Ademetionine 1,4-Butanedisulfonate

在中国生物医药风起云涌的创新浪潮中，十年，足以见证一个领域从萌芽、生长到蜕变的完整过程。 2015年，本土创新随着药政改革如火如荼。自浙江大学博士毕业并在制药公司历练数年之后，郎国竣便前瞻性地锚定创新药产业链最上游的源头环节，领衔创立了三优生物。 十年后的今天，站在新起点上，郎国竣感慨：“时间过得真快，感觉只是眨眼之间。” 然而，正是这“眨眼之间”，三优生物实现了跨越式发展：从最初仅有百亿级抗体库和100平方米的实验场地的初创企业，迅速成长为行业引领者。如今，三优生物已建成智能超万亿分子库，打造了超万亿规模、智能化与一体化结合的研发平台，投产和布局的研发及GMP场地达20000多平方米。 完成自身积淀后，2026年，三优生物将进一步发布“原创新药创新工场”平台，其核心是整合全球创新药研发资源，利用智能超万亿分子库，打造从药物发现到临床前研发，再到构建原创新药网络生态圈，并与生态圈各方协同践行“让天下没有难做的创新生物药”的使命。 郎国竣坦言，十年前并未预料到三优生物能发展到今天的规模，他的初心仅是“尽己所能为新药研发做点事”。这或许说明，当人坚持做一件对的事，时间自会给予意想不到的回报。  郎国竣  三优生物创始人兼CEO TONACEA 01 10年，构建10万亿级分子库 创立三优生物前，郎国竣已经深耕医药行业多年。 他先是供职于复星医药旗下复宏汉霖，历经研发、质控等岗位；此后，又转身投入另一家公司，从事产品和营销方面的工作。这种横跨“研发-生产-市场”的实践经验，赋予他审视产业的多元视角与全局思维。 得益于如此积累，郎国竣敏锐地捕捉到2015年前后中国生物医药产业链中的一个关键“断层”。 “当时，行业的商业化和工艺生产，包括临床，都已有了不错的基础，但在最源头的药物发现与临床前研发环节，却极度缺乏平台型公司。”郎国竣回忆说。 彼时，中国医药行业正从“仿制”向“创新”艰难转身，大多数人仍在观望，郎国竣却看到未来十年创新的核心瓶颈与巨大机遇——解决药物发现的源头难题，正是平台型公司最根本的价值所在。 回首创新药跌宕起伏的十年，无数公司崭露头角又黯然离场，三优生物却步步为营，走得稳健。 郎国竣总结，三优生物之所以能安全穿越周期，最重要的原因是始终坚持“研发创新”与“商业运营”双轮驱动，避免单一模式的脆弱性。他观察到：“很多公司只埋头研发，一旦资金链断裂，再好的项目也只能搁置。” 这一“双轮驱动”的策略，在近年资本寒冬中展现出惊人的韧性。郎国竣分享：“我们在2024年已实现盈亏平衡，2025年有了一定的利润基础，现金流持续为正。这不仅确保我们能活下去，而且能健康地活下来。” 生存下来，才有资格谈论未来。不过，让三优生物真正站稳脚跟的终极竞争力，是其坚实的技术护城河。 三优生物用十年的时间，构建了10万亿级别的智能超万亿分子库，这是三优生物竞争力的核心基石。 当前，国内主流分子库的库容多在百亿级（10^10）至千亿级（10^11）之间。三优生物的分子库达到十万亿级（10^13），库容是其他分子库的100至1000倍。 “一个百亿级库或许只能筛出几十个候选分子，”郎国竣解释道，“而十万亿级的库，理论上可筛选数千甚至上万个先导分子。这意味着，面对任何靶点，我们都能提供丰富的优质候选分子，从源头上极大提升了后续研发的成功率。” 但十万亿仍然不是终点，三优生物未来还将继续升级扩容分子库。 借助AI技术，新的分子库将告别早期库的随机性，实现对入库分子的“理性设计”，主动筛选分子、构建最优序列，让库更加智能、高效。三优生物还计划在分子形式上进行升级，从最初的全长抗体库，到领先的单域抗体库，再到目前已达万亿级别的多肽库，不断突破分子形式的边界。 未来，三优生物更计划探索将多肽与核酸结合，构建下一代分子库。 TONACEA 02 创新的范式革命 提出“原创新药创新工场”构想，致力于构建行业生态圈，这背后的底气，远不止于三优生物技术层面的超大容量分子库。 过去十年间，三优生物为超过2000家客户成功交付两千多个项目，建立了可靠且深度的客户网络，这是支撑起其战略转型的坚实基石。 这一转型的核心，是从传统CRO的单向服务关系，向一个开放、高效、互利共赢的生态网络跃迁。 “原创新药创新工场”旨在打造一个开放式平台，通过整合上下游产业资源，为客户提供生态化的综合解决方案。其根本价值在于深度介入并系统性地赋能客户的创新进程，从而提升合作伙伴的长期核心竞争力。 这种“深度赋能”在实践中体现为多层级的价值创造，其中最直接的层面是充当资源与信任的枢纽。 郎国竣与写意君分享了一个案例。一家初入中国的美国生物技术公司急需对接本地CDMO和安评机构，却因不熟悉市场而进展缓慢。 三优生物凭借其深厚的本土市场认知与合作网络，不仅为其精准匹配、介绍了合作伙伴，更全程陪同考察、协助谈判，最终确保项目成功落地。 这个过程中，三优生物为客户提供的已远非单纯的技术服务，而是弥合信息差、建立信任关系的综合赋能。 更进一步来说，通过与领先CDMO等战略伙伴构建的深度绑定、互为背书的生态关系，三优生物嵌入一个强大的产业信用网络：战略伙伴为其专业技术背书，而三优生物则保障其推荐客户与项目的质量，从而更高效地触达优质资源，形成良性循环。 在此基础上，“创新工场”扮演着更为关键的双重赋能角色，系统性连接创新链条上的关键断点。 一方面，“创新工场”可以成为科研成果转化的关键推手。 很多高校实验室手握具有潜力的早期靶点，却苦于转化无门；而制药企业的精力也不足以逐一接触、评估、孵化这些早期项目。三优生物的卡位，就是要成为连接上游创新与下游开发的可靠桥梁。 “三优生物拥有干湿结合的一体化智能平台与上万平米研发基地及配套设施，服务链条涵盖原材料制备、抗体筛选，到抗体工程、体外药效、动物模型，再到细胞株构建、纯化工艺开发、制剂工艺开发及质量控制的全流程。”郎国竣指出，“我们虽不做大规模生产，但已构建起从前端发现到临床前研究的完整闭环能力。” 另一方面，“创新工场”还可以成为深度连接资本与产业的“红娘”。 凭借在大量早期研发中积累的深度认知，郎国竣表示，三优生物能高效地为投资机构筛选、对接优质项目，显著降低尽调成本与信息风险。在完成投资决策后，三优生物还可以继续提供专业的研发支持，为项目“护航”，实现从对接到落地的全过程赋能。 由此可见，“原创新药创新工场”正是三优生物将十年积累的技术、经验与网络，升维为动态生态系统的战略体现。 郎国竣希望，三优生物不再仅是技术供应商，而是串联学术界、产业界与资本界的核心枢纽，通过提供资源对接、信任背书与全链条技术支持，系统性地推动整个新药研发生态向更开放、高效、共赢的方向演进。 TONACEA 03 提前探秘“创新工场”：如何一键打通研发资源 2026年，“原创新药创新工场”平台正式上线。 该平台以集成化门户网站的形式，将超过70个专项服务平台或产品页面汇聚于统一界面，使客户得以通过“一个入口”实现“一站获取”所需的技术服务或产业资源，极大降低多供应商管理的复杂性与成本。 同时，“创新工场”借助oneClick+在线设计平台、CRS商城等工具，将深度的专家经验转化为标准化、易用的产品与服务，从而显著缩短项目启动周期。 据郎国竣介绍，首批上线的39个服务模块中，以三优生物自有的全链条技术平台为主，下一步，“原创新药创新工场”将选择性引入并深度整合外部战略合作伙伴的技术平台。 三优生物会为合作伙伴提供免费的产品定义、界面设计与内容重塑服务，旨在将不同来源的技术模块标准化，确保用户获得结构清晰、体验流畅、视觉统一的使用体验。 作为所有合作的枢纽，“创新工场”将成为所有项目交互数据的天然沉淀池，形成全球稀缺的多模态生物医药研发数据资产，为未来训练专属的AI驱动药物发现算法与预测模型提供了独一无二的基础，构筑起长远的竞争壁垒。 “创新工场”平台的生态价值主张体现在两个层面。它首先是一个开放创新平台，通过“前沿创新”模块持续汇聚并输出深度行业洞见，旨在吸引并留住高价值客户，树立行业思想领袖品牌形象。 在此基础上，平台致力于激发价值网络效应，系统连接客户、供应商、科学家与创新者，促进生态内的高效协作。其长期愿景是，通过这种深度连接与信任机制，未来可在平台上直接撮合合作，解决传统产业链中的信任与效率瓶颈，进而将协作关系沉淀为可衡量的“品牌协同”资产，最终形成一种全新的平台化、网络协同商业模式。 TONACEA 04 “不去想20年后，先把2026年做到最好” 站在下一个十年的新起点，郎国竣坦言此刻的心境是“激情澎湃”与“战战兢兢”复杂交织。 “10年，我们走到今天相当不容易，”郎国竣说，“我们从最初的3个人发展到现在的200多人，这十年的努力为公司打下了特别好的发展基础。” “战战兢兢”则源于他对企业生存规律的清醒认知和敬畏感。“95%的企业活不过10年，存活超过15年的企业更是少之又少。要想维持企业的基业长青，非常艰难。” 正是这份深刻的危机感，让郎国竣始终保持冷静与务实。“先不去想15年、20年后会怎样，我们首先要把2026年做到最好的状态。” 未来，三优生物将继续坚持“研发创新”与“商业运营”双轮驱动的核心战略，在短期增长的迫切性与长期可持续发展之间寻求平衡。 在郎国竣的构想中，三优生物会成为中国新药研发的重要策源地之一，而驱动这一切的根本动力，始终是他创业之初那份朴素的使命感。 “让天下没有难做的创新生物药。” Sanyou Bio’s Decade: From an Intelligent Super-Trillion Antibody Library to the “Innovation Hub for Original Drug Discovery” Amid the surging wave of innovation in China’s biopharmaceutical industry, a decade is long enough to witness the full evolution of a sector—from its initial emergence and growth to fundamental transformation. In 2015, as regulatory reform ignited local innovation, Lang Guojun, after earning his PhD from Zhejiang University and gaining several years of experience in pharmaceutical companies, foresightedly focused on the core upstream segment of the industry value chain and founded Sanyou Biopharmaceuticals (hereinafter referred to as “Sanyou Bio”). Standing at a new starting point ten years later, Lang Guojun reflects, “Time has passed incredibly fast—it feels like just the blink of an eye.” Yet within this “blink of an eye,” Sanyou Bio has achieved a leapfrog development,rapidly evolving from a startup with an initial antibody library at the 10^10 scale and a laboratory space of just 100 square meters into an industry front-runner. Today, Sanyou Bio has established an Intelligent Super-Trillion Antibody Library (STAL), the R&D platforms of super-trillion, intelligence and integration, as well as more than 20,000 square meters of operational and planned R&D and GMP-compliant facilities. Building on this foundation, in 2026, Sanyou Bio will officially launch its Innovation Hub for Original Drug Discovery (ClickLinks). At its core, the platform integrates global innovative drug R&D resources and leverages the Intelligent Super-Trillion Antibody Library to deliver a one-stop solution spanning drug discovery, preclinical development, and industrialization, while building an ecosystem network for original drug innovation. in collaboration with ecosystem partners, the platform will jointly fulfill the mission of “To provide solutions to accelerate biological drugs innovation globally.” Lang Guojun admits that a decade ago, he never imagined Sanyou Bio would reach its current scale. His original intention was simply “to do whatever I can to contribute to new drug development.” Perhaps this illustrates a simple truth: when one persists in doing the right thing, time often delivers rewards beyond expectation. TONACEA 01 A Decade to Build a 10-Trillion–Scale Molecular Library Before founding Sanyou Bio, Lang Guojun had already spent many years in the pharmaceutical industry. He first worked at Henlius, a subsidiary of Fosun Pharma, holding positions across R&D and quality control. He later joined another company, focusing on product and marketing functions. This rare combination of experience across R&D, manufacturing, and commercialization endowed him with a multidimensional and holistic view of the industry. Thanks to this background, Lang Guojun keenly identified a critical “gap” in China’s biopharmaceutical value chain around 2015. “At that time, commercialization, process development, and even clinical capabilities had already made solid progress,” Lang Guojun recalls. “But at the very source—drug discovery and preclinical research—there was a severe shortage of platform-based companies.” As China’s pharmaceutical industry struggled to transition from generics to innovation, most players held a wait-and-see attitude. Lang Guojun, however, recognized both the core bottleneck and the greatest opportunity of the next decade: solving the fundamental challenges of drug discovery, which he believed represented the true value of platform companies. Looking back at the turbulent decade of innovative drug development, countless companies emerged and then faded away. Sanyou Bio, by contrast, advanced steadily and prudently. According to Lang Guojun, the company’s ability to navigate industry cycles safely stems from its unwavering commitment to a dual-engine strategy of “R&D innovation” and “commercial operations,” avoiding the fragility of a single-track model. “Many companies focus solely on R&D,” he observes. “Once their funding dries up, even the best projects come to a halt.” This dual-engine strategy has demonstrated remarkable resilience amid the recent capital crunch in the industry. “In 2024, we achieved break-even; in 2025, we built a solid profit base with sustained positive cash flow,” Lang Guojun shares. “This not only ensures survival, but healthy survival.” Survival is merely the prerequisite for the future. What truly anchors Sanyou Bio’s position is its formidable technological moat. Over ten years, Sanyou Bio has built a 10-trillion–scale Intelligent Super-Trillion Antibody Library, which serves as the cornerstone of its competitiveness. Today, most domestic molecular libraries range from 10^10 to 10^11 in size. Sanyou Bio’s library, at 10^13, is 100 to 1,000 times larger. “A 10^11-scale library may yield only dozens of candidate molecules,” Lang Guojun explains. “A 10^13-scale library can theoretically generate thousands or even tens of thousands of lead candidates. This means that for any target, we can provide a rich pool of high-quality candidates, significantly increasing downstream R&D success rates from the very beginning.” Yet even 10 trillion is not the endpoint. Sanyou Bio plans to continue expanding and upgrading its libraries. With the integration of AI technologies, next-generation libraries will move beyond randomness toward rational design, proactively selecting molecules and constructing optimal sequences to make libraries smarter and more efficient. In parallel, Sanyou Bio continues to expand molecular modalities—from full-length antibody libraries, to leading single-domain antibody libraries, and now trillion-scale peptide libraries—constantly pushing the boundaries of molecular forms. Looking ahead, the company also plans to explore the integration of peptides with nucleic acids to create the next generation of molecular libraries. TONACEA 02 A Paradigm Shift in Innovation The confidence behind the Innovation Hub for Original Drug Discovery and its ambition to build an industry ecosystem extends far beyond Sanyou Bio’s ultra-large molecular libraries. Over the past decade, Sanyou Bio has successfully delivered more than 2,000 projects for over 2,000 clients, building a deep and reliable customer network that underpins its strategic evolution. At the heart of this transition is a shift from traditional one-way CRO services to an open, efficient, and mutually beneficial ecosystem network. The Innovation Hub is designed as an open platform that integrates upstream and downstream resources to provide ecosystem-based comprehensive solutions. Its fundamental value lies in deeply engaging with and systematically empowering clients’ innovation processes, thereby enhancing partners’ long-term core competitiveness. This “deep empowerment” manifests across multiple layers of value creation, beginning with Sanyou Bio’s role as a hub of resources and trust. Lang Guojun shares an example: a U.S.-based biotech company newly entering China urgently needed to connect with local CDMOs and safety evaluation institutions but faced delays due to unfamiliarity with the market. Leveraging its deep local insights and partnership network, Sanyou Bio not only precisely matched and introduced suitable partners, but also accompanied site visits and supported negotiations, ultimately ensuring successful project implementation. In this process, Sanyou Bio provided far more than technical services—it bridged information gaps and established trust. Moreover, through deep, mutually endorsed ecosystem relationships with strategic partners such as leading CDMOs, Sanyou Bio has embedded itself within a powerful industrial credit network. Strategic partners vouch for Sanyou Bio’s technical expertise, while Sanyou Bio ensures the quality of recommended projects and clients, enabling more efficient access to premium resources and forming a virtuous cycle. On this foundation, the Innovation Hub assumes a dual empowerment role by systematically connecting critical breaks along the innovation chain. On one hand, it serves as a catalyst for translating scientific discoveries into development programs. Many academic laboratories hold promising early-stage targets but lack pathways for translation, while pharmaceutical companies often lack the bandwidth to evaluate and incubate such early projects individually. Sanyou Bio positions itself as a reliable bridge between upstream innovation and downstream development. “Sanyou Bio has built a dry–wet integrated intelligent platform, over 10,000 square meters of R&D facilities, and a comprehensive service chain spanning raw material preparation, antibody screening, antibody engineering, in vitro pharmacology, animal models, cell line development, purification process development, formulation development, and quality control,” Lang Guojun notes. “While we do not engage in large-scale manufacturing, we have established a complete closed-loop capability from discovery through preclinical research.” On the other hand, the Innovation Hub also functions as a matchmaker connecting capital and industry. Drawing on deep insights accumulated through extensive early-stage R&D involvement, Sanyou Bio can efficiently help investors screen and connect with high-quality projects, significantly reducing due diligence costs and information risks. After investment decisions are made, Sanyou Bio continues to provide professional R&D support, safeguarding projects from matchmaking through execution. In this sense, the Innovation Hub represents Sanyou Bio’s strategic elevation of a decade of accumulated technology, experience, and networks into a dynamic ecosystem. Lang Guojun envisions Sanyou Bio evolving beyond a technical service provider into a central hub linking academia, industry, and capital—driving the new drug R&D ecosystem toward greater openness, efficiency, and shared success through resource connectivity, trust endorsement, and end-to-end technical support. TONACEA 03 A First Look at the Innovation Hub: One-Click Access to R&D Resources In 2026, the Innovation Hub for Original Drug Discovery will officially go live. Presented as an integrated portal, the platform consolidates over 70 specialized service platforms or product pages into a unified interface, enabling clients to access required technical services or industrial resources through a single interface, dramatically reducing the complexity and cost of managing multiple vendors. At the same time, tools such as the OneClick+ Online Design Platform and the CRS marketplace, transform deep expert knowledge into standardized, user-friendly products and services, significantly shortening project initiation cycles. According to Lang Guojun, of the first 39 service modules launched, most are based on Sanyou Bio’s proprietary end-to-end technology platforms. Moving forward, the Innovation Hub will selectively introduce and deeply integrate technology platforms from external strategic partners. Sanyou Bio will provide partners with free product definition, interface design, and content restructuring services, aiming to standardize technology modules from diverse sources and ensure a clear structure, seamless experience, and unified visual identity for users. As the central hub for all collaborations, the Innovation Hub will naturally accumulate interaction data from all projects, forming a globally scarce multimodal biopharmaceutical R&D data asset. This will lay a unique foundation for training proprietary AI-driven drug discovery algorithms and predictive models, establishing a long-term competitive moat. The platform’s ecosystem value proposition operates on two levels. First, it serves as an open innovation platform that continuously aggregates and disseminates deep industry insights through its “Frontier Innovation” module, aiming to attract and retain high-value clients while building thought leadership. Building on this, the platform aims to unlock network value effects by systematically linking clients, suppliers, scientists, and innovators to promote efficient collaboration. Its long-term vision is that, through deep connectivity and trust-based mechanisms, the platform will ultimately enable direct partnership matchmaking, address long-standing trust and efficiency bottlenecks in the traditional value chain, and transform collaborative relationships into measurable “brand synergy”assets—culminating in a brand-new platform-based, network-collaborative business model. TONACEA 04 “Don’t Think About 20 Years Ahead—First Make 2026 the Best It Can Be” Standing at the threshold of the next decade, Lang Guojun describes his current mindset as a complex mix of “excitement” and “cautious vigilance.” “Reaching where we are today has not been easy,” he says. “We’ve grown from a team of three to more than 200 people. These ten years have laid a very solid foundation for the company.” The caution stems from his clear understanding of business survival dynamics. “Ninety-five percent of companies don’t survive beyond ten years, and even fewer last more than fifteen. Sustaining long-term success is extremely difficult.” This acute sense of crisis keeps Lang Guojun grounded and pragmatic. “Rather than thinking about what things will look like in 15 or 20 years, our priority is to make 2026 the best version it can be.” Going forward, Sanyou Bio will continue to uphold its dual-engine strategy of R&D innovation and commercial operations, striving to strike a balance between short-term growth imperatives and long-term sustainable development. In Lang Guojun’s vision, Sanyou Bio will become one of China’s key sources of innovative drug development—and the fundamental driving force behind it all remains the simple mission that inspired him at the very beginning: “To provide solutions to accelerate biological drugs innovation globally.” 关于三优生物 三优生物是一家以“让天下没有难做的创新生物药”为使命，以超万亿分子库和智能科技驱动的生物医药高科技企业。 公司致力于打造全球顶尖的原创新药创新工场。公司以智能超万亿分子库（AI-STAL）为核心；以干湿结合、国际领先的创新生物药智能化及一体化研发平台为依托；以多样化的业务模式推动全球创新药物的研发及产业化。 公司总部位于中国上海，在亚洲、北美洲、欧洲等多地建立了业务中心，形成了全球化的业务网络，现有投产及布局的研发及GMP场地20000多平方米。 公司已与全球2000多家药企、生技公司等建立了良好的合作关系，已赋能1200多个新药研发项目；已完成50多个合作研发项目，其中10多个协同研发项目已推至IND及临床研发阶段。 公司已申请130多项发明专利，其中30多项发明专利已获得授权，并获得了国家级高新技术企业、上海市专精特新、ISO9001、ISO27001等10余项资质及体系认证。 推荐阅读 不用部署模型，也能做蛋白结构预测？ 盘点诺奖技术“噬菌体展示”孕育的抗体药物 三优十周年｜平台篇-抗体工程平台 三优十周年｜平台篇-体外药效平台 特应性皮炎：从免疫紊乱到精准靶向的治疗革命 三优十周年｜抗衰老-从长生神药到系统与精准干预 战略跃迁｜打造全球顶尖的原创新药创新工场 三优十周年｜恶病质的代谢紊乱与靶向药物的研发态势 三优十周年｜中国药企撬动千亿美元赛道 三优十周年｜平台篇-创新抗体产生平台 三优十周年｜眼科治疗前沿：用抗体药物精准点亮未来 三优十周年｜磁阵列全人源小鼠抗体发现平台 三优十周年｜新“药王”加冕：减重增肌 三优十周年｜抗体药物如何实现精准“狙击”炎症？ 三优十周年｜血液肿瘤治疗革命-精准分层治疗新时代 三优十周年｜抗体药免疫原性全方位解析

点击蓝字，关注我们 放射科主要依靠低年资医生先读片、高年资医生再复核，需要花费大量人力和时间，因此乳腺癌X线成像筛查长期以来一直是人工智能的目标应用案例。21世纪初以来，计算能力的飞速发展、数字化乳腺X线成像技术的普及，生成大量适用于人工智能算法训练和验证的存档影像数据库，乳腺X线成像筛查读片识别可疑病变这项劳动密集型工作成为人工智能自动化的典型应用，有望提高筛查效率并同时减轻筛查工作负担。过去十年，人工智能算法的快速发展已被见证。由于早期采用非现代人工智能软件进行乳腺X线成像读片的效果令人失望，最新人工智能深度学习模型迅速发展，为循证实施提供了依据：从根据乳腺癌大数据库分析人工智能的准确性，到采用连续参加筛查者有代表性定群报告筛查指标的回顾研究，以及目前首个将人工智能加入全国10万女性乳腺癌筛查项目的随机对照试验：MASAI研究。 MASAI (NCT04838756): Mammography Screening With Artificial Intelligence Official Title: A Randomized, Single-blinded, Controlled Trial on the Efficacy of Mammography Screening With Artificial Intelligence - the MASAI Study 2023年8月，MASAI研究中期分析报告发表于英国《柳叶刀》肿瘤学分册，结果表明人工智能可将放射科医生读片工作量减少44%。2025年3月，英国《柳叶刀》数字健康分册发表MASAI研究另一项早期分析发现，人工智能辅助乳腺癌筛查检出率提高29%，且假阳性率并未增加。那么，人工智能再聪明，能比得过两位资深放射科医生一起读片的双保险吗？能否减少两次筛查之间发生或者上次按计划筛查后两年内诊断出筛查时漏诊的间隔期乳腺癌？ 2026年1月29日，国际四大医学期刊之首、创刊203年的英国《柳叶刀》正刊在线发表荷兰拉德堡德大学医学中心、瑞典隆德大学斯科讷医院、挪威癌症登记中心、挪威北极圈大学、丹麦大学王国医院的MASAI研究主要分析报告，发现人工智能已经悄悄超越人类专家双人读片模式。 该全国多中心人群筛查准确性非劣效单盲随机对照试验于2021年4月12日至2022年12月7日将瑞典西南部四个筛查点10万5934例女性按1比1随机分入两组：干预组5万3052例（年龄中位53.8岁，四分位46.5至63.3岁）采用荷兰拉德堡德大学医学中心研发的人工智能系统辅助乳腺X线成像筛查，对照组5万2882例（年龄中位53.7岁，四分位46.5至63.2岁）采用标准双人读片。人工智能用于将检查结果按低风险（多数）或高风险分流至放射科医生单人或双人读片并标记可能异常区域及其评分。本次对研究方案预设主要结局指标间隔期乳腺癌发生率进行分析，非劣效界值为20%。次要结局指标包括间隔期乳腺癌特征、真阳性率、真阴性率以及按年龄、乳腺密度和癌症类型（原位癌和浸润癌）划分的真阳性率。 结果，干预组和对照组相比： 间隔期乳腺癌：1.55‰比1.76‰（95%置信区间：1.23～1.92、1.42～2.15） 非劣效比例比：0.88（95%置信区间：0.65～1.18，P=0.41） 间隔期浸润癌：75例比89例 乳腺癌≥2厘米：38例比48例 三阴性等非管腔A型：43例比59例 真阳性率：80.5%比73.8%（95%置信区间：76.4～84.2、68.9～78.3，P=0.031）该差异不受年龄和乳腺密度的影响，仅对浸润癌有效，对原位癌无效 真阴性率：98.5%比98.5%（95%置信区间：98.4～98.6、98.4～98.6，P=0.88） 因此，该研究结果表明，人工智能辅助乳腺X线筛查与标准双人读片相比，筛查结果不但非劣效，而且略胜一筹，包括间隔期乳腺癌较少、间隔期乳腺癌不良特征病例较少、真阳性率较高、真阴性率相同，同时还减轻读片工作量，几乎解放将近一半人力，能够有效提高筛查效率，值得考虑在其他国家临床实践中推广应用。 对此，澳大利亚悉尼大学发表同期评论：人工智能对提高人群乳腺癌筛查效率和效果的应用。 Lancet. 2026 Jan 29;407(10527):505-514. IF: 88.5 Interval cancer, sensitivity, and specificity comparing AI-supported mammography screening with standard double reading without AI in the MASAI study: a randomised, controlled, non-inferiority, single-blinded, population-based, screening-accuracy trial. Jessie Gommers, Veronica Hernstrom, Viktoria Josefsson, Hanna Sartor, David Schmidt, Annie Hjelmgren, Anna-Maria Larsson, Solveig Hofvind, Ingvar Andersson, Aldana Rosso, Oskar Hagberg, Kristina Lang. Radboud University Medical Centre, Nijmegen, Netherlands; Lund University, Malmo, Sweden; Skane University Hospital, Malmo, Sweden; Lund University, Lund, Sweden; Cancer Registry of Norway, Oslo, Norway; The Arctic University of Norway, Tromso, Norway; Rigshospitalet, Copenhagen, Denmark. BACKGROUND: Evidence indicates that artificial intelligence (AI) can improve mammography screening by increasing cancer detection and reducing screen reading workload, but its effect on interval cancers (primary breast cancers diagnosed between two screening rounds or within 2 years after the last scheduled screening that were not detected at screening) is unknown. We aimed to compare the interval cancer rate in AI-supported mammography screening with standard double reading without AI. METHODS: In this Swedish randomised, controlled, non-inferiority, single-blinded, population-based screening accuracy trial, participants were allocated in a 1:1 ratio to either AI-supported mammography screening (the intervention group) or standard double reading without AI (the control group). AI was used to triage examinations to single or double reading by radiologists and for detection support. This is a protocol-defined analysis of the primary outcome, interval cancer rate, with a 20% non-inferiority margin. Secondary outcomes reported in this analysis are interval cancer characteristics, sensitivity, specificity, and sensitivity by age, breast density, and cancer type (in-situ and invasive). Other secondary outcomes from the trial that have been previously reported are referenced in the Methods section of this Article. The trial is registered with ClinicalTrials.gov (NCT04838756) and is complete. FINDINGS: Between April 12, 2021, and Dec 7, 2022, 105934 women were randomly assigned to the intervention or control group, of whom 19 were excluded from the analysis. Median age was 53.8 years (IQR 46.5-63.3) in the intervention group and 53.7 years (46.5-63.2) in the control group. Interval cancer rates were 1.55 (95% CI 1.23-1.92) and 1.76 (1.42-2.15) per 1000 participants in the intervention and control group respectively, a non-inferior proportion ratio of 0.88 (95% CI 0.65-1.18; p=0.41). Descriptively, the intervention group had fewer interval cancers that were invasive (75 vs 89), T2+ (38 vs 48), or non-luminal A (43 vs 59) than the control group. Sensitivity was higher in the intervention group (80.5% [95% CI 76.4-84.2]) than the control group (73.8% [68.9-78.3]; p=0.031), an effect consistent across age and breast density, and for invasive cancer but not for in-situ cancer. Specificity was 98.5% (95% CI 98.4-98.6) for both groups (p=0.88). INTERPRETATION: AI-supported mammography screening showed consistently favourable outcomes compared with standard double reading, with a non-inferior interval cancer rate, fewer interval cancers with unfavourable characteristics, higher sensitivity, and the same specificity, while also reducing screen reading workload. These findings imply that AI-supported mammography screening can efficiently improve screening performance compared with standard double reading and may be considered for implementation in clinical practice. FUNDING: Swedish Cancer Society, Confederation of Regional Cancer Centres, Swedish governmental funding for clinical research. DOI: 10.1016/S0140-6736(25)02464-X Lancet. 2026 Jan 29;407(10527):471-473. IF: 88.5 AI for enhancing efficiency and effectiveness of population breast cancer screening. M Luke Marinovich, Nehmat Houssami. The Daffodil Centre, University of Sydney, Cancer Council NSW, Sydney, NSW, Australia; School of Public Health, Faculty of Medicine and Health, University of Sydney, Sydney, NSW, Australia. Breast cancer screening has long been a target use case for artificial intelligence (AI). Developments in computational power coincided with the adoption of digital mammography in the early 2000s and the generation of large, data-rich, archival imaging datasets suitable for algorithm training and validation. As such, the labour-intensive task of mammography screen-reading to identify suspicious lesions was a model application of AI automation, with potential for improving screening effectiveness and simultaneously reducing screening workload burden. Consequently, the last decade has seen rapid growth in AI algorithm development. Against a backdrop of disappointing outcomes from early adoption of pre-contemporary AI software for mammographic interpretation, evaluation of more recent, deep learning-based AI models has progressed swiftly to inform evidence-based implementation: from studies of AI accuracy in selected cancer-enriched datasets, to retrospective studies reporting screening metrics using representative cohorts of consecutive participants, and now the first randomised controlled trial (RCT) of AI embedded into a population breast cancer screening programme. DOI: 10.1016/S0140-6736(26)00092-9 （来源：SIBCS） 声 明 凡署名原创的文章版权属《肿瘤瞭望》所有，欢迎分享、转载。本文仅供医疗卫生专业人士了解最新医药资讯参考使用，不代表本平台观点。该等信息不能以任何方式取代专业的医疗指导，也不应被视为诊疗建议，如果该信息被用于资讯以外的目的，本站及作者不承担相关责任。