ST凯利（300326）公告，公司于近日收到涌金投资发来的《上海凯利泰医疗科技股份有限公司要约收 购报告书摘要》。涌金投资拟以5.18元/股的价格，向除收购人以外的上市公司全体持有无限售条件流 通股的股东发出部分要约收购，收购股份数量为7170.13万股，占上市公司总股本的10%。本次要约收 购旨在进一步提高对上市公司的持股比例，坚定投资者对公司未来发展信心。 ...

闻乐 发自 凹非寺 量子位 | 公众号 QbitAI AI现在真能救命！ 传统的心脏MRI（磁共振成像）检查可能遗漏的关键风险信号，现在能够被AI捕捉了。 登上《Nature》子刊的最新研究—— 多模态 AI模型 MAARS ，首次用深度学习方法处理原始MRI图像，实现了对心源性猝死风险 （SCDA）的高精度预测，准确率高达 89% 。 这项研究来自约翰霍普金斯大学。 目前， 肥厚型心肌病 的临床诊断准确率仅50%，这使得医生在决定患者是否需要植入除颤器等保护措施时，常常面临"掷骰子"般的选择。 MAARS模型采用3D视觉 Transformer 架构，直接分析LGE-CMR（增强心血管磁共振成像）原始信号强度，避免了人工解读的主观性。 该模型能够从原始MRI图像中挖掘 隐藏的 纤维化瘢痕模式。 这些以往被医生忽略的细微结构变化，正是触发致命性心律失常的关键信号。 现在，MAARS可以将这些隐藏的病理信息转化为可量化的风险指标，提前发现连资深心脏专家都有可能忽略的致命风险。 在40岁到60岁人群中诊断准确率高达93% MAARS的核心优势在于通过多模态数据融合与 3D-ViT 技术，首次实现对心脏纤维化瘢痕的 ...

Core Insights - The emergence of medical AI agents marks a significant shift in the healthcare industry, transitioning from a focus on large models to intelligent agents that enhance service delivery and operational efficiency [1][2][3] Group 1: Development of Medical AI Agents - The relationship between large models and intelligent agents is characterized by a division of labor, where large models handle cognitive tasks while agents perform actions based on environmental inputs [2] - Medical AI agents are being deployed across various healthcare scenarios, including patient services, diagnostic assistance, hospital management, and even in research and education [3][4] Group 2: Application Scenarios and Key Players - In-hospital applications include patient services, diagnostic assistance, and hospital management, with companies like 惠每科技, 金域医学, and 微医 leading the charge [4] - Out-of-hospital applications feature AI family doctors and health management systems, with key players such as 京东健康 and 平安好医生 [4] - Specialized AI agents are emerging in fields like drug development and insurance claims, with companies like 健康之路 and 中康科技 involved [4] Group 3: Personalization and Specialization - The trend towards personalization in AI agents includes the creation of virtual personas and unique names to enhance user engagement and trust [5][6] - Specialized diagnostic agents are being developed to address specific medical conditions, such as the "腹痛诊疗智能Agent" for abdominal pain management [8][9] Group 4: Challenges and Future Directions - Current medical AI agents are primarily iterative upgrades and have not yet achieved breakthroughs in foundational models [12][14] - The development of specialized agents requires deep involvement from clinical experts to ensure effectiveness and relevance [15] - The advancement of foundational AI capabilities is crucial for the realization of fully functional medical AI agents [16]

编者按： 更好发挥市场作用，更好激发发展潜力。行政审批的"减法"，换来市场活力的"加法"。 4月24日，《市场准入负面清单（2025年版）》对外发布，事项数量从2018年首版的151项压减至106 项。"一单尽列、非禁即入"，让经营主体有了更明晰的认知和更稳定的预期，让"无形的手"释放更多活 力。 市场，当今世界最稀缺的资源。拥有超大规模且极具增长潜力的市场，是我国发展的巨大优势和应对变 局的坚实依托。建设全国统一大市场是构建新发展格局、推动高质量发展的需要。当前，各地区各部门 贯彻落实以习近平同志为核心的党中央决策部署，不断推动全国统一大市场建设走深走实。 即日起，新华社推出"加快建设全国统一大市场一线观察"系列报道，立足一线采访调研，围绕完善规则 制度、推动设施联通、促进要素流通、加强市场监管等方面，展现各地区各部门聚焦重点难点，以有力 有效、可感可及的举措激发市场活力、畅通经济循环、推动高质量发展的生动实践。 市场准入是经营主体参与经济活动的前提，不断完善市场准入制度是建设全国统一大市场的必然要求。 新版负面清单再"瘦身"，让企业可进的领域更多；清理整治市场准入壁垒，让要素资源顺畅流动……近 段时间以来 ...

Core Viewpoint - The RuiPath pathology model, developed by Ruijin Hospital in collaboration with Huawei, aims to enhance the efficiency and accuracy of pathology diagnostics in China by leveraging AI technology [1][5]. Group 1: Model Development and Features - The RuiPath model is a clinical-grade multimodal pathology model that covers 90% of the annual cancer incidence in China, addressing 19 common cancer types and hundreds of auxiliary diagnostic tasks [1][5]. - The model has achieved state-of-the-art (SOTA) performance in 7 out of 14 auxiliary diagnostic tasks tested against 12 mainstream public datasets, surpassing the performance of Harvard's UNI2 model [4]. - The model's core "visual foundation model" was developed using over one million high-quality digital pathology slides from Ruijin Hospital, utilizing Huawei's AI toolchain for annotation, training, and fine-tuning [2][4]. Group 2: Efficiency and Impact - The implementation of the RuiPath model allows pathologists to increase their daily workload from 200-300 slides to 400-500 or more, significantly improving diagnostic efficiency [5]. - The model aims to standardize digital pathology practices across hospitals in China, enabling easier deployment and reducing training costs for other institutions [5][10]. - The collaboration between Ruijin Hospital and Huawei has streamlined the model training process, allowing for the completion of the RuiPath model development with only a 16-card cluster, making it more accessible for hospitals [10][11]. Group 3: Industry Challenges and Solutions - There is a significant shortage of pathology doctors in China, with only about 20,000 available and a gap of 140,000 needed, highlighting the importance of AI solutions in addressing this challenge [5]. - The partnership has evolved through two phases: digitalization and smart pathology, focusing on data standardization and collaborative model development [7][8]. - The use of Huawei's ModelEngine has transformed the annotation process, allowing pathologists to review over 700 slides in a day, thus enhancing both efficiency and accuracy [10].

Core Viewpoint - Brainlab, a leading company in the German medical technology sector, has unexpectedly postponed its IPO plans due to ongoing geopolitical uncertainties affecting capital markets, despite having received oversubscription for the offering [2][4]. Company Overview - Founded in 1989 and headquartered in Munich, Germany, Brainlab specializes in developing image-guided surgery, radiation therapy, and digital operating room solutions, with a core philosophy of "software-defined medicine" [7]. - The company has installed over 5,000 systems globally, covering 85 countries, and holds a 60% market share in the neurosurgery navigation field, being recognized as the "gold standard" in surgical navigation [7]. IPO Details - Brainlab planned to issue 2 million new shares and transfer up to 3.2 million existing shares, with a pricing range of €80-100, aiming to raise approximately €416 million [2]. - The overall valuation of the company was estimated at around €1.7 billion (over 14.3 billion RMB) [2]. Financial Position - The company emphasized that its cash flow situation does not rely on IPO funding for organic growth, providing financial confidence for the postponement [5]. - Brainlab participated in a €27 million Series A funding round for neurosurgical robotic developer Robeauté, indicating its strong cash flow position [5]. Management Perspective - CEO Rainer Birkenbach stated that the current geopolitical uncertainties and market volatility led management to believe that it was not the optimal time for the IPO, despite positive business performance in the previous quarter [4]. Market Impact - Brainlab is the second German company to postpone its IPO this year, following Autodoc, which has negatively impacted the IPO market in Germany and Europe [10]. - A survey indicated that 74% of European investment bankers expect an increase in IPO activity in 2025, but uncertainties from the German elections and potential U.S. tariffs are affecting market confidence [10]. - The DAX index has risen by 15.63% this year, while the expected growth rate for the real economy is only 0.2%, highlighting a divergence between capital markets and the real economy [10]. Strategic Considerations - The case of Brainlab illustrates that even in favorable conditions such as oversubscription, strategic determination and timing are crucial considerations for capital operations in the medical technology sector [11].

Group 1 - The 2025 Global Digital Economy Conference will be held in Beijing from July 2-5, focusing on "Building Digital Friendly Cities" and showcasing the application of digital technology in high-quality industry development [1] - The year 2025 is defined as the year of large model application landing, with global large model technology accelerating from "technical verification" to "commercial closed loop" [3] - The integration of artificial intelligence with business scenarios is becoming an inevitable trend, with the current supply-demand collaboration catalyzing the performance upgrade of AI products and cost reduction in various scenarios [5] Group 2 - The Dr.GPT health model developed by the company integrates general knowledge, clinical cases, medical research, health records, and online consultation data to support health decision-making and provide personalized health management services [7] - The company officially joined the Beijing Artificial Intelligence Industry Alliance in June 2025, collaborating with major tech firms to promote industrial synergy and ecological development [8] - The company aims to deepen the application of AI in the healthcare sector and expand cooperation with various stakeholders, contributing to the digital transformation of the healthcare industry and enhancing public health levels [8]

撰文丨王聪 编辑丨王多鱼 排版丨水成文 肺癌 是全球癌症相关死亡的首要原因。对于能够检测特定基因突变以实现靶向治疗且经济实惠、无创的方法的需求，以及预测患者生存结果的需求，凸显了提升 诊断和预后能力的重要性。当前的肺癌诊断模型常常无法整合多样化的患者数据，导致临床评估不全面。 2025 年 7 月 2 日， 温州医科大学附属第一医院 黄晓颖 教授、 北京大学未来技术学院 王劲卓 、温州医科大学 张康 、四川大学华西医院 王成弟 等，在 Cell 子刊 Cell Reports Medicine 上发表了题为 ： AI-enabled molecular phenotyping and prognostic predictions in lung cancer through multimodal clinical information integration 的研究论文。 该研究开发了一款多模态 集成 AI 模型 —— LUCID ， 通过多模态临床信息整合，实现了肺癌分子表型分析及预后预测。 除了突变识别之外，准确的生存时间预测仍是优化肺癌治疗策略的关键组成部分。这种预后信息使临床医生能够制定更个性化的治 ...

Core Viewpoint - The announcement highlights the abnormal trading fluctuations of the convertible bond "塞力转债," which experienced a cumulative price increase of over 30% over three consecutive trading days, prompting a disclosure to investors regarding potential valuation risks [2][8]. Group 1: Convertible Bond Trading Situation - "塞力转债" recorded a closing price of 175.078 yuan per bond as of July 3, 2025, representing a premium of 75.08% over the face value and a conversion premium rate of 9.65% [2][8]. - The bond experienced a cumulative price increase of over 30% from July 1 to July 3, 2025, qualifying as an abnormal trading situation under the Shanghai Stock Exchange regulations [2][8]. Group 2: Company Operations and Major Events - The company confirmed that its production and operational activities are normal, with no significant changes in market conditions or industry policies that could affect the bond's trading price [9]. - There are no undisclosed major events related to asset restructuring, share issuance, or significant transactions that could impact the company's stock or bond prices [10]. Group 3: Share Pledge and Control - The controlling shareholder,赛海科技, holds 21,642,540 shares, representing 11.33% of the total share capital, and has recently unpledged 6,700,000 shares while pledging 6,000,000 shares, resulting in a total of 15,100,000 shares pledged [19][20]. - The actual controller, 温伟, holds 9,634,208 shares, representing 5.04% of the total share capital, and has similarly unpledged 3,850,000 shares while pledging 3,000,000 shares, leading to a total of 8,300,000 shares pledged [20].

本公司董事会及全体董事保证本公告内容不存在任何虚假记载、误导性陈述 或者重大遗漏，并对其内容的真实性、准确性和完整性承担法律责任。 一、募集资金基本情况 根据中国证券监督管理委员会《关于核准合富（中国）医疗科技股份有限公 司首次公开发行股票的批复》（证监许可〔2021〕4074 号）核准，合富（中国） 医疗科技股份有限公司（以下简称"公司"）首次向社会公开发行人民币普通股 证券代码：603122 证券简称：合富中国 公告编号：临 2025-030 合富（中国）医疗科技股份有限公司 关于募集资金专户完成销户的公告 ），并使用部分募集资金向合益信息提供借款以 实施募投项目，公司于 2022 年 6 月与全资子公司合益信息、保荐机构海通证券 股份有限公司、存放募集资金的富邦华一银行有限公司上海徐汇支行签署了《募 集资金专户存储四方监管协议》（以下简称"四方监管协议"）。三方监管协议、 四方监管协议明确了各方的权利和义务，与上海证券交易所《募集资金专户存储 三方监管协议（范本）》不存在重大差异，公司在使用募集资金时已经严格遵照 履行。 为提高募集资金收益率，公司于 2024 年 6 月 19 日召开第二届董事会第十一 ...