiomics package and combined with biological indicators. Feature selection and dimensionality reduction were performed using linear discriminant analysis and Pearson correlation algorithms. A support vector machine was used to establish the predictive model. Results: The proposed segmentation model achieved a Dice coefficient of 90.09%, a positive predictive value of 89.23%, and an intersection over union of 82.15%, outperforming mainstream segmentation models. The proposed predictive model achieved an area under the curve of 85.05%, accuracy of 77.59%, specificity of 81.68% and sensitivity of 73.52%, all superior to models based solely on single-time computed tomography images or lacking biological features. Conclusion: The proposed method provides an effective approach for predicting the efficacy of immunotherapy in non-small cell lung cancer patients and offers a  valuable tool to support clinical decision-making.