Volume 3, Issue 1

Objective To synthesize the international evidence, implementation models, and barriers related to Hospital at Home (HaH), and to propose a China-oriented pathway for respiratory HaH by integrating family bed services, internet hospitals, internet-based medical services, Internet Plus Nursing Services, and community health services. Contemporary evidence suggests that HaH has evolved beyond a simple substitute for inpatient admission and is increasingly understood as a composite model of hospital-level acute care delivered outside the hospital through remote monitoring, virtual review, in-home nursing, home-based treatment, and rapid escalation pathways. Methods This review draws on recent systematic reviews, randomized trials, real-world studies, implementation research, and national and local policy documents. The analysis focuses on conceptual boundaries, the international evidence base, key respiratory indications, digital infrastructure, nursing and community coordination, and multidimensional evaluation across clinical safety, environmental footprint, social benefit, and economic value. Results Among appropriately selected patients, HaH appears comparable or superior to conventional inpatient care with respect to mortality, readmission, patient experience, functional recovery, and some cost-related outcomes. Respiratory conditions, particularly post-exacerbation management of chronic obstructive pulmonary disease, home oxygen therapy, home noninvasive ventilation, post-pneumonia transitional care, and intensified post-discharge follow-up, represent high-priority and operationally feasible scenarios for HaH. China already has several institutional components relevant to HaH, including family bed services, internet hospitals, internet-based diagnosis and treatment, Internet Plus Nursing Services, and community health services. However, these components remain only partially connected and require specialty-led integration, digital coordination, nursing execution, and community continuity to form an operational, evaluable, and scalable pathway. Environmental gains should not be assumed; instead, transport, hospital bed utilization, household energy use, consumables, waste, and digital infrastructure should all be assessed within a full-pathway framework. Conclusions HaH has developed into a model of hospital-level care outside the hospital that integrates remote monitoring, virtual rounds, in-home nursing, home-based treatment, and rapid escalation and referral. For China, the key issue is not whether HaH is conceptually feasible, but how existing institutional mechanisms can be reorganized into an integrated hospital–community–home pathway for respiratory care. Future implementation should be evaluated across four domains: clinical safety, environmental footprint, social benefit, and economic value.

Key Words: hospital at home; home hospitalization; respiratory disease; family bed services; internet hospital; Internet plus nursing services; environmental footprint; social benefit; economic value.

To systematically summarize the major challenges in developing medical GPT systems and, with reference to recent international reviews, evaluation frameworks, ethical and regulatory guidance, as well as Prof Chunxue Bai' s BAIMGPT White Paper, to outline practical solutions for translating large language models into clinically usable systems. Recent high-impact systematic reviews, methodological studies, real-world workflow evaluations, and governance guidance were synthesized to examine the main issues in medical GPT development, including factual reliability, knowledge updating, data governance, multimodal integration, workflow adaptation, explainability, bias, fairness, and accountability. Current evidence indicates that the bottlenecks of medical GPT go well beyond imperfect accuracy. Major challenges include hallucinations and factual inconsistency, limited ability to absorb newly updated medical knowledge, heterogeneous clinical data and unstable labels, insufficient support for multimodal decision-making, weak adaptation to real-world workflows, incomplete explainability and accountability, and concerns regarding bias, fairness, and ethics. Current LLMs remain sensitive to information order and quantity and are not ready for autonomous clinical decision-making. The mission of medical GPT development is not simply to improve language generation, but to transform large models into trustworthy medical intelligence systems with reliable knowledge, workflow compatibility, traceability, and governance readiness. At present, medical GPT should be positioned as a tool for cognitive augmentation and workflow support rather than a substitute for clinical judgment.

Key Words: BAIMGPT/medical GPT; large language model; clinical decision support; retrieval-augmented generation; data governance; human-AI collaboration; disease-specific agent; BAIMGPT

The deep integration of the metaverse and digital healthcare is fundamentally reshaping the nature of health data. Traditional electronic health records (EHRs), characterized as discrete, event-driven "solid records," are evolving into "digital life streams"—continuously generated, cross-platform flowing, and algorithmically reconfigured data processes. This ontological shift poses profound challenges to existing data governance frameworks. On one hand, the heightened fluidity of data renders traditional privacy mechanisms, which rely on clear data boundaries, increasingly ineffective. On the other hand, dominant platforms are "reterritorializing" this fluid data through technical standards, protocols, and hardware monopolies, thereby constructing novel power structures that extend from the software layer to the hardware layer. Employing "liquefaction" and "reterritorialization" as core analytical lenses, this paper systematically investigates the structural dilemmas of health data governance in the metaverse. It poses three central research questions: (1) How does the emergence of "digital life streams" alter the foundational premises of data governance; (2) How does the dialectic between liquefaction and reterritorialization shape the power dynamics of health data; (3) In the face of this dual movement, what form should a new governance paradigm take through a critical analysis of technological solutions (federated learning, zero-knowledge proofs, trusted execution environments), legal-regulatory models (GDPR, HIPAA), and market-based mechanisms—examined across five dimensions: power distribution, economic equity, value choices, accountability, and digital sovereignty—this paper reveals the inherent limitations of single-pronged governance approaches in addressing a highly dynamic data ecosystem. Building on this critique, the paper proposes a relational data governance paradigm and constructs a multi-level framework integrating micro-level technological architecture (privacy by default, interoperability standards, explainable algorithms), meso-level institutional innovations (data trusts, digital commons, participatory auditing), and macro-level legal reforms (digital personality rights, digital gatekeeper regulation, global minimum standards). This framework moves from "data ownership" to "data relational rights" and supplements "individual privacy protection" with "collective digital well-being," achieving a dual conceptual elevation. It further embeds operational mechanisms such as the "four elements" of data trusts (trustee composition, decision-making, benefit distribution, supervision) and Ostrom's eight principles for governing digital commons. Drawing on case studies including the Apple Health ecosystem, VR psychotherapy platforms, and the European Health Data Space (EHDS), the paper offers both theoretical integration and policy references for the governance of health data in the metaverse.

Key Words: metaverse; health data governance; digital life stream; liquefaction; reterritorialization; data trusts; digital personality rights.

With the rapid evolution of precision oncology, particularly in non-small cell lung cancer (NSCLC), therapeutic decision-making is increasingly shaped by histologic subtype, driver alterations, immune biomarkers, and minimal residual disease (MRD). Under this paradigm, conventional pathology based solely on morphologic interpretation is no longer sufficient for modern clinical needs. The integration of artificial intelligence (AI) and digital pathology has transformed whole-slide imaging (WSI) from static glass slides into computable, sharable, and traceable data objects, enabling automated tumor region detection, histologic classification, tumor cell proportion estimation, PD-L1 quantification, tumor microenvironment analysis, and even prediction of potential molecular phenotypes. In parallel, molecular testing has expanded from a limited number of actionable genes to broad multigene panels, while liquid biopsy and circulating tumor DNA (ctDNA) provide complementary options for molecular profiling when tissue is limited. MRD monitoring further shifts lung cancer management from one-time pretreatment stratification toward dynamic peri-treatment risk assessment. This review systematically summarizes the roles of AI-assisted pathology interpretation, the integration of driver mutations with PD-L1, TMB, ctDNA and MRD, the coupling of digital pathology with molecular subtyping, the importance of data standardization in precision medicine, and the major barriers to clinical translation, including insufficient external validation, platform heterogeneity, limited interpretability, regulatory concerns, and fragmented workflows. We argue that the true value of AI-enabled digital pathology and molecular testing lies not merely in improving the accuracy or efficiency of individual diagnostic steps, but in establishing an intelligent companion diagnostic system spanning the entire continuum of lung cancer care. Such a system can continuously integrate pathology, molecular profiling, liquid biopsy, MRD surveillance, and clinical decision-making. Looking forward, the field is expected to evolve from single-task algorithms to multimodal foundation models, from static companion diagnostics to dynamic companion diagnostics, and from isolated laboratory tools to regionalized, platform-based intelligent ecosystems, ultimately promoting data-driven precision lung cancer care.

Key Words: lung cancer; digital pathology; artificial intelligence; molecular testing; companion diagnostics

This paper addresses the full-cycle medical education system that integrates "undergraduate medical education—postgraduate medical education—continuing medical education." It proposes the design and development approach of the "Huisheng Intelligent Education" metaverse platform to address current challenges, including outdated educational resources, unequal clinical practice opportunities, the tension between standardized training and personalized development, and fragmented management across different stages. The platform leverages extended reality and artificial intelligence as core technologies to construct four key components: a medical knowledge ecosystem, virtual teaching scenarios, a comprehensive teaching intelligence system, and a full-cycle teaching management hub. These components form a closed-loop path for knowledge delivery, situational training, process tracing, assessment and feedback, personalized guidance, and governance improvement. The early practice example is the metaverse-based multi-disciplinary team (MDT) teaching unit for difficult-to-diagnose pulmonary nodules, aimed at training residents in image interpretation, risk stratification, consensus formation, and doctor-patient communication. It supports synchronized participation of residents from both Shanghai and Xiamen hospitals in the same virtual space, providing consistent teaching and review, thus laying the foundation for subsequent objective evaluations and scalable implementation.

Key Words: full-cycle medical education; training and management; huisheng intelligent education; metaverse platform

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 systematically review recent advances in artificial intelligence (AI) and multimodal fusion for differentiating benign from malignant pulmonary nodules, with a focus on the theoretical basis, key technologies, clinical utility, and practical boundaries of integrated decision-making based on imaging, clinical data, and blood-based biomarkers. Methods　International guidelines for pulmonary nodule management, classic risk prediction models, recent AI-based imaging studies, multi-omics and liquid biopsy studies, and methodological consensus documents were reviewed. Evidence was synthesized from six perspectives: the significance of multimodal assessment, integration of imaging and clinical variables, synergistic value of blood biomarkers including ctDNA and circulating genetically abnormal cells (CAC), clinical potential of multimodal models, the boundary between decision support and decision replacement, and current challenges with possible solutions. Results　Current pulmonary nodule management still relies primarily on nodule size, volume, density, margin characteristics, growth dynamics, and conventional clinical risk factors such as age, smoking history, and prior malignancy, under the framework of established guidelines and prediction models. However, in subcentimeter nodules, subsolid nodules, multiple nodules, inflammation-related nodules, and intermediate-risk nodules, single-modality imaging features and conventional models remain inadequate in calibration and net clinical benefit. AI-based radiomics, deep learning, and multimodal machine learning can extract high-dimensional CT features beyond human visual recognition and improve risk stratification when combined with clinical variables. Meanwhile, liquid biopsy approaches, including cfDNA/ctDNA methylation, fragmentomics, CAC, and proteomic classifiers, provide additional molecular and cellular evidence for intermediate-risk nodules, thereby helping reduce unnecessary invasive procedures and accelerating precision diagnosis in truly high-risk cases. Nevertheless, real-world implementation remains limited by data heterogeneity, insufficient external validation, lack of assay standardization, high-dimensional low-sample-size issues, and regulatory and reimbursement barriers. Conclusion　The differential diagnosis of pulmonary nodules is evolving from single-modality imaging judgment toward multimodal integrated decision-making based on imaging, clinical data, and biomarkers. At the current stage, AI should be positioned as a decision-support tool rather than a decision-replacement tool. Future practice-changing systems will likely be prospectively validated, interpretable, auditable, guideline-concordant multimodal platforms that can be seamlessly embedded into pulmonary nodule clinics and multidisciplinary workflows.

Key Words: pulmonary nodule; artificial intelligence; multimodal fusion; radiomics; deep learning; cfDNA methylation; circulating genetically abnormal cells; proteomics; decision support

With the increasing adoption of low-dose computed tomography (LDCT) screening and the widespread use of chest CT in health examinations, chronic disease management, and multidisciplinary care, the detection rate of pulmonary nodules has risen substantially. Accordingly, the focus of pulmonary nodule management has shifted from simple lesion detection to refined evaluation, risk stratification, and longitudinal follow-up. Traditional radiologic assessment mainly relies on nodule diameter, density, margin characteristics, and interval growth. Although these approaches remain clinically valuable, they are limited by interobserver variability, suboptimal reproducibility, and difficulty in longitudinal comparison, especially in small nodules, juxta-vascular nodules, pleural-based nodules, subsolid nodules, and multiple nodules. Deep learning can automatically extract multi-scale imaging representations from two-dimensional, three-dimensional, and longitudinal CT data. In recent years, it has been widely applied to pulmonary nodule detection, precise segmentation, phenotypic characterization, malignancy risk prediction, dynamic follow-up, and progression forecasting. This review summarizes the clinical basis of refined pulmonary nodule imaging assessment and the current management framework, and systematically discusses recent advances in deep learning for nodule detection and segmentation, radiologic phenotype characterization, malignancy risk stratification, longitudinal follow-up, and clinical translation. In addition, based on the Fleischner Society guidelines, ACR Lung-RADS, BTS guideline, and recent consensus statements on subsolid nodules, this review analyzes the major barriers to real-world implementation, including insufficient external validation, heterogeneity of reference standards, limited interpretability, poor probability calibration, and incomplete workflow integration.

Key Words: pulmonary nodule; deep learning; low-dose computed tomography; ground-glass nodule; refined imaging assessment; risk stratification