本研究由 解放军总医院牵头，联合浙江大学医学院附属第二医院、复旦大学附属华山医院等共 11 家国内顶尖三甲医院，携手南京大学、吉林大学两所重点高校， 并汇聚 Pi3Lab、上海三友医疗器械股份有限公司 等产学研多方力量，共同完成了首个面向脊柱诊疗领域的大模型研发。 论文共同第一作者包括 赵明、董文辉博士、张阳医生， 核心贡献者包括来自 浙江大学医学院附属第二医院的陈其昕教授、夏顺楷医生，以及复旦大学附属华山医 院的马晓生教授、管韵致医生等。 通讯作者为 解放军总医院骨科医学部副主任孙天胜教授， 共同通讯作者为 南京大学智能科学与技术副院长单彩峰教授。 脊柱疾病影响全球 6.19 亿人，是致残的主要原因之一 。然而，现有 AI 模型在临床决策中仍存在「认知鸿沟」。缺乏椎体级别（level-aware）、多模态融合的指令 数据和标准化基准，是制约 AI 辅助诊断的关键瓶颈。 本文提出了一套统性的解决方案，包括 首个大规模、具有可追溯性的脊柱指令数据集 SpineMed-450K，以及临床级评测基准 SpineBench。 基于此训练出的专科 大模型 SpineGPT，在所有任务上均实现了显著提升， 仅仅 7B 参 ...

Core Points - The launch of the national medical insurance imaging cloud for cross-province retrieval marks a significant advancement in medical imaging services, allowing for seamless access to patient imaging data across different regions [1][2][3] - The initiative aims to enhance diagnostic accuracy and efficiency by overcoming traditional barriers associated with physical film storage and retrieval, thus facilitating better patient care [2][3][4] Group 1: Technological Advancements - The transition from traditional film to cloud-based imaging solutions addresses the limitations of physical films, such as their bulkiness, susceptibility to damage, and inability to capture detailed images [2][3] - The national medical insurance imaging cloud has indexed over 170 million imaging records, with a goal to establish a "national network" by 2027 [3] Group 2: Operational Efficiency - The implementation of a unified platform and indexing standards allows for efficient storage and retrieval of imaging data, significantly reducing costs associated with data transmission [3] - The cloud service is priced at 5 yuan per person per imaging instance, promoting local storage and cross-regional transmission as a medical service [3] Group 3: Patient-Centric Care - Patients can now access, download, and share their imaging data without incurring additional fees, which enhances the convenience of tiered medical services and remote consultations [3][4] - The ability to consult with high-level medical institutions remotely improves diagnostic quality and resource utilization across the healthcare system [4][5] Group 4: Future Innovations - The data generated from imaging services is expected to drive innovations in AI-assisted diagnostics and remote medical applications, further benefiting patients and healthcare professionals [4]

付超奇：我们有全国医保统一的信息平台，覆盖100多万家的定点医药机构，今年专门出台了医保 影像云索引，把影像的本地存储以及跨地传输作为一项医疗服务进行云胶片的集采。 央广网北京11月21日消息（记者杜希萌 李思默）据中央广播电视总台中国之声《新闻和报纸摘 要》报道，全国医保影像云跨省调阅启动仪式20日在北京举行，未来将加速推动实现影像检查结果"患 者可阅、同行可查、医保可核"。 国家医保局透露，目前，北京、河北、山西等24个省份和新疆生产建设兵团已完成医保影像云软件 部署。国家医保局大数据中心主任付超奇说，此次跨省调阅，通过医保打包购买服务的方式，困扰医疗 机构的影像存储等问题得到有效解决。目前，国家医保局归集的医保影像云索引数据超过1.7亿条， 2027年有望实现"全国一张网"。 随着医学影像全面迈入"云时代"，患者可以对自己的影像检查数据进行多次浏览、下载和分享，且 无需重复付费，极大便利分级诊疗和异地转诊。在北京大学第三医院院长付卫看来，不止于临床诊断， 这些数据还有望转化为技术创新的新增量。 付卫：进一步挖掘影像大数据的价值，进一步探索AI辅助诊断、远程医疗等方面的科学研究与创 新应用。 ...

Core Points - The national medical insurance imaging cloud cross-province retrieval officially launched on November 20, aiming to eliminate repeated examinations for patients and facilitate access to medical imaging data across different hospitals [1][2] Group 1: Benefits of Medical Imaging Cloud - The medical imaging cloud addresses the pain points of patients by reducing the frequency of repeated examinations, which has been a significant issue in cross-province medical services, with over 650 million direct settlements during the 14th Five-Year Plan period [2] - The integration of personal health data allows for a comprehensive medical history, enhancing personalized health management and improving diagnostic accuracy by providing doctors with access to historical imaging data [2] - The cloud-based system is expected to significantly reduce misdiagnosis and missed diagnoses, as it offers a vast pool of standardized data for AI-assisted diagnostics, improving the quality of medical services [2] Group 2: Challenges and Resistance - Concerns have been raised regarding the potential resistance from medical institutions due to the impact on revenue from imaging fees, which are a significant source of income for hospitals [3] - The national medical insurance bureau estimates that the unified platform could save approximately 80 billion yuan annually by reducing redundant examinations and physical film costs, which could be redirected towards medical innovation and digital technology reforms [3] Group 3: Information Security - Patient data security is a priority, with strict authorization protocols in place for accessing imaging data from other medical institutions, ensuring that patient consent is obtained before any data retrieval [4] - The medical insurance information platform maintains high security standards to protect patient information, and future measures will further enhance data protection [4] Group 4: Future Scope of Cross-Province Retrieval - The goal is to establish a nationwide medical insurance imaging cloud by the end of 2027, which will enable access to imaging data wherever there is medical insurance coverage [5] - The national medical insurance bureau has developed a unified indexing system for imaging data, allowing for efficient storage and retrieval across over 1 million designated medical institutions, covering more than 1.3 billion insured individuals [5] - The medical insurance cloud has already indexed 170 million imaging records, demonstrating the feasibility and effectiveness of this model [5]

11月20日，国家医保局在北京大学第三医院举行了全国医保影像云跨省调阅启动仪式。医保影像云跨省 调阅，即通过将患者数字影像资料上传至医保部门认可的影像存储中心，实现跨地区跨机构调阅检查资 料。这意味着，未来患者看病就医，不必重复检查，医疗机构可查看患者保存在云存储中的所有影像检 查资料。 北医三院可跨省调阅雄安新区等五地医保云影像 患者看病就医，刚做过的检查，到了另一家医院，就得再做一遍，这样的情况许多人都碰到过。尤其是 在跨省异地就医的场景下，检查多、检查复杂、检查重复成为反映强烈的痛点问题。 全国医保影像云跨省调阅就是针对这一问题，化解群众的"急难愁盼"，实现医保支付影像检查数据跨省 互通互认。 简单地说，患者在一家医疗机构做过的检查，其影像资料可上传至医保部门认可的影像存储中心，患者 可通过便捷方式阅读本人检查资料，同行可跨地区跨机构调阅检查资料，医保部门可核查已上传的检查 资料。 "我们聚焦影像数据找不到、存不起、跨省难、阅不快等难点，通过进一步制定政策、标准，明确实现 路径，高效推进医保影像云的建设工作。"国家医保局副局长王文君介绍。 在全国医保影像云跨省调阅启动当天，北医三院的五条调阅路线已实现实 ...

Core Insights - The Chinese government has approved the "Implementation Plan for Strengthening Basic Medical and Health Services," aiming to enhance grassroots medical capabilities and establish a 15-minute accessible healthcare service circle for residents within five years [1] Group 1: Policy and Market Trends - The AI-assisted diagnosis market in China is projected to exceed 80 billion yuan by 2025, with a compound annual growth rate of 58.3%, particularly in intelligent diagnosis and health management sectors, which are expected to grow over 70% [2] - The government is promoting AI-assisted diagnosis through various policies, including the inclusion of AI in medical insurance guidelines for radiology, ultrasound, and rehabilitation projects starting November 25, 2024 [1][2] Group 2: Technological Innovations - New technologies such as AI-assisted diagnosis, microfluidic chips, and point-of-care testing (POCT) are emerging to address the challenges faced by grassroots medical facilities, including limited personnel and high costs [2] - The introduction of the Minasis intelligent diagnostic platform allows for significant space and cost savings, integrating multiple diagnostic functions into a single device occupying only 1 square meter [2][3] Group 3: Implementation and Impact - The Minasis platform enables residents to complete blood tests in under 18 minutes, significantly reducing the traditional diagnostic cycle by over 55% and lowering testing costs by 30% [3] - The Chinese government plans to support the construction of 125 national regional medical centers and allocate 10 billion yuan for the development of tightly-knit county medical communities by 2025 [4] Group 4: Future Outlook - The equipment update policy is expected to enhance the diagnostic capabilities of county medical communities, with a focus on high-end and intelligent medical devices [5] - The new policies will likely drive demand for imaging, testing, and ICU equipment procurement in grassroots medical institutions [5]

Core Insights - The deployment of the DeepGEM model by Guangzhou Kingmed Diagnostics Group aims to enhance cancer diagnosis through accurate and timely gene mutation predictions [1][2] - The collaboration involves Tencent and Guangzhou Medical University First Affiliated Hospital, focusing on developing a multimodal model for pathology and genetics [1][2] Group 1: DeepGEM Model Development - DeepGEM provides accurate predictions of gene mutations related to lung cancer, achieving a prediction accuracy of 78% to 99% within one minute [1] - The model addresses the challenges of conventional gene testing methods, which are often complex, time-consuming, and costly, particularly in resource-limited areas [1] Group 2: Clinical Application and Future Plans - Following successful validation, the three parties will promote the clinical application of DeepGEM for lung cancer gene mutation prediction [2] - There are plans to further develop a multimodal model that integrates various omics data, including pathology, proteomics, and metabolomics, for AI-assisted diagnosis across multiple cancer types [2] Group 3: Vision and Collaboration - The initiative aims to serve as a model for translating clinical research into practical applications, benefiting the public [2] - Kingmed Diagnostics expresses a desire to collaborate with more partners to create intelligent and accessible clinical diagnostic solutions [2]

Core Insights - The collaboration between Guangzhou Jinyu Medical, Tencent, and Guangzhou Medical University aims to develop an AI model called DeepGEM for predicting gene mutations in tumor patients using conventional pathological images [1][5][6] - The DeepGEM model has shown promising results in accurately predicting lung cancer gene mutations, achieving a precision rate between 78% and 99% [2][3] - The partnership is expected to enhance the clinical application of DeepGEM and expand its capabilities to other cancer types, integrating various omics data for a comprehensive diagnostic approach [6] Group 1: Development of DeepGEM Model - The DeepGEM model was developed by a team from Guangzhou Medical University and Tencent, utilizing AI to predict lung cancer gene mutations from pathological images [2][3] - The model can process different types of biopsy samples and generate spatial distribution maps of gene mutations, enhancing the understanding of mutation patterns within tissues [3][4] - The model's performance has been validated with a large dataset from Jinyu Medical, covering 4,260 lung cancer patient samples across various medical institutions [4] Group 2: Clinical Implications and Future Directions - The collaboration aims to provide timely and cost-effective gene diagnostics, especially for patients in resource-limited areas, by combining AI screening with targeted gene confirmation [3][6] - The successful deployment of DeepGEM at Jinyu Medical marks a significant milestone in the exploration of multi-modal AI models for pathology and genetics [6] - Jinyu Medical's extensive data repository and commitment to integrating AI in medical testing are expected to lead to advancements in diagnosing not only tumors but also rare and complex diseases [5][6]

Core Insights - The collaboration between Kingmed Medical, Tencent, and Guangzhou Medical University First Affiliated Hospital aims to develop the AI model DeepGEM for predicting gene mutations in cancer patients using routine pathological images [1][5][6] - DeepGEM has demonstrated a high accuracy rate of 78% to 99% in predicting common lung cancer driver gene mutations, significantly improving the efficiency and accessibility of genetic diagnostics [2][4] Group 1: Development and Technology - DeepGEM is developed by a collaboration between Guangzhou Medical University First Affiliated Hospital, Guangzhou Respiratory Health Research Institute, and Tencent, marking a significant advancement from traditional pathology to genetic insights [1][3] - The model utilizes innovative techniques such as Multiple Instance Learning (MIL) and an end-to-end architecture that enhances prediction accuracy without the need for manual tumor region annotation [3][4] Group 2: Clinical Application and Validation - Kingmed Medical is providing a large-scale dataset for validating DeepGEM, with over 15,000 NGS tests conducted annually and a sample size of 4,260 lung cancer patients across various medical institutions [4][5] - The model has reached clinical auxiliary diagnostic levels for identifying mutations in genes like EGFR, KRAS, and ALK, showcasing its robustness and compatibility for clinical use [4][6] Group 3: Future Prospects and Expansion - The partnership aims to expand the application of DeepGEM beyond lung cancer to other cancer types, integrating various omics data for a comprehensive diagnostic approach [5][6] - The collaboration is seen as a milestone in the exploration of AI-driven pathology-genetics models, with aspirations to enhance the efficiency of clinical research and diagnostics in both cancer and rare diseases [6]

Group 1: Company Strategy and Focus Areas - The company aims to enhance its core competitiveness through industrial investments, focusing on the medical imaging and non-destructive testing sectors [2][3] - The company is actively seeking acquisition opportunities that align with its strategic development [2][3] - The company plans to expand its product portfolio through both depth and breadth, targeting cross-sector development [2][3] Group 2: Product Development and Market Position - The company has developed a portable multi-modal medical imaging system for emergency rescue applications, integrating portable DR and ultrasound devices [3][4] - The company’s industrial ultrasonic testing equipment is widely used in aerospace, petrochemicals, energy, and transportation sectors [4][5] - Key products include multi-modal medical imaging systems, specialized ultrasound diagnostic equipment, and automated non-destructive testing devices, which are expected to drive future revenue growth [5][6] Group 3: Research and Development - The company has participated in multiple national and provincial major research projects, with details disclosed in its prospectus and periodic reports [4][5] - The company has made advancements in AI-based breast cancer screening technology, enhancing the accuracy and efficiency of early detection [5][6] Group 4: Marketing and Sales Strategy - In 2025, the company will adjust its marketing strategy, focusing on major cities like Guangzhou, Shenzhen, Shanghai, and Beijing to strengthen its domestic and overseas marketing efforts [7][8] - The company aims to enhance its marketing network by attracting high-quality domestic and international distributors and improving local customer service capabilities [7][8] Group 5: Revenue Contribution - The sales revenue from surgical robots and automated non-destructive testing equipment constitutes a significant portion of the company's overall revenue [8]