西南证券主要观点如下： 智通财经APP获悉，西南证券发布研报称，中国医疗人工智能行业空间较大，政策推动行业快速发展； 讯飞医疗科技(02506)具备自研基座大模型，性能上限持续抬升。星火医疗大模型X1能够显著降低幻觉 现象，提升推理的正确性、专业性与可解释性；基于大模型和大数据技术平台，构建多元化GBC客户 产品矩阵。首次覆盖，建议关注。 2025年3月，公司正式发布基于深度推理技术的"讯飞星火医疗大模型X1"。该模型是当前唯一完全采用 全国产算力训练完成的医疗深度推理大模型。在真实场景测试中，其全科辅助诊断合理率达到94.0%。 2025年7月，讯飞星火医疗大模型完成能力升级，多项性能指标得到进一步提升。具体而言，全科辅助 诊断合理率提升至95.0%，体检报告解读合理率提升至86.3%，健康咨询解答率提升至91.5%，院端心血 管内科主要诊断合理率提升至91.2%。测试表明，在关键医疗场景任务上，其表现继续大幅领先于 OpenAI o3及DeepSeek R1等国际主流模型。 基于大模型和大数据技术平台，构建多元化GBC客户产品矩阵 截至2024年底，公司已向超过7万家基层医疗机构提供产品和服务，业务覆盖全 ...

91 Q F 9 t m 0 8 F Q 中国最 x 7 ● C e ma F E FFF "I"|I .... Lo) Al 000 ปี (D 今天分享的是：人工智能+医疗专题：2025年医疗人工智能产业报告 报告共计：99页 《2025年医疗人工智能产业报告》聚焦医疗AI的价值计量与商业化破局，展现其在政策、技术与需求驱动下的发展态势。2024年中国医疗AI解决方案市场 规模达164亿元，预计2030年将增至353亿元，CAGR为13.63%，应用已覆盖临床专科、医技支撑、基层医疗等全场景。临床专科中，胸外科、心内科、骨 科、神外科、内分泌科等AI应用成熟，从诊断、手术规划到术后康复全流程赋能，提升诊疗效率与精准度，如骨科手术机器人实现复杂手术规模化落地， 内分泌科AI盘活慢病管理。医技科以影像科、放疗科、病理科、检验科为核心，AI通过自动分割、智能阅片、流程优化等功能，破解人力短缺与效率瓶 颈，其中影像科AI辅助诊断覆盖多病种，病理科大模型提升泛化能力与可解释性。基层医疗AI凭借政策支持与场景适配，率先实现商业化闭环，通过CDSS 等系统弥补人才缺口，推动分级诊疗落地。行业发展面临价值分歧、数据治理成 ...

本文源自：市场资讯 作者：公告君 回购日期回购均价回购股数回购金额本年累计回购股数2025-12-185.12819.00万97.44万723.13万2025-12- 174.98140.00万199.23万704.13万2025-12-164.94464.08万316.83万664.13万2025-12-125.10115.50万79.07 万600.05万2025-12-115.01253.50万268.16万584.55万2025-12-095.15134.50万177.70万531.05万2025-12- 085.23722.50万117.82万496.55万2025-12-055.18328.00万145.12万474.05万2025-12-045.14718.10万93.16 万446.05万2025-12-035.09039.20万199.53万427.95万2025-10-245.6462.12万11.97万388.75万2025-10- 235.54335.00万194.02万386.63万2025-10-225.6391.77万9.98万351.63万2025-10-215.7015.96 ...

今天分享的是：2025年医疗人工智能产业报告-蛋壳研究院 报告共计：99页 2025年医疗人工智能产业虽未实现规模化盈利，但发展势头强劲，2024年中国医疗AI解决方案市场规模达164亿元，预计2030 年将扩至353亿元，CAGR为13.63%。产业发展受资本、政策与医生三向驱动，大模型技术突破降低了应用门槛，头部医院纷 纷部署并参与专科模型研发，政策层面也出台多项举措明确发展路径。然而，产业仍面临价值分歧这一核心困局，患者疗效 与科室效益难以平衡，医院支付意愿和能力不足，商业化进程受阻。临床应用上，医疗AI已广泛渗透至胸外科、心内科、骨 科等临床专科及影像科、病理科等临床支撑科室，在诊断辅助、手术规划、流程优化等方面发挥重要作用，提升了诊疗效率 与精准度，降低了并发症发生率。基层医疗领域因政策支持与需求匹配，成为商业化相对成功的场景，AI有效弥补了基层人 才与能力缺口。数据资产化是产业可持续增长的关键破局路径，医疗数据治理智能化迭代降低了研发成本，数据场内交易与 可信数据空间建设推动了数据流通复用。深睿医疗、东软集团、京东健康等企业通过多模态大模型、智能化解决方案等创 新，在临床全流程赋能、科研转化、基层 ...

当各行各业因为经济周期的压力而出现回调时，医疗AI行业在尚未突破规模化盈利的情况下，仍维持着高增速发展。最新市场调研数据显示，2024年中 国解决方案市场共计164亿元，预期于2030年扩大至353亿元，CAGR13.63%。 庞大且增长的市场不仅吸引了更多创业者，亦将带动大量医生进入AI的科学研究与产品研发，当前的医疗AI已不是某个单一学科的造物，而是计算机、 工业、医学等学科的深度融合。 聚焦2025年，医疗AI最显著的变化可归纳于两个要点：大模型的突破式演变与医疗机构的规模化参与。 2025年年初DeepSeek发布的DeepSeek-R1为大模型产业带来。参数高效微调（PEFT）、混合专家架构（MoE）等创新设计支持下，DeepSeek成功降低了大 模型的入场门槛，让医院管理者们开始主动部署基础设施。 动脉智库统计数据显示：截至2025年5月，第三方榜单中国内排名前100的医院已全部完善大模型部署，有38家医院在通用模型的基础上进一步展开研发， 打造出55个符合自身需求的垂直医疗模型，其中22个为专科模型。 同时，医生对于大模型的实操热情也远超传统AI。当深圳市卫健委开始限制医院自行算力采购（算力集中 ...

PART1 课程背景 医疗人工智能 实训营 第一期 当前，医疗AI在影像诊断、精准治疗和药物研发等场景广泛应用，但医疗机构普遍面临跨学科 人才短缺、技术落地难的挑战，制约了医疗智能化转型进程。 上海交通大学医学院联合上海交通大学生物医学工程学院、学生创新中心、附属医院等优势技 术与临床资源，针对人工智能在医疗领域快速渗透的行业趋势，聚焦医生、工程技术人员及科 研人员以及企业界对AI应用技能的迫切需求，推出医疗人工智能实训营。本课程通过系统化实 战训练，精准衔接临床需求与AI技术，为培养复合型人才提供核心支撑。 PART2 课程特色 01 临床痛点精准匹配 01 临床痛点精准匹配 直击医生在诊断、治疗，科研人员在药物研发中的数据处理与模型部署难题，提 供可直接落地的AI技能训练。 02 实战导向的完整课程体系 从AI基础、医学数据处理到临床案例实操，形成"理论-分析-应用-实操"闭环学习 路径。 03 双领域权威协同 医学专家与AI工程师联合设计并授课，内容严格遵循临床规范与技术前沿，保障 专业性与实用性。 PART3 开课信息 PART4 课程安排 图片 第一天：《人工智能基础》 结合医疗场景讲解核心算法原理 ...

当前，医疗人工智能看似重回热潮，但底层逻辑已发生根本转变。"'取代医生'是个伪命题。AI无法独立 开具处方、撰写诊断报告，更不能执行手术操作。医生的价值远不止于解读影像或数据，更贯穿于医生 对患者整体状态的综合观察--从步态、神情到精神状态，这些细微信息共同构成诊断的重要依据。"一 医学的本质是人学。在可预见的未来，人类不会到无人工厂看病--这一观点已成为行业共识。自2017年 国内首款AI医学影像产品问世以来，人工智能在医疗领域的探索不断深入。科技部同年将医疗影像纳 入国家级人工智能平台重点方向，进一步推动了该领域的发展。 多年来，科技企业在医疗领域的探索逐渐呈现出各自不同的路径与边界。但核心始终是回归医疗本质， 聚焦真正影响诊疗效果的关键问题。 医疗AI的理性回归 伴随人口结构变化，医疗系统面临持续压力。人工智能被视为可能带来改变的技术要素之一，但必须明 确其定位。 "未来，AI在医疗领域的应用或将呈现"二八格局"：针对80%的常见疾病，可基于通用模型进行微调， 实现风险防控；针对20%的专病与疑难杂症，则训练高质量的垂直模型，提升辅助诊断精度。在这一过 程中，技术提供方始终是工具的赋能者，而医院、医生与 ...

Core Insights - Yidu Technology reported a total revenue of RMB 358 million for the first half of the 2026 fiscal year, representing an 8.7% year-on-year growth, with adjusted EBITDA doubling to approximately RMB 54 million [1][3] - The company has achieved near breakeven in its financial statements, ahead of management's previous expectations by one year [1] Business Growth - The core business segments showed strong growth, with new order amounts in the big data platform and solutions segment increasing by 19.7%, and the life sciences solutions segment seeing a remarkable 61.1% increase [3] - The company’s AI technology has been implemented in over 30 top-tier hospitals, with the Doctor Copilot achieving nearly 1,000 daily calls per hospital [3] YiduCore Developments - YiduCore, the company's core algorithm engine, processed over 1.3 billion patient visits and nearly 7 billion authorized medical records, covering over 10,000 hospitals [3] - The accuracy of the TNM staging assessment for tumors improved significantly, with T-stage accuracy rising from 58% to 90% and N-stage accuracy from 62% to 80% [4] Industry Recognition - The company received accolades at the 11th China Health Information Processing Conference (CHIP 2025), winning the championship in "Medical NLP Code Automatic Generation Evaluation" and the "Best Paper Award" [5] Business Segments Performance - The AI for Medical segment generated RMB 153 million in revenue, a 14.6% increase year-on-year, providing solutions to 127 top hospitals and 44 regulatory bodies [6] - The AI for Life Science segment achieved RMB 138 million in revenue, supporting the accelerated approval of several innovative drugs [7] - The AI for Care segment reported revenue of RMB 66.67 million, a 30.3% increase, with significant participation in health insurance projects across multiple provinces [8] Health Management Initiatives - The company's diabetes digital therapy expanded to multiple regions, managing nearly 100,000 patients, with a 27.04% increase in fasting blood sugar compliance [9] - The active transaction user count on the health management platform exceeded 22 million [9] Future Outlook - The company aims to deepen the integration of large model technology with real-world scenarios, leveraging its cash reserves and operational efficiency to capture future market opportunities [9]

Core Insights - Recently, the AI-assisted evaluation software for children's hand X-ray images and the CT image triage software for liver focal lesions developed by the company have been officially approved by the National Medical Products Administration of China as Class III medical devices [1] - The company has launched over 100 medical AI products, with 17 applications approved by the National Medical Products Administration of China and 15 applications certified by the FDA, leading globally with 31 applications certified by the EU [1][2] - The company aims to drive medical progress through technological innovation, focusing on clinical needs and practical applications [1][4] Company Overview - Established at the end of 2017, the company is dedicated to integrating advanced AI technology into medical imaging analysis, intelligent diagnostic assistance, and medical data management [2] - The company has achieved significant milestones in the medical imaging field, which is a key area for AI application, due to the global DICOM standard that allows data from different devices to be read by AI [2] Product Development - The company has received 5 additional Class III certificates from the National Medical Products Administration since 2025, totaling 17 certificates [2] - The company has developed over 100 AI products that optimize hospital processes, reduce marginal costs, and transition from single-point AI applications to comprehensive digital solutions [2] Market Penetration - The company's AI products are implemented in over 4,000 hospitals across China, including top-tier hospitals and grassroots healthcare facilities, enhancing the quality of medical services [3] Technological Innovation - The company has launched the YuanZhi medical large model, which integrates text, voice, and visual understanding capabilities, adapting to various medical scenarios [4][5] - The YuanZhi model has produced over 10 medical intelligent agents covering multiple scenarios, including imaging diagnosis and clinical treatment [5] Collaboration and Research - The company collaborates with hospitals to develop intelligent agents capable of detecting multiple diseases from imaging data, showcasing significant advancements in diagnostic efficiency [5][6] - The company completed a Series A financing round of 1 billion yuan, which will accelerate innovation and product deployment [6]

Core Insights - Recently, the AI-assisted evaluation software for children's hand X-ray images and the CT image triage software for liver focal lesions developed by the company have been officially approved by the National Medical Products Administration of China as Class III medical devices [1][2] - The company has launched over 100 medical AI products, with 17 applications approved by the National Medical Products Administration of China and 15 applications certified by the FDA, leading in the EU with 31 CE-certified applications [1][2] - The company aims to continuously drive medical AI innovation and transform technology into practical medical productivity, focusing on clinical needs and application scenarios [1][5] Company Development - The company was established at the end of 2017 in Shanghai and has been dedicated to integrating advanced AI technology into medical imaging analysis, intelligent diagnostic assistance, and medical data management [1][2] - As of now, the company's AI products have been implemented in over 4,000 hospitals across China, including top-tier hospitals and grassroots healthcare facilities, promoting high-quality medical service development [3][4] AI Application in Medical Imaging - The medical imaging sector has become a significant area for AI applications, facilitated by the global DICOM standard, which allows data from different manufacturers and hospitals to be read by AI [2][3] - The company has introduced a multi-modal medical model, "YuanZhi," which integrates text, voice, and visual understanding capabilities, adapting to various medical scenarios [3][4] Technological Advancements - The company has developed intelligent models capable of precise detection of multiple diseases in imaging diagnostics, utilizing advanced technologies like Transformer and multi-modal techniques [4][5] - The company completed a Series A financing round of 1 billion yuan, which will accelerate innovation and product implementation, focusing on both horizontal and vertical innovations in AI applications [5]