截至2025年3月31日，在线自营店有效GMV(商品交易总额)保持健康增长，年度活跃用户数和自营会员 数量持续提升，自营SKU同比增长33.6%至123万。同时，我们持续提升慢病用户体验，慢病用户数持 续提升的同时，人均DOT时长同比增长9%。报告期内，我们进一步整合优化药品物流网络布局，致力 于提升整体物流体验。药品次日达水平持续保持高位的同时，次日送货上门服务逐步拓展至覆盖广州、 宁波、金华、嘉兴、绍兴、湖州、台州和温州等13个城市。 报告期内，本集团收入实现稳定增长，总收入为305.98亿元，同比增长13.2%。报告期内净利润额为 14.32亿元，同比增长62.2%。调整后净利润额为19.5亿元，同比增长35.6%，调整后净利润率从5.3%提 升至6.4%。报告期内，天猫健康平台业务持续稳健发展，年度GMV(商品交易总额)保持稳健增长，年 度活跃用户数(过往于十二个月内在天猫健康平台实际购买过一次或以上商品的消费者)持续提升。报告 期内，本集团积极运营天猫健康类目的广告业务，持续赋能商家，提升品牌商经营服务能力，其收入实 现高速增长。在医疗大模型应用领域，医药电商业务场景下，从流量、供给和服务承接等实现 ...

Group 1 - The core viewpoint of the article highlights the rapid integration of large language models into the healthcare system, with 98 out of the top 100 hospitals in China claiming to have completed model deployment by April 30, 2025 [1] - Among these hospitals, 38 have developed 55 vertical medical models tailored to their specific needs based on general models [1][4] - The shift from "buyers" to "developers" indicates that doctors are becoming crucial contributors to AI development in healthcare [2][12] Group 2 - The emergence of specialized models marks a new era in healthcare AI, with hospitals increasingly focusing on vertical models that address specific diseases and clinical needs [4][6] - The DeepSeek model has significantly lowered the barriers for hospitals to deploy large models, allowing even those with no prior experience to customize and implement them [4][11] - As of April 30, 2025, there are 22 specialized vertical models in development, covering various medical fields such as cardiology, nephrology, and oncology [7][9] Group 3 - The collaboration between hospitals and enterprises remains the dominant model for developing vertical medical models, although independent hospital development is on the rise [13][15] - Major companies like Huawei, China Telecom, and iFlytek are actively involved in the development of these vertical models, indicating a trend towards increased collaboration between hospitals and tech firms [16][19] - The regulatory landscape for large models is evolving, with potential classification as medical devices requiring rigorous approval processes [19][20]

Core Insights - OpenAI has released a new medical-focused evaluation dataset called HealthBench, which consists of 5,000 core test dialogues created by 262 doctors from 60 countries, enhancing the dataset's difficulty, authenticity, and richness [1] - The performance of large models in healthcare has significantly improved, with GPT-3.5Turbo scoring 16%, GPT-4o at 32%, and o3 achieving 60%, indicating substantial progress overall [1] - Smaller models have shown remarkable advancements, with GPT-4.1nano outperforming GPT-4o while reducing costs by 25 times [1]

本报讯(记者袁传玺)在日前由中国医院协会举办的"公立医院运营效能提升与创新发展"研讨会上，温州 医科大学附属第一医院(以下简称"温医大附一院")全面介绍了该院智慧医院建设的进展情况。据了解， 从2024年起，温医大附一院与京东健康股份有限公司(以下简称"京东健康")围绕"未来数字医院"建设项 目，率先将新一代医疗大模型应用到医院全场景之中，切实优化了患者服务、提升了医生效率，让医院 管理更加高效。 双方将融合视觉、触觉及生物传感技术，打造"全感官数字医生"。通过可穿戴设备实时获取患者的心 率、血压等体征数据，数字分身能够动态调整问诊策略，为患者提供更加精准的建议。双方还将共同打 造数字孪生医院平台，将其建设为医院的数字分身。未来将实现亚毫米级建模与纳秒级数据同步，数字 孪生医院不仅能够精准复刻实体医院的运行状态，还能预测实体世界的未来变化。 目前，温医大附一院和京东健康凭借强大的技术实力和专业的医疗团队，成功上线了一系列创新的医疗 服务平台，涵盖就医主入口、诊前智能预约、预问诊病史采集、诊中数字陪诊、诊后一清单、数字人宣 教等核心应用，全面覆盖患者就医各个环节，实现了门诊患者服务流程的闭环管理。 数字人宣教系 ...

Deepseek框架优化医疗影像品质，具身智能则提升机器人环境适应性，歌锐将二者集成于"牛顿Endo"手 术机器人。自有算法+Deepseek大模型蒸馏+具身智能相结合： • 实现低剂量曝光控制与图像品质控制，有效提升图像质量的同时降低辐射风险 而在精准医疗时代，我们既看到手术机器人系统实现 0.1mm 级精度突破，也目睹着同质化企业陷入增量 不增利的困境。 量级扩张与价值困局的剧烈碰撞，正将整个行业推向质变临界点 —— 如何将技术突破转化为临床价值， 并突破低端内卷与国际化竞争的壁垒？ 在 2025 年 4 月 24 日举办的首届全球骨科大会上，思宇MedTech通过专访 歌锐科技高级副总裁刘建勋 ，深入了解这家企业如何以务实创新、全球布局和开放合作实现从原研创新到临床价值创造、从中国到世 界的跨越路径。 # 技 术落地 ， 从 来 不 是 P P T 空谈 在全球手术机器人技术版图中，中国已跃身第一梯队。 成立于 2021 年，凭借其技术创新和以临床价值为核心的理念，歌锐科技已迅速崛起为行业标杆企业，成 为了推动行业发展的全新力量。 在首届全球骨科大会上，歌锐荣获 2025 全球骨科技术创新奖， 以"牛顿 ...

Core Insights - Google Health and DeepMind have released a groundbreaking AI model, AMIE, which outperforms primary care physicians in simulated diagnostic scenarios, marking a milestone in conversational medical AI [1] - The AI-assisted medical consultation is gaining traction, with healthcare professionals supporting its use as an effective supplementary tool [3][5] - Left Medical Technology has launched a 24-hour online pediatric "AI family doctor," receiving positive feedback from thousands of parents [1][12] Group 1: AI in Healthcare - The AI model AMIE is recognized for its superior performance compared to primary care doctors in diagnostic simulations [1] - The number of medical AI models in China has surged from 50 in 2023 to over 200 in 2024, reflecting a 300% increase [8] - Left Medical Technology's "Left Hand Doctor" brand has become a leading smart doctor in China, providing various AI healthcare solutions across over 200 top-tier hospitals [10][13] Group 2: Impact on Medical Practice - AI consultation tools are seen as valuable for enhancing efficiency and quality in medical practice, particularly for younger doctors [3][5] - The "Left Hand Doctor" system has improved consultation efficiency by over 50%, allowing doctors to focus more on analyzing reports rather than basic symptom inquiries [7] - The introduction of AI in pre-consultation has addressed common issues in traditional consultations, such as incomplete medical history collection [5] Group 3: Market Dynamics and Investment - Investors are looking for "killer application" scenarios in the medical AI space, emphasizing the importance of addressing clinical pain points and providing user-friendly solutions [9] - The collaboration between Left Medical Technology and experts from Peking Union Medical College aims to create an "AI avatar" that learns from specialists' experiences, enhancing decision-making for both experienced and novice doctors [15] - The company plans to expand its AI services to underserved areas, optimizing resource allocation and improving primary healthcare capabilities [14]

Summary of Conference Call Notes Company and Industry Overview - The conference call involved **iFlytek Medical**, a company specializing in AI-driven healthcare solutions, particularly in the field of medical big models and patient management systems [1][2]. Key Points and Arguments Financial Performance - In 2024, iFlytek Medical achieved a revenue of **734 million** (CNY), representing a **32% year-on-year growth**. Gross profit reached **404 million** (CNY), with a **28% increase** year-on-year [3][4]. - The revenue structure improved, with G-end revenue growing by **10%**, B-end (hospital services) revenue increasing by **103%**, and C-end (patient services) revenue rising by **57%** [4]. - Sales collections grew by **33%**, and operational cash flow improved by **57%** year-on-year. Accounts receivable at year-end stood at **930 million** (CNY), with **76%** from government and state-owned enterprises [5]. Technological Advancements - iFlytek Medical has developed the **X1 medical big model**, which has surpassed the capabilities of existing models like GPT-4O in medical applications. The model can achieve diagnostic accuracy comparable to that of a deputy chief physician [6][7]. - The company has integrated various data types (voice, images) into its model, enhancing its ability to provide comprehensive patient services [8]. - iFlytek Medical's big model has been trained on a vast dataset, creating a **data flywheel** effect that continuously improves its performance through real-world applications [19][21]. Market Position and Strategy - iFlytek Medical aims to leverage its AI capabilities to enhance healthcare delivery across G-end (government), B-end (hospitals), and C-end (patients) sectors, establishing a diversified product matrix [9][10]. - The company has established partnerships with top hospitals, enhancing its credibility and market reach. It has covered **31 provinces** and provided services to over **910 million** patients [10][11]. - The government has shown support for AI in healthcare, with policies promoting the application of big models in various sectors, which bodes well for iFlytek Medical's growth [10]. Challenges and Opportunities - The company faces competition from other AI healthcare providers but believes its proprietary technology and data capabilities give it a competitive edge [31][34]. - iFlytek Medical is focused on expanding its services in both urban and rural healthcare settings, addressing the need for balanced medical resource distribution [10][36]. - The company is also exploring opportunities in post-discharge patient management, which is becoming increasingly important as healthcare systems emphasize continuity of care [53][54]. Other Important Content - iFlytek Medical's AI solutions have been implemented in various hospitals, including **Qilu Hospital** and **Shandong Provincial Hospital**, demonstrating the practical application of its technology [15][16]. - The company has developed a comprehensive patient management system that includes follow-up care and personalized health plans, which has shown to reduce readmission rates significantly [55][56]. - iFlytek Medical's collaboration with **Huawei** focuses on providing computational power and data centers, enhancing its operational capabilities in the healthcare sector [43][44]. This summary encapsulates the key insights from the conference call, highlighting iFlytek Medical's financial performance, technological advancements, market strategies, and the challenges and opportunities it faces in the evolving healthcare landscape.

Core Insights - The company, known for its pioneering work in medical AI, has launched the "ShuKunkun" multimodal healthcare model and Digital Human 4.0 technology platform, aiming to transform the entire medical ecosystem from imaging diagnosis to hierarchical treatment [1][3]. Company Development - Since the introduction of the world's first coronary CTA product in 2017, the company has developed over 100 digital doctor products and obtained 17 Class III NMPA certifications, significantly innovating the diagnostic paradigm in radiology [3]. - The "ShuKunkun" model, which can analyze images, understand text, and apply clinical logic, is a key component of the company's strategy to evolve AI from a supportive tool to a core driver of the diagnostic ecosystem [3][4]. Technological Capabilities - The "ShuKunkun" model has demonstrated superior diagnostic accuracy, outperforming human doctors in complex liver disease diagnoses during an industry competition [3][4]. - The model integrates knowledge, reasoning, and experience, enabling it to assist doctors in making informed clinical decisions by processing various types of medical data [4]. Application Scenarios - The Digital Human 4.0 platform supports 12 core workflows in imaging, enhancing efficiency in hospitals by reducing report writing time by 50% and accelerating research data processing [4][5]. - The company has introduced an AI solution for ultrasound that supports comprehensive body examinations and has improved computational power by 50% [5]. Collaborative Initiatives - The company is collaborating with hospitals to implement the model across various departments, providing comprehensive patient management and operational support [5]. - A "Digital Doctor Intelligent Team" has been established to enhance the efficiency of chronic disease management and integrate medical prevention at the grassroots level [5]. Hardware Innovations - The company has developed AI-native hardware that allows seamless deployment across different healthcare settings, from top-tier hospitals to community clinics [6]. - A partnership with Huawei has led to the creation of a medical model computing machine that optimizes computational resources for AI applications [6]. Future Directions - The company has initiated the "ShuKunkun Model Open Co-construction Plan," inviting industry partners to collaboratively build a medical AI ecosystem [6]. - The Digital Human 4.0 platform is envisioned as an open platform that integrates AI across the entire healthcare chain, contributing to the Healthy China strategy [6].

持续推动医疗影像领域变革。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 据 IPO早知道消息， 数坤科技 日前在 CMEF 上亮相了升级后的 "数坤坤" 多模态医疗健康大模型 （以下简称"数坤坤"大模型） 。 依托这一前沿大模型，数坤科技推出了数字人体 4.0技术平台以及平台之上被"数坤坤"大模型全面赋 能升级后的数智影像、数智超声、数智医院、数智基层解决方案，和一系列加载了"数坤坤"能力的AI 原生硬件。 数坤科技创始人、董事长毛新生 表示， 数字人体 4.0将带动医疗健康产业升级进入新阶段，希望通 过数字化的人体，让所有的医生在为病人服务时能拥有跟今天完全不同的智能化手段。数坤在大模型 本身以及模型应用方面，都走在全世界的前列。 作为医疗垂类大模型， "数坤坤"已经可以识别出CT、MR、DR、X-ray、钼靶以及超声等模态的影 像数据，深刻理解患者的生化检查、诊断报告、既往病史和现病史等文本信息。同时，在学习了公域 这种创新模式相当于为医生配备了处理繁琐工作的 "智能助手"，显著提升诊疗效率与质量；同时为 患者提供24小时在线的"专属医生"，实现个性化的健康管 ...

Group 1 - Shenzhen People's Hospital, Shukun Technology, and Huawei have formed a deep collaboration to deploy the DeepSeek-R1 model and the "Shukun Kun" multi-modal healthcare model, enhancing the hospital's medical and research capabilities [3] - The platform equipped with DeepSeek and "Shukun Kun" medical models will provide comprehensive research support, enabling rapid construction of multi-modal databases and offering data analysis tools for researchers [5] - The global first peritoneal dialysis model was launched by Shenzhou Medical and Sun Yat-sen University First Affiliated Hospital, utilizing advanced DHC+DeepSeek dual-engine architecture for precise understanding and processing of complex information in the field [7] Group 2 - Yingxi Intelligent has deployed its first bipedal humanoid robot named "Supervisor" in its AI-driven fully automated drug discovery laboratory, aimed at enhancing laboratory operations [10] - The innovative implantable neurostimulation system by Boruikang has passed the review by the National Medical Products Administration, marking a significant advancement in medical technology [12] - The commercial production base for small nucleic acid drugs by Zhaowei Technology has officially commenced operations, featuring advanced equipment and a total designed capacity of 48 production lines [20] Group 3 - Zhengxu Bio's self-developed base editing drug CS-101 has successfully treated a Malaysian patient with beta-thalassemia, marking another significant achievement in gene editing therapy [23] - The OCT intravascular imaging catheter developed by Weiguang Medical has received registration approval for dual indications, enhancing diagnostic efficiency in coronary and carotid artery diseases [26] - The first platelet-related cell new drug XJ-MK-002 has been granted orphan drug designation by the FDA for treating congenital thrombocytopenia, showcasing innovation in rare disease treatment [28] Group 4 - The new generation PI3Kα inhibitor JYP0035 capsule has received approval for clinical trials, aimed at treating PIK3CA mutated, HR-positive, HER2-negative breast cancer, indicating strong R&D capabilities [30] - The first domestically developed antibacterial fishbone suture line has been approved by the NMPA, representing a breakthrough in surgical materials [43] - The flexible ultrasound probe project led by Suzhou Nolaisheng Technology has been approved as a national key research and development project, highlighting advancements in ultrasound technology [39]