Iterative Health
FDA 许可的 GI AI 叠加多专科研究中心网络规模,但收入不透明,2026 年独角兽估值仍难支撑
Iterative Health 在网络表现和战略相关性上有罕见强的公开证据,但当前独角兽估值仍需要收入和利润率证据支撑;这些数据尚未披露。
封面要素
公司概况
Iterative Health 是一家位于 Cambridge 的医疗技术与服务公司,最早从 Iterative Scopes 面向 GI 的 AI 工具演进而来,如今定位为更宽的临床研究基础设施平台。公司把 FDA 许可的实时息肉检测产品 SKOUT,与服务胃肠病学、肝病学、肥胖和心脏病学研究的多专科研究中心网络模式结合起来。公开证据能支撑其运营动能:100+ 个研究中心、40+ 个申办方/CRO/器械合作伙伴,以及 IBD 试验执行中有记录的速度优势。但公司尚未披露足以仅凭基本面清晰支撑独角兽估值的收入质量或利润率结构。
- 成立时间
- 2017-01-01
- 创始人
- Jonathan Ng
- 创立地点
- Cambridge, MA, USA
- 总部
- Cambridge, MA, USA
- 产品
- 产品栈把 SKOUT——一款 FDA 许可、用于结肠镜的实时 AI CADe 工具——与以研究中心为核心的临床研究运营平台配在一起,帮助医疗服务机构和申办方筛选、启动、配置人员并招募高表现试验中心。配套 AI 资产支持 IBD 试验招募、内镜评分,以及 AI 与人工结合的审阅流程。
- 客户
- 医疗服务网络、GI/肝病学/心脏病学研究中心、药企申办方、生物技术公司、医疗器械公司,以及希望更快在社区场景执行试验的 CRO。
- 商业模式
- 服务与软件混合模式:来自申办方/CRO 和医疗服务合作伙伴的研究中心网络与试验运营收入,加上 SKOUT 及配套 AI 流程工具的产品变现。公开价格和合同条款大多未披露。
- 阶段
- Series C
- 融资情况
- 2026 年 4 月官方披露完成 $77M Series C;公开估值背景约为 $1.3B-$1.4B,累计披露资本约 $270M+,但收入、利润率和完整优先股堆叠的可见度仍有限。
执行摘要
主要优势
- 站点网络跑赢同行已有记录:公开研究和公司公告反复显示,其启动速度更快,入组表现明显好于已发表 IBD 基准。
- 双资产定位把 FDA 批准的 GI AI 设备和嵌入式研究网络运营绑在一起,比纯单点方案厂商更能控制工作流。
- 投资人与合作伙伴质量高:GV、Intrepid、Insight、GI Alliance、One GI 和 U.S. Heart & Vascular 支撑分发、可信度和品类入口。
主要风险
- 收入、毛利率、烧钱速度和留存仍未披露,当前私有估值缺少公开市场投资人会要求的核心运营指标。
- AI 内镜报销、成本效益和真实世界泛化能力仍未解决,限制了今天能计入多少 SKOUT 上行情景。
- Iterative 从 GI 扩到心脏病学和肥胖症时,还要在 100+ 站点网络和多个锚定合作体系里守住表现,执行风险随之上升。
未决问题
- 当前收入 / ARR、毛利率、运营烧钱速度和现金跑道没有公开披露。
- 头部客户集中度、续约行为、合同期限和申办方组合没有公开披露。
- Series C 完整优先股堆叠细节、稀释条款和任何老股成分,公开渠道看不到。
目录
01公司概况
1.1 身份、定位与运营模式
Iterative Health 成立于 2017 年,如今对外呈现为一家医疗技术与服务公司,总部位于马萨诸塞州 Cambridge 和纽约州 New York。公司当前首页和关于页面给出的框架,已经不再像单一产品 AI 初创公司,而更像多专科临床研究基础设施平台。其运营重点现在覆盖 GI、肝病学、肥胖和心脏病学,并把研究直接嵌入社区护理场景,而不是只集中在学术中心。 这种定位清楚显示出公司从早期 Iterative Scopes 身份向外扩张;早期叙事强调精准胃肠病学、计算机视觉和 SKOUT 等产品。当前模式把医疗服务机构和研究中心接入集中化运营、申办方资源和 AI 驱动的流程工具。落到实践里,Iterative 向申办方和 CRO 出售的是速度、启动可靠性和招募表现;给医疗服务集团的则是研究基础设施、人员支持和患者识别工具。 SKOUT 仍然重要,因为它证明公司保留了真实的商业化器械足迹,而不是只做服务中介。但 2026 年的核心投资叙事更宽:已成规模的研究网络、自有流程技术,以及医疗服务中心和申办方需求之间更紧的连接。后续章节应把这一组合视为公司的真实业务模型。[CO001, CO002, CO003, CO004, CO005, CO008]
| 指标 | 数值 / 状态 | 日期 | 置信度 | 缺口 / 备注 |
|---|---|---|---|---|
| 成立时间 | 2017 | 2017 | 高 | 成立年份得到当前关于页面和早期融资材料交叉印证。 |
| 总部 | Cambridge, MA 和 New York, NY | 当前 | 高 | 公司描述了双总部布局;更广办公室清单未公开列举。 |
| 当前阶段 | Series C 轮 / 后期私有 | 2026 | 中 | 阶段由 Series C 完成支撑,但未披露公开 IPO 文件。 |
| 最新轮次 | $77M Series C | 2026 | 中 | 已正式宣布;公司材料未披露估值。 |
| 最新估值 | $1.3B-$1.4B 隐含估值(二级来源) | 2026 | 低 | 区间来自 TechCrunch、Forge 和 AI2.work,而非公司。 |
| 累计融资 | $271M-$273M 隐含 | 2026 | 中 | 第三方追踪器大体一致,但官方未公布累计总额。 |
| 员工数 | 约 250,公司披露;166,Tracxn 估计 | 2025-2026 | 中 | 当前员工规模方向上清楚,但各来源尚未对齐。 |
| 研究网络规模 | >100 个站点,覆盖多个大陆 | 2026 | 高 | 官方 2026 年发布支持 >100 个站点;首页首屏指标在抓取文本中没有显示精确患者触达数字。 |
| 外部商业交易对手 | 40+ 制药 / 生物技术 / 器械 / CRO 伙伴 | 2026 | 高 | 官方 Series C 发布给出最清楚的赞助方侧规模披露。 |
| 收入 / ARR | 当前 | 低 | 未审阅到公开来源披露当前收入或 ARR。 | |
| SKOUT 监管状态 | 已获 FDA 许可的 CADe 产品 | 2022 / 2025 更新 | 高 | 官方发布和 FDA 510(k) 材料支持其持续监管状态。 |
主要来源为 Iterative Health 发布和 FDA 材料;null 表示审阅过的来源集中未公开披露。
[CO001, CO002, CO011, CO012, CO015, CO017]Iterative Health 将社区医疗服务集团、集中化站点运营、申办方需求和 AI 工具串成一个研究执行模式。
[CO003, CO004, CO005, CO017, CO018, CO039]1.2 创始人、领导层与治理信号
创始人兼 CEO Jonathan Ng 仍是核心战略人物。公开材料一直把他标为创始人/CEO;更早的 Iterative Scopes 融资史也把公司起点连到一个 MIT 孵化项目,该项目围绕 Ng 的医生创业者履历,以及 MIT/Harvard 训练背景搭建。这让公司在医学、AI 和临床试验运营交叉处具备可信的创始人-市场匹配。 2025 年底,领导层厚度明显提升。Bill Kayser 在 GI Alliance、Prospero Health 和 McKesson 担任高级财务职位后加入,出任总裁兼首席财务官。这一点值得注意,因为他带来的不是纯软件财务经验,而是医生执业基础设施和医疗 M&A 运营经验。公开团队材料还显示 Dana Feuchtbaum 担任 COO,Nadege Gunn 担任肝病学和肥胖首席医疗官,进一步强化公司在研究中心扩张和相邻治疗领域进入上的运营重心。 治理可见度仍是局部而非完整。公开融资公告披露了与投资人有关的董事席位或观察员角色演进——包括 Series A 时的 Nan Li 和 Lotus Mallbris、Series B 时的 Lonne Jaffe,以及 Series C 时的 Ajay Agrawal 和 Anthony Philippakis——但没有给出完全对齐的当前董事会名单或所有权地图。因此,治理结论只能是方向性的:创始人主导,机构监督在增强,但完整控制权仍不透明。[CO006, CO007, CO009, CO010, CO013, CO014]
| 人员 | 职位 | 背景 / 职能覆盖 | 创始人-市场匹配或依赖 |
|---|---|---|---|
| Jonathan Ng, MBBS | 创始人 / CEO | 医生创业者;公司起源与 MIT 分拆公司和 MIT/Harvard 开发背景相连 | 高 — 仍是公司的战略门面和融资叙事核心 |
| Bill Kayser | 总裁兼首席财务官 | 曾任 GI Alliance 和 Prospero Health CFO;曾任 McKesson 战略 / M&A 高管 | 中高 — 扩大医疗服务方合作的关键财务 / 运营招聘 |
| Dana Feuchtbaum | 首席运营官 | 与 DDW 2026 网络表现披露相关的公开运营高管 | 中 — 对站点网络执行很关键,但相比 CEO 职位,创始人集中度更低 |
| Nadege Gunn, MD | 首席医疗官,肝病学与肥胖 | 医疗领导职位显示,治疗领域正扩展到 GI 核心之外 | 中 — 强化专科宽度,而非核心公司控制 |
覆盖 2026 年尽调中最影响决策的公开领导层职位,而不是团队页面列出的每一位 VP。
[CO006, CO007, CO009, CO010, CO013, CO014]1.3 融资历史、规模信号与利益相关方地图
公开披露的融资历史足以证明这是一家可信的后期公司,尽管精确估值和股权结构表细节仍有部分不透明。Iterative 在更早的 $13.5M 种子轮之后,于 2021 年完成 $30M Series A;2022 年 1 月又完成由 Insight Partners 和 Clearlake 共同领投的 $150M Series B。2026 年 Series C 再增 $77M,由 Intrepid Growth Partners 和 GV 领投,EDBI 作为新投资人进入,Insight 和 Obvious Ventures 再次参与。这条融资序列支撑了一个持久的机构投资人联盟,也说明公司成功把自己从 AI 驱动的 GI 初创公司,重新定位为更宽的临床研究平台。 2025-2026 年,规模主张也更强。官方新闻稿现在描述的是横跨多个大陆的 100+ 个研究中心、40+ 个药企/生物技术/医疗器械/CRO 合作伙伴,以及约 250 人组织。这些都是有意义的信号,因为业务模型依赖的是运营密度和可重复性,而不只是算法新颖性。GI Alliance、One GI、US Heart & Vascular,以及 NextStage 心脏病学研究中心收购,说明医疗服务网络合作伙伴如今与财务投资人同样具有战略意义。 关键但书在于披露质量。TechCrunch、Forge、AI2.work 和 Tracxn 都暗示累计融资约 $270M,最新估值约在 $1.3B-$1.4B 区间,但这些数字仍是二手估计,因为公司没有公开披露 Series C 估值。第三方追踪平台给出的员工数也不一致。因此,投资人可以承认其动能和投资人联盟质量,但还无法精确判断经济性或所有权结构。[CO009, CO011, CO012, CO015, CO016, CO017]
| 利益相关方 | 关系 | 角色 / 重要性 | 控制权或经济重要性 | 尽调问题 |
|---|---|---|---|---|
| Intrepid Growth Partners | Series C 领投方 | 领投 2026 年增长轮 | 将 Ajay Agrawal 加入董事会,暗示有意义的治理影响力 | 确认持股比例、董事会权利和保护性条款 |
| GV (Google Ventures) | Series C 共同领投 | 支持 AI 赋能工作流逻辑的成长投资方 | Anthony Philippakis 拥有董事会观察员席位,说明可见度较高但没有完整控制权 | 确认信息权和跟投资金能力 |
| EDBI | Series C 新投资方 | 为 2026 年投资团增加新加坡背景战略资本 | 经济重要性清楚;治理角色未披露 | 澄清出资规模和任何地理商业支持协议 |
| Insight Partners | Series B 共同领投方和持续投资人 | 长期软件 / 成长投资方,有董事会角色历史 | Lonne Jaffe 在 Series B 的董事会席位暗示持久治理影响力 | 确认 Series C 后董事会席位是否仍有效 |
| Clearlake Capital | Series B 共同领投方 | 在公司规模化阶段带来跨界成长资本 | 审阅过的 2026 年材料未披露当前董事会角色 | 澄清当前持股、优先股堆叠和退出预期 |
| GI Alliance | 医疗服务方网络伙伴 | 2025 年增加 21 个活跃站点和 80+ 项进行中试验 | 商业上重要,因为它加深 GI 密度,而不是带来直接股权控制 | 确认排他性、经济条款和续约条款 |
| One GI | 医疗服务方网络伙伴 | 增加 34 家诊所和 13 个活跃研究站点 | 战略重要性在于社区站点供给,而非治理 | 确认表现义务和扩张经济性 |
| US Heart & Vascular | 医疗服务方网络伙伴 | 锚定进入心脏病学研究 | 对多专科增长和赞助方组合多元化很重要 | 确认收入分成模型和心脏病学管线承诺 |
同时纳入财务投资人和医疗服务方网络利益相关方,因为公司规模如今同时依赖资本和站点供给伙伴。
[CO015, CO016, CO019, CO021, CO022, CO028]可投性在网络规模、资本获取和监管验证上最强,但透明度和未解决的类别风险较弱。
[CO009, CO010, CO013, CO014, CO015, CO017]1.4 里程碑、监管证明点与关键风险
里程碑记录显示公司有两个清晰阶段。第一阶段从 2017 年成立延续到 2022 年 Series B,重点是面向胃肠病学的 AI,包括收购 CRSG/Precision,以及 2022 年 SKOUT 获得 FDA 许可。第二阶段在 2025-2026 年最明显,重点是借助 GI Alliance、One GI、US Heart & Vascular,以及从 NextStage 收购三个心脏病学研究中心,搭建多专科临床研究网络。 SKOUT 仍是最清楚的监管和产品证明点。官方材料和 FDA 文件支持这样的事实:该器械是一款 FDA 许可、用于结肠镜的实时计算机辅助检测工具;公司报告还称每次结肠镜腺瘤检出增加 27%,近端 5-9 mm 息肉检出增加 44%。这些事实很重要,因为它们说明 Iterative 能把 AI 工作转化为受监管的临床产品,而不只是流程软件。 主要反向背景不是公司特定执法事件,而是品类风险。独立文献在评估 CADe 系统时仍强调假阳性、自动化偏见、流程整合负担、泛化限制,以及长期结果数据不确定。再加上公开收入、债务和股权结构表披露缺失,公司的执行叙事强于其公开财务透明度。因此,后续章节应把 Iterative 视为运营上可信,但在经济性和采用耐久性上仍需要重尽调的公司。[CO005, CO024, CO025, CO026, CO027, CO028]
| 日期 | 事件 | 类型 | 金额 / 状态 | 参与方 | 含义 |
|---|---|---|---|---|---|
| 2017 | Iterative Scopes 作为 MIT 分拆公司创立,聚焦 GI AI | 创立 | Jonathan Ng | 确立公司“医生 + 计算”起源故事 | |
| 2021-08 | 完成 Series A 融资 | 融资 | $30M | Obvious Ventures、Eli Lilly、JJDC、Breyer、Seae 及其他投资方 | 为早期商业化和董事会搭建提供资金 |
| 2022-01 | 完成 Series B 融资 | 融资 | $150M | Insight Partners、Clearlake 和现有投资人 | 推动公司进入后期成长融资阶段 |
| 2022-04 | 收购 CRSG 和 Precision Research | 扩张 | Iterative Scopes、Chris Fourment | 扩展临床研究运营深度和站点知识 | |
| 2022-09 | SKOUT 获得 FDA 许可 | 监管 | FDA 已许可 | Iterative Scopes、FDA、Provation | 形成受监管产品证据和商业可信度 |
| 2025 | 独立 CADe 文献突出假阳性、自动化偏差、集成和可推广性限制 | 反向 | 品类风险仍在 | 独立临床审阅者 | 提醒投资人,器械采用仍有工作流和证据风险 |
| 2025-04 | 宣布 GI Alliance 合作 | 合作 | 21 个站点 / 80+ 项试验 | GI Alliance、Iterative Health | 加速社区 GI 网络密度 |
| 2025-09 | 宣布 One GI 合作 | 合作 | 34 家诊所 / 13 个活跃站点 | One GI、Iterative Health | 增加社区诊所研究能力 |
| 2025-10 | Bill Kayser 加入,出任总裁兼 CFO | 治理 | Bill Kayser、Jonathan Ng | 增加经验丰富的医生诊所财务领导力 | |
| 2026-03 | US Heart & Vascular 合作启动 | 合作 | >100 个站点,覆盖 3 个大陆 | USHV、Iterative Health | 标志着心脏病学扩张和多专科定位 |
| 2026-04 | 宣布 Series C | 融资 | $77M | Intrepid Growth Partners、GV、EDBI 和现有投资人 | 为超越 GI / 肝病学、进入肥胖和心脏病学的规模化提供资金 |
| 2026-05 | DDW 2026 数据发布显示,扩张后规模仍实现 >3x 基准入组表现 | 扩张 | 0.33 pts/site/month(按站点月计) | Iterative Health、GI Alliance 作者 | 公开支撑网络表现投资逻辑 |
| 2026-05 | 收购 NextStage 心脏病学站点 | 扩张 | 3 个德克萨斯站点 | Iterative Health 与 NextStage Clinical Research | 在关键市场增加自有心脏病学研究能力 |
该时间线意在记录截至运行日期的主要公司、监管、合作和品类风险公开时间线。
[CO001, CO007, CO009, CO019, CO021, CO024]公开里程碑显示,公司正从 GI AI 产品公司转向多专科临床研究基础设施。
[CO007, CO009, CO015, CO016, CO019, CO021]1.5 图表
02市场分析
2.1 市场边界与需求基础
Iterative Health 暴露在两个临床上重叠、但买方完全不同的市场里。一个是 AI 辅助结肠镜质量工具:SKOUT 这类软件进入内镜室,在质量门槛不断抬高的背景下尝试提升病变检出。另一个是 GI/社区临床试验研究中心网络服务:主要卖给需要更快启动、更好招募和触达社区患者人群的申办方与 CRO 的运营基础设施。若把两者并成一个 TAM,会掩盖一个事实:前者像医疗服务机构运营或技术预算,后者则是外包试验执行支出。 质量工具市场的纳入支出范围很窄。它覆盖病变检测软件、部署/支持,以及在筛查或监测性结肠镜中的质量流程适配;不包括更宽的 CRC 筛查宇宙,后者现在还包括血液检测和居家粪便检测;也不包括不改变检测表现的通用内镜软件。研究中心网络市场的纳入支出同样具体:研究中心启动、合同、培训、患者识别支持,以及申办方接入高表现社区 GI 研究中心。不包括完整 CRO 技术栈,也不包括每一项 GI 研究支出。 需求真实但不均匀。ACS 称,美国仍有 20M+ 符合条件的人没有按时完成 CRC 筛查;Medicare 政策也仍会把粪便或血液检测阳性者导向后续结肠镜。但新近扩龄的 45-49 岁人群远未饱和:2023 年只有 33.7% 按时完成筛查,采用集中在保险和教育水平更高的成年人。因而,Iterative 面向结肠镜的市场锚定在巨大的预防缺口上,但采用仍取决于可及性、外展触达,以及直接可视化筛查相对非侵入式替代方案是否仍具吸引力。[CM001, CM002, CM003, CM004, CM005, CM006]
| 细分 / 品类 | 纳入支出 | 排除支出 | 买方 / 支付方 | 现状替代 | 重要性 |
|---|---|---|---|---|---|
| CRC 筛查需求外层 | 最终可能需要结肠镜的适龄成年人,包括粪便或血液筛查阳性后的随访结肠镜 | 非内镜 CRC 筛查供应商、肿瘤治疗支出及无关 GI 诊疗 | 患者、支付方、医疗系统 | 不筛查,或只做非侵入式检测 | 任何结肠镜质量产品的外层需求 |
| AI 辅助结肠镜质量工具 | 实时病灶检测软件、部署、支持,以及筛查或监测结肠镜的检查室工作流适配 | 通用 EHR / 报告软件、非视觉筛查检测,或宽泛 GI IT 预算 | GI 诊所、ASC、医院或 VA 付费;内镜医生使用 | 手工质量改进、第二观察者,或不用 AI | SKOUT 类工具的直接产品市场 |
| 内镜质量衡量与报告 | ADR/SSLDR 跟踪、GIQuIC 或与 MIPS 对齐的质量工作流,以及绑定结肠镜表现的反馈系统 | 不改变结肠镜质量或报告行为的通用分析工具 | GI 质量负责人或管理员 | 手工病历审计和延迟反馈 | 重要邻近领域,因为质量标准上升会创造工具需求 |
| 面向赞助方的 GI / 社区试验站点网络服务 | 站点启动、签约、协调员支持、患者识别,以及赞助方进入社区 GI 站点的能力 | 完整 CRO 外包、实验室 / 影像服务和非站点网络试验功能 | 赞助方或 CRO 付费;PI 和站点人员使用 | 仅使用学术中心的站点策略,或碎片化直接签约站点 | Iterative 网络业务的直接服务市场 |
| 医疗服务方网络研究赋能 | 诊所合作、PI / 协调员支持,以及把试验嵌入社区 GI 诊疗的基础设施 | 不包含研究赋能的通用诊所管理服务 | 医疗服务集团在运营上合作;经济上常由赞助方预算支付 | 独立本地研究团队,或没有研究服务 | 对网络规模至关重要的供给侧层,并可能形成护城河 |
| 有意排除的邻近领域与替代品 | 只纳入与 Iterative 当前模型有关的检查室 / 站点网络切入点 | 宽泛非侵入式筛查、完整 CRO 支出和无关 GI 软件类别 | 因品类而异 | 血液 / 粪便检测、仅学术中心试验、自建站点运营,或不用 AI | 让市场规模测算扎根,避免把全品类支出当成虚高 TAM |
纳入支出按工作流和预算归属定义,而不是按每一美元是否触及 CRC 筛查或 GI 研究来定义。
[CM001, CM002, CM031, CM032, CM042, CM049]可防守的市场视角从宽泛的 CRC 筛查需求,收窄到更小的质量工具切口,以及另一个由申办方付费的站点网络服务切口。
各层使用混合运营单位,而不是可相加的美元口径,因为公开记录不足以把两类市场暴露统一到同一计量口径。
[CM001, CM003, CM008, CM013, CM026]2.2 受证据约束的测算口径
公开记录能支撑市场重要性,但无法给 Iterative 两块暴露市场中的任一块算出清晰的美元 TAM。对结肠镜 AI 来说,最有用的外层口径不是收入,而是质量压力如何作用于一个筛查仍不足的人群。2024 年 ACG/ASGE 更新把 ADR 衡量范围扩大到几乎所有 45 岁及以上成人的筛查、监测和诊断性结肠镜,并把整体基准抬到 35%,另设 6% 的 SSLDR 目标。这显著提高了承诺让检测更稳定的工具的适用性。Yale 的证据摘要和 Olympus 的 EAGLE 试验,都支持 CADe 能在不明显增加时间成本的情况下提升腺瘤检出的核心主张。 但买方仍必须围绕不确定性下判断。同一组文献认为,晚期肿瘤改善更小,长期间期癌和真实世界成本效益证据仍有限,CADe 商业化推进速度可以说快于上市后共识。因此,检查室业务更合适的测算口径是一个受约束的楔子:只有一部分结肠镜项目会因质量压力足够强,即便报销模糊也愿意购买。 对研究中心网络业务来说,最强的公开口径是申办方痛点,而不是市场研究 TAM。Iterative 的 IBD 数据显示,启动时间为 74 天,而基准为 122-171 天;每个研究中心每月 0.34 名患者,而基准生产率约 0.10;在更大 DDW 2026 规模下也有相近的 0.33-0.32 表现。这些都是有意义的服务市场指标,但它们不是可发布的 SAM/SOM,因为公开记录缺少定价、签约量、积压订单和研究中心层面的收入。[CM007, CM008, CM011, CM012, CM013, CM017]
| 发布方 / 视角 | 年份 | 地区 / 场景 | 数值 | 单位 | 方法 | 置信度 | 限制 |
|---|---|---|---|---|---|---|---|
| American Cancer Society 筛查缺口 | 2026 | 美国 | >20 | 百万未按期筛查成年人 | 基于指南更新和全国筛查缺口的人群需求视角 | 中 | 大需求外层,但不是 Iterative 特定收入 TAM |
| NHIS / JAMA 45-49 岁采用率 | 2023 | 美国 45-49 岁成年人 | 33.7 | % 已按期筛查 | 针对新扩龄队列的人群采用视角 | 中 | 年龄段特定,且不是所有适龄成年人的采用率 |
| NHIS / JAMA 模式组合 | 2023 | 美国 45-49 岁成年人 | 27.7 结肠镜;7.1 粪便 | % | 用于观察新适龄需求目前如何被捕捉的检查组合视角 | 中 | 没有把仅筛查结肠镜从所有诊断路径中拆出 |
| ACG / ASGE 质量门槛 | 2024-2026 | 结肠镜项目 | 总体 35;男性 40;女性 30;SSLDR 6 | % | 显示买方质量门槛的视角 | 高 | 质量目标不是以美元计的市场规模,但会强烈影响紧迫性 |
| Olympus EAGLE CADe 临床有效性 | 2025-2026 | 欧洲筛查和监测结肠镜检查 | ADR 提升 7.3;93 个大腺瘤;230 个 SSL | 百分点 / % 变化 | 评估检查室 ROI 的临床有效性视角 | 中 | 厂商赞助的产品证据,不是直接经济回报 |
| Iterative IBD 网络启动视角 | 2024-2025 | 2-3 期 UC/CD 试验 | 74 对 122-171;首次随机化 83 对 140 | 天 | 衡量赞助方痛点和节省时间的运营基准视角 | 中 | IBD 专属指标,不是直接价格或市场份额衡量 |
| Iterative 网络规模视角 | 2026 | 全球研究网络 | >100 个站点;>40 个合作伙伴 | 数量 | 来自公司和独立报道的当前产能视角 | 高 | 覆盖范围说明有规模,但不说明合同价值、利用率或收入分成 |
这些行是相互竞争的运营视角,不是可相加的 TAM 组成部分。它们有用,是因为公开记录不足以支撑一个可辩护的单一美元市场规模估计。
[CM003, CM004, CM005, CM007, CM008, CM013]新扩龄的 45-49 岁人群到 2023 年观察到采用率改善,但保留的所有指标仍指向少数人采用的市场。
低、中、高分别对应 2021、2019 和 2023 年观察到的流行率点,而不是预测情景。
[CM004, CM005, CM006]2.3 买方、用户与付款方地图
两块业务的买方地图分得很清楚。对 CADe 来说,日常用户是内镜医生,但经济讨论通常经过执业机构管理者、内镜医学主任、医院服务线负责人,或 VA 与卫生系统质量负责人。付款方通常是医疗服务机构本身。已审阅的支付来源没有显示 AI 辅助结肠镜有专门 Medicare 报销,因此预算负责人必须靠质量提升、竞争差异化、培训支持或流程效率来证明支出合理。Olympus 的云采购叙事和 Yale 的一键式实施描述都说明,供应商知道这个市场必须适配现有检查室经济性,而不能等新的支付代码。 对研究中心网络业务来说,经济买方更明显在申办方一侧。Iterative 自有材料和 Clinical Trials Arena 都描述,申办方和 CRO 可以集中接入表现更高、更具多样性的社区研究中心。GI Alliance 等医疗服务机构很重要,因为它们提供研究者、协调员和患者关系,但核心支出更接近外包试验执行,而不是内镜报销。社区诊所也得到共同收益:研究可以嵌入普通 GI 护理,让医生和患者获得接入机会,而不必把所有人都拉回学术中心流程。 这种分裂对估值很重要。结肠镜工具业务受医疗服务机构预算和房间级 ROI 约束。研究中心网络业务受申办方需求、研究中心表现,以及让社区诊所在多项研究中持续保持启动状态的能力约束。把它们称作一个市场,会错过它们经由不同采购动作、由不同付款方买单的事实。[CM016, CM033, CM040, CM041, CM042, CM043]
| 细分市场 | 买方 | 用户 | 付款方 | 工作流 | 预算负责人 | 采用触发因素 |
|---|---|---|---|---|---|---|
| 独立社区 GI 诊所 / ASC | 诊所管理员或内镜医疗主任 | 内镜医生和检查室人员 | 医疗服务机构 | 筛查或监测结肠镜检查时在检查室层面部署 | 诊所损益或 ASC 运营预算 | 需要在不拖慢周转的情况下提高检出质量 |
| 医院门诊内镜服务 | 服务线负责人、GI 科主任或质量委员会 | 内镜医生、护士、质量人员 | 医院 | 技术评审、IT 集成,然后进入检查室使用 | 医院资本开支或运营预算 | 需要拉开质量差异,并统一不同医生的表现 |
| VA 或学术医疗中心内镜单元 | 质量负责人或培训项目负责人 | 内镜医生、专科培训医生和护士 | 机构预算 | 基于现有软硬件栈增量加装 | 临床运营或质量预算 | 需要第二双眼睛和一致的教学工作流 |
| 购买 GI 站点准入的赞助方或 CRO | 临床运营或试验外包负责人 | 研究团队、站点经理、研究者 | 赞助方或 CRO | 站点筛选、签约、启动、招募和试验执行 | 临床开发预算 | 需要更快启动、入组,并触达社区患者 |
| 大型 GI 医疗服务网络伙伴 | 研究负责人和诊所高管 | 研究者、协调员、医生、患者 | 经济上由赞助方预算买单;运营上由医疗服务网络受益 | 把研究嵌入附属站点的日常 GI 诊疗 | 网络研究负责人和站点运营预算 | 需要把研究作为增长和留住医生的杠杆 |
| 社区 GI 试验站点团队 | 主要研究者和协调员负责人 | 协调员、PI、转诊临床医生 | 赞助方资助的研究预算 | 预筛、知情同意、启动、随机化并留住患者 | 在网络标准内本地管理的研究预算 | 需要降低行政负担,让研究流程更可预测 |
两条业务线的预算归属差异很大:CADe 主要看医疗服务方经济性,站点网络服务则主要由赞助方临床运营预算驱动。
[CM016, CM033, CM040, CM041, CM042, CM043]同一家 GI 机构可能出现在两条业务线中,但预算负责人、付款方和主要阻碍会因细分市场大幅变化。
角色标签根据公开工作流证据泛化;具体头衔会因机构和合同结构而异。
[CM033, CM037, CM040, CM041, CM042, CM043]2.4 增长驱动因素与采用约束
最大的共同顺风因素很直接:CRC 筛查现在从 45 岁开始,美国仍有庞大未筛查人群,结肠镜质量基准更紧,申办方也仍极度需要更快、更可靠的招募。纸面上,这对 SKOUT 和研究中心网络平台都是强配置。社区 GI 护理尤其重要,因为它集中了一批与慢性病患者的持续关系;这些患者同时影响筛查量和 GI 试验招募。 约束同样重要。CADe 一侧,文献仍提示假阳性、自动化偏见、泛化能力和有限长期结果;证据足以支撑兴趣,但不足以消除买方怀疑。同样关键的是,报销背景不友好。GI 学会称 2026 年 CMS 提案会削减 ASC 和医院内镜支付,Becker's 和 ACG 则描述了 Medicare 支付与诊所成本之间长期存在的缺口。MACRA 式成本衡量进一步让医疗服务机构对服务地点、麻醉、病理和后续利用更敏感。在这种语境下,AI 工具必须靠质量或流程胜出,而不是靠明显的报销套利。 研究中心网络业务面对的是不同但相连的约束。它受益于社区研究中心规模和申办方紧迫性,但公开证明仍集中在 IBD。进入肝病学、肥胖和心脏病学在战略上有吸引力,却证据更少。尽调结论因此是不对称的:市场真实,Iterative 也似乎在销售真实痛点;但在管理层披露定价、装机基础、申办方合同经济性和非 GI 复制数据之前,精确 TAM/SAM/SOM 数学应刻意保持收窄。[CM029, CM030, CM033, CM034, CM035, CM036]
| 驱动因素 / 约束 | 方向 | 时点 | 影响 | 尽调问题 |
|---|---|---|---|---|
| 45 岁起筛查扩容,加上 >20M 成年人未按时筛查 | 利好 | 当前 | 维持庞大筛查人群,既能支撑结肠镜量,也能带动下游质量工具需求 | 按付款方和站点类型量化新增筛查增长中仍流向结肠镜的比例 |
| ADR 和 SSLDR 预期提高 | 利好 | 当前 | 能提升或记录检出表现的工具和工作流,商业价值随之提高 | 要求客户数据,验证 SKOUT 是否足以改变上报质量指标并影响诊所行为 |
| CADe 有效,且几乎没有或没有明显时间惩罚 | 利好 | 当前 | 买方重视质量但承受不了吞吐损失,AI 因此更容易成立 | 收集检查室层面对退镜时间、假阳性和员工接受度的参考反馈 |
| 赞助方急于加快启动和入组 | 利好 | 当前 | 节省数月可压缩试验周期和成本,因此服务买方很清晰 | 询问赞助方续约率、积压订单和每个已启动站点收入 |
| 社区 GI 成为慢病诊疗中心 | 利好 | 当前 | 支撑一个判断:社区诊所是有战略价值的研究渠道,不是边缘场景 | 按适应症和地理位置绘制活跃社区站点密度 |
| AI 结肠镜没有专门报销 | 逆风 | 当前 | CADe 采购只能从医疗服务方预算中证明合理性,而不是走清晰的付款方资助 ROI 路径 | 按场景调取准确的 CPT 层面经济性和客户回本案例 |
| CMS 服务地点和医生支付压力 | 逆风 | 2025-2026 | ASC 和 HOPD 经济性被削弱,增量技术支出更难解释 | 按当前和拟议 Medicare 费率建模检查室经济性 |
| CADe 的工作流、报告和证据摩擦 | 逆风 | 当前 | 即使兴趣很高,假阳性、自动化偏差、测量负担和长期结果有限也会拖慢采用 | 审查上市后结果、GIQuIC 集成和质量报告自动化 |
| 定价和合同经济性公开透明度不足 | 约束 | 当前 | 公开信息无法支撑任一业务线的可信 SAM 或 SOM,也限制估值精度 | 要求提供装机基础、定价、赞助方合同价值和站点利用率队列 |
| 非侵入式筛查增长和可及性不均 | 约束 | 当前 | 血液和粪便选项扩大上游捕获,但可及性差异会限制服务不足群体中真正转化成的结肠镜需求 | 按付款方和人群测试非侵入式筛查阳性到结肠镜的漏斗转化 |
驱动因素解释这个品类为什么存在;逆风解释买方转化为什么可能落后于表面需求机会。
[CM029, CM030, CM033, CM034, CM035, CM036]两项业务都从识别痛点开始,走到预算归属、合规放行、部署,最后落到可衡量的质量或入组提升。
漏斗数值是顺序可视化权重,不是实测转化概率;有来源支撑的事实在细节文本和引用声明中。
[CM016, CM033, CM041, CM044, CM045, CM050]03竞争格局
3.1 内镜 AI 对手与 OEM 渠道权力
在检查室里,Iterative 竞争的对象与其说是另一家小型 GI 软件初创公司,不如说是大型医疗技术现任者加上一家专注型纯玩家。Medtronic 将 GI Genius 定位为 AI 辅助结肠镜模块;Olympus 则在更宽的 EVIS X1 生态内覆盖 CADDIE、OLYSENSE 和 ENDO-AID。Fujifilm 把 CAD EYE 放进自有处理器和内镜组合;Odin Vision 仍是最接近的纯玩家云竞争者,产品为 CADDIE。这些产品解决的临床任务与 SKOUT 相同:在结肠镜过程中提供实时息肉检测支持。 战略差异在于分销权力。Olympus 和 Fujifilm 可以把 AI 捆进已经存在于医院和 ASC 内部的品牌内镜设备路径、服务关系和升级周期。Medtronic 看起来更模块化,因为 GI Genius 作为独立智能内镜模块销售,但它仍带着规模化商业组织和买方熟悉的器械采购动作。因此,即便临床定位可信,Iterative 进入检查室时的 OEM 杠杆也低于拥有装机基础的现任者。 独立证据让这个品类保持吸引力,但并未尘埃落定。Yale 综述和 AGA 指南都支持 CADe 提高腺瘤检出,但也保留了对晚期病变、长期癌症结局、成本效益和普遍采用的不确定性。这在竞争上很重要,因为它留下了一个持久的现状替代方案:标准结肠镜加传统质量改进。还没有相信长期结果提升的买方,仍可以完全推迟 AI,或只在质量压力最急的地方试用。[CP001, CP003, CP004, CP005, CP006, CP008]
| 竞争对手 | 类别 | 公开规模 / 融资信号 | 目标细分市场 | 差异化 | 局限 |
|---|---|---|---|---|---|
| Iterative Health | AI + GI / 站点网络混合运营商 | 2026 年 $77M Series C 轮;公开报道 >100 个研究站点、>40 个合作伙伴 | 需要社区站点执行的 GI 医疗服务方、赞助方和 CRO | 已审阅对象中,唯一公开把 CADe 周边 AI 工具与嵌入式站点网络运营配在一起的公司 | 公开装机基础、价目表和赞助方赢单 / 输单数据仍未披露 |
| Medtronic GI Genius | 老牌医疗技术厂商的 CADe 模块 | 全球医疗技术龙头;将 GI Genius 作为 AI 辅助结肠镜模块销售 | 医院、ASC 和 GI 内镜项目 | 商业触达已有规模,定位为模块化 AI 加装件 | 已审阅公开材料未显示 GI 专属站点网络服务或透明定价 |
| Olympus OLYSENSE / CADDIE / ENDO-AID | 老牌 OEM 内镜生态 | 全球内镜龙头;2025 年推出更广的 OLYSENSE CAD/AI 产品组合 | 已安装 Olympus 的内镜套间 | EVIS X1 生态内整合硬件、云应用和工作流 | 锁定强度取决于 Olympus 硬件覆盖,可能限制多厂商灵活性 |
| Fujifilm CAD EYE | 老牌 OEM AI 内镜工具 | 大型影像 OEM;CAD EYE 绑定 EX-1 及兼容 Fujifilm 系统 | 基于 Fujifilm 的内镜项目 | 自有硬件集成,加上 AI 辅助检出 / 表征叙事 | 已审阅公开证据未显示广泛服务网络或赞助方执行能力 |
| Odin Vision CADDIE | 纯 AI 内镜专科厂商 | 规模较小的云 AI 专科厂商,主打临床医生案例和低培训门槛 | 寻求软件驱动 CADe 支持的站点 | 基于云的工作流,专科叙事更简单 | 公开可见的分销能力不如大型 OEM 明显 |
| IQVIA | 全球 CRO / 站点网络龙头 | 声称 II 期 IBD 成功率高 88%、III 期高 25%;100 个国家 >1,100 个站点 | 需要规模、数据和全球运营覆盖的赞助方 | 赞助方触达、站点覆盖和数据基础设施规模巨大 | 定位不围绕检查室 GI AI,也不强调嵌入社区诊所 |
| ICON + Accellacare | 全球 CRO + 站点网络龙头 | 五年内 114+ 项 GI 研究;Accellacare 拥有 8.1M+ 患者和 50+ 个站点 | 寻求可预测入组和启动支持的赞助方 | GI 研究历史广,加上自有站点网络基础设施 | GI 专注度不如 Alimentiv,AI 产品导向不如 Iterative |
| Alimentiv | 专科 GI CRO / 站点网络对手 | 深耕 GI 研究 30+ 年;60+ 个国家 5,000+ 个站点;支持 >70% 的 IBD 化合物 | 需要专科深度的 GI 和 IBD 赞助方 | 只聚焦 GI,具备集中化内镜传统和站点中心模式 | 已审阅证据未显示可与 SKOUT 相比的检查室 AI 产品 |
| 现状 / 碎片化直连站点运营 | 替代方案 / 内部自建 | 装机基础巨大但没有统一供应商;许多站点仍入组不足 | 避免新增供应商支出的医疗服务方和赞助方 | 前期无需增量软件或外包支出 | 启动更慢、入组可预测性更弱,也没有 CADe 提升 |
| 医疗服务方控制的研究附属业务 | 潜在对手 / 供给侧替代 | GI Alliance 和 One GI 披露了活跃研究站点和广泛诊所覆盖 | 能够内化或重定向研究流量的大型 GI 集团 | 在诊所层面掌握研究者关系和患者触达 | 今天可以与 Iterative 合作,之后仍可能变成议价对手 |
公开规模和差异化来自已审阅的官方及独立来源。定价单独处理,因为几乎所有竞争对手都靠联系销售型企业包装成交,而不是标价。
[CP001, CP004, CP006, CP009, CP014, CP016]序数轴在 x 轴衡量分销能力、在 y 轴衡量跨场景广度,显示 Iterative 被挤在硬件龙头和更大 CRO 网络之间。
点位是方向性的 1-10 序数评分,来自分销杠杆和两个相关场景广度的公开证据。它们不是实测市场份额或收入位置。
[CP001, CP004, CP006, CP014, CP024, CP028]3.2 GI 与 IBD 试验网络对手
在面向申办方的一侧,Iterative 遇到的竞争者比检查室 AI 里大得多。IQVIA 和 ICON 都在全球研究中心网络和患者触达基础设施之上,销售 GI 或 IBD 专项开发能力。Alimentiv 从相反方向切入:它比全球 CRO 窄,但把自己定位为深度专精 GI 的运营商,拥有内镜、组织学和 IBD 流程厚度。当买方更看重申办方覆盖、地域触达或成熟外包关系,而不是 Iterative 的 GI 优先运营叙事时,这三类对手都能与 Iterative 对卖。 规模差距是公开的。IQVIA 宣称在 100 个国家拥有 1,100+ 个高表现研究中心,并声称在外包 II 期和 III 期 IBD 工作中占有超额份额。ICON 宣传过去五年有 114+ 项 GI 研究,加上 Accellacare 网络覆盖 8.1M 患者、50+ 个研究中心和更快研究中心启动指标。Alimentiv 宣称在 60+ 个国家拥有 5,000+ 个研究中心关系,并称其支持了 70%+ 在研 IBD 化合物。Iterative 自有材料显示出有意义的动能,但公开的申办方触达规模仍小于这些现任者网络。 医疗服务机构控制的研究中心供给是另一股竞争压力。GI Alliance 和 One GI 今天是合作伙伴,在狭义上不是对手;但它们披露的网络说明大型 GI 集团为何具有战略意义:它们控制研究者、协调员、患者流和本地试验接入。如果这些关系不能长期排他,同一个让 Iterative 差异化的资产,也可能遭到争夺、让医疗服务集团内部化,或流向另一个 CRO 或运营商。[CP024, CP025, CP026, CP027, CP028, CP029]
| 购买标准 | Iterative | Medtronic | Olympus | Fujifilm | IQVIA | ICON | Alimentiv |
|---|---|---|---|---|---|---|---|
| 实时结肠镜 CADe | 是 | 是 | 是 | 是 | 否 | 否 | 否 |
| OEM 硬件捆绑 / 已安装生态 | 低 | 中 | 高 | 高 | None | None | None |
| 云端或可持续更新的 AI 栈 | 中 | Unknown | 高 | 中 | None | None | None |
| 专属 GI / IBD 试验执行深度 | 高 | None | None | None | 高 | 中 | 高 |
| 规模化站点网络或患者触达基础设施 | 中 | None | None | None | 高 | 高 | 高 |
| 多专科赞助方覆盖 | 中 | None | 低 | 低 | 高 | 高 | 低 |
| 嵌入式医疗服务网络关系 | 高 | 低 | 中 | 低 | 中 | 中 | 中 |
| 存在强渠道锁定风险、对 Iterative 不利的公开证据 | 中 | 中 | 高 | 高 | 低 | 中 | 低 |
是 / 否 / 低 / 中 / 高取值是基于证据的公开能力序位判断,不是测得的市场份额。未知表示已审阅公开来源集未支撑该单元格。
[CP001, CP004, CP006, CP008, CP014, CP015]最高杠杆的能力被拆开:OEM 掌握手术室层面的硬件套装,CRO 掌握申办方和站点规模;在被审视玩家中,只有 Iterative 有两侧公开证据。
高 / 中 / 低 / 无 / 未知评级概括护城河关键能力广度,而不是每个产品功能。未知表示本章审阅的公开证据未能支持该单元格。
[CP024, CP025, CP028, CP031, CP034, CP036]3.3 打包、定价不透明与买方替代方案
公开商业打包比公开定价清楚得多。Iterative、Medtronic、Olympus、Fujifilm、IQVIA、ICON 和 Alimentiv 展示的都是联系销售驱动的企业销售,而不是透明标价。落到实践里,买方评估这些方案靠的是设备捆绑、服务工作说明书、研究中心网络接入或定制企业合同,而不是公开的单次检查或单研究中心费率表。不透明本身就具有战略意义,因为它有利于拥有更重一线团队、也更有谈判打包空间的规模化卖方。 替代集合也比常规单品类对比表更宽。GI 医疗服务机构可以从 Medtronic、Olympus 或 Fujifilm 购买 CADe 模块,而不买 Iterative 的研究中心网络服务。申办方可以从 IQVIA、ICON 或 Alimentiv 购买 GI 或 IBD 试验执行,而不买 Iterative 的检查室 AI。买方也仍可以留在现状:内镜室不使用 AI,或由内部团队和通用 CRO 流程支持碎片化的直接对接研究中心运营。 这些替代方案让两块业务的多归属程度不均。试验服务更容易多归属,因为申办方本来就按方案、地域或职能拆分工作。内镜 AI 一旦某个研究中心在硬件生态、流程和服务伙伴上标准化,多归属就更难。这种不对称解释了为什么硬件捆绑和医疗服务网络接入,比任何一张公开疗效图表都更关系护城河耐久性。[CP023, CP030, CP041, CP042, CP043, CP044]
| 公司 / 替代方案 | 公开合同模式 | 包含能力 | 公开价格可见度 | 含义 |
|---|---|---|---|---|
| Iterative Health | 演示驱动的企业销售 | AI 工具,加上集中化试验运营和站点支持 | 已审阅来源未见公开标价 | 买方可能围绕打包结果和网络准入谈判,而不是围绕标价 |
| Medtronic GI Genius | 模块 / 设备式企业销售 | 用于息肉检出的 AI 辅助结肠镜模块 | 已审阅来源未见公开标价 | 作为临床加装件竞争,可嵌入现有采购流程 |
| Olympus OLYSENSE / CADDIE / ENDO-AID | 通过 Olympus 设备和中枢打包的生态方案 | 云端 CAD/AI 应用,加上集成硬件工作流 | 已审阅来源未见公开标价 | 打包会把 AI 经济性藏在更大的主机、镜体和服务决策里 |
| Fujifilm CAD EYE | 硬件绑定的软件包 | 通过 EX-1 和兼容 Fujifilm 系统提供 CAD EYE | 已审阅来源未见公开标价 | 包装更有利于已经标准化使用 Fujifilm 处理器和镜体的站点 |
| IQVIA | CRO / 站点网络服务协议 | GI 开发专长、站点合作、人员配置、招募和技术 | 已审阅来源未见公开合同费率 | 规模和范围可能使定价按方案和地区高度定制 |
| ICON / Accellacare | CRO / 网络 SOW | GI 专长,加上站点 / 患者解决方案和全球站点网络准入 | 已审阅来源未见公开合同费率 | 可把招募和站点表现打包进更大的临床开发合同 |
| Alimentiv | 专科 GI CRO 合同 | 聚焦 GI 的运营、内镜和站点网络服务 | 已审阅来源未披露公开合同费率 | 当申办方需要 GI 深度能力时,专科化可能支撑溢价 |
| 现状 / 内部自建 | 内部人力与碎片化供应商栈 | 直接管理站点、手工提升质量、通用 CRO 支持 | 只能从内部预算看见,而非供应商价目表 | 前期看似更便宜,但可能更慢、噪声更大,也更难对标 |
已审阅来源对产品包装叙事较充分,但几乎没有公开价格。缺少公开费率表本身就是竞争图景的一部分,因为比较重心会转向谈判后的 ROI、捆绑采购和客户背书销售。
[CP023, CP030, CP041, CP042, CP043, CP044]公开 KPI 信号解释了 Iterative 为什么可信,也说明大型在位者仍可能从两条战线挤压其护城河。
每个 KPI 都是有来源支撑的公开信号,不是完整的市场份额模型。指标组合混合了公司、竞争对手和独立证据,因为竞争问题要相对比较,而不是单看 Iterative。
[CP003, CP006, CP012, CP028, CP031, CP034]3.4 护城河耐久性与反向竞争风险
Iterative 最好的公开竞争故事,不是它拥有单个最好的 CADe 产品,也不是它拥有单个最大的试验网络;而是它似乎把两种通常分开的能力拼在一起:能帮助 GI 流程的 AI 工具,以及能帮助申办方更快招募和启动研究的嵌入式社区研究中心运营。这个混合定位真实存在,也足够重要,而且在本章审阅的公开集合中并不常见。 但这条护城河的耐久性仍有条件。AI 一侧,Olympus 和 Fujifilm 可以把更多智能装进已安装的内镜生态,从而加深渠道锁定;Medtronic 可以用模块化分销,在买方想要更轻插件的任何地方竞争。试验一侧,IQVIA 和 ICON 可以在申办方覆盖和地域触达上用规模压过 Iterative;Alimentiv 则能挑战「只有 Iterative 提供 GI 专项深度」这一说法。GI Alliance 和 One GI 等医疗服务网络还提醒投资人,研究中心供给是谈判来的,不是拥有的。 反向证据基础让这种风险更尖锐,而不是消除它。AGA 仍没有对 CADe 辅助结肠镜给出支持或反对推荐,试点文献仍显示额外的非肿瘤性息肉检出,公开材料仍未披露 SKOUT 装机份额、相对具名对手的申办方胜率,或任一业务的实际定价。投资结论因此更有层次:整合模式不同,但公开证明尚不足以强到可以无视现任者反击或商品化。[CP019, CP020, CP022, CP024, CP041, CP043]
| 护城河主张 | 主要威胁 | 严重性 | 缓释措施 / 尽调要求 |
|---|---|---|---|
| AI 与站点网络一体化栈 | 买方仍可把 CADe 和试验运营从不同供应商拼起来 | 高 | 要求披露交叉销售率、附加率,以及 SKOUT 是否显著提升站点网络销售(或反向提升)的证据 |
| 社区 GI 站点供给与医疗机构伙伴关系 | 大型医疗机构网络可将附属业务留作独家,或转给其他运营商 | 高 | 获取 GI Alliance 和 One GI 的合同条款、排他期、续约权和伙伴集中度指标 |
| GI 优先的运营可信度 | Alimentiv 等专科对手可主张更深的 GI 和 IBD 专长 | 中 | 将试验胜率、内镜能力和研究者留存与 Alimentiv 及其他 GI 专科机构对标 |
| AI 疗效叙事 | AGA 仍未给出推荐,长期结局证据仍有限 | 中 | 要求提供 SKOUT 部署后的真实世界结局数据、成本效果分析和高级别病变表现 |
| 不依赖内镜 OEM 的分销 | Olympus 和 Fujifilm 可将 AI 捆进硬件生态与服务周期 | 高 | 核实互操作路线图、转售或集成伙伴关系,以及锁定在 Olympus / Fujifilm 账户中的敞口 |
| 申办方关系增长 | IQVIA 和 ICON 在全球申办方覆盖和多专科项目上可用规模压过 Iterative | 高 | 要求披露头部申办方集中度、重复预订队列数据,以及相对具名 CRO 的胜负分析 |
| 借助 AI 预筛与人员配置形成运营差异化 | 公开数据在 IBD 最强,GI / 肝病之外明显更薄 | 中 | 在承保多专科护城河耐久性前,要求提供肥胖、肝病和心脏病学队列证据 |
严重性反映每个威胁对 Iterative 当前公开护城河的可能冲击,而不是公司失败的预测。每一行都列出最可能把叙事差异化转化为可承保证据的尽调事项。
[CP019, CP020, CP024, CP041, CP043, CP045]3.5 图表
04财务情况
4.1 收入模式、变现触点与定价不透明
Iterative Health 的公开材料支持一个多触点变现模式,但无法审计其公开收入结构。公司现在销售的是面向申办方和研究中心的研究基础设施主张:集中接入不断扩大的社区研究中心网络,提供运营支持、试验管理和 AI 驱动的流程工具。GI Alliance、One GI 和 US Heart & Vascular 都把 Iterative 描述成不止是软件供应商。每个案例里,公司贡献的是试验管理能力、运营骨干、研究中心赋能和申办方接入,而不仅是仪表盘或算法。这使得申办方/CRO 服务和医疗服务网络运营合作,成为最清楚可观察的收入流。 SKOUT 增加了第二个商业化触点。它证明 Iterative 仍有一个可销售给 GI 诊所和内镜场景的真实受监管产品,但公开记录没有显示标价、报销表或装机基础经济性。服务侧也同样不透明:没有费率表,没有披露合同条款,也没有说明收入在申办方预算、医疗服务合作和器械销售之间如何拆分。因此,最强结论是方向性的,而不是数字性的。Iterative 几乎肯定横跨申办方服务、医疗服务网络经济性、AI 驱动流程和 SKOUT 变现,但公司没有披露区分软件式经常性收入与项目制或服务驱动收入所需的定价架构。[CI001, CI002, CI003, CI004, CI005, CI006]
| 收入来源 | 机制 | 单位 / 定价代理 | 当前公开信号 | 收入质量 | 尽调要求 |
|---|---|---|---|---|---|
| 申办方 / CRO 试验执行服务 | 集中接入高表现站点,并提供启动、激活、招募和试验管理支持 | 可能按研究预算 / 工作说明书计费;无公开费率表 | >40 家制药 / 生物技术 / 器械 / CRO 伙伴;>100 个站点;声称激活速度快 2 倍、入组快 3 倍 | 中 — 若表现属实则粘性较强,但仍受研究预算和方案流影响 | 披露申办方 / CRO 合同贡献的收入、复购率,以及任何待交付订单或已签约未确认收入 |
| 医疗机构网络运营伙伴关系 | Iterative 为 GI Alliance、One GI、US Heart & Vascular 等集团管理或支持研究附属业务 | 可能来自共享服务、管理费或附属业务经济;条款未公开 | 21 个 GI Alliance 站点、13 个 One GI 活跃站点,并为 USHV 站点提供端到端支持 | 中 — 若具有排他性且嵌入运营,可能较耐久,但取决于伙伴留存和经济分成 | 按伙伴类型披露合同期限、排他性、收入分成和现金回本情况 |
| AI 赋能的招募与工作流服务 | 嵌入临床试验运营的 AI 招募、预筛和工作流工具 | 独立软件价格未披露;可能捆绑进研究服务 | 2022 年 CRSG/Precision 交易将 AI Recruitment 与药企伙伴、医疗机构和站点绑定 | 中高 — 可能具备软件式粘性,但公开证据未把该收入流与服务拆开 | 披露附加率、独立定价,以及包含 AI 工具的合同占比 |
| SKOUT 设备 / GI 产品收入 | 获 FDA 许可、面向结肠镜流程销售的实时息肉检测 AI | 设备 / 服务协议 / 订阅经济未公开 | 产品营销活跃且有疗效主张;无公开报销表或标价 | 中 — 产品可通过服务 / 支持形成复现收入,但医疗机构付费意愿可能受报销影响 | 披露装机量、平均售价、服务收入,以及设备经济中依赖医院或 ASC 预算的比例 |
Iterative 披露了商业触点,但未披露收入结构。各行根据公开证据梳理变现机制;若没有明确数字,定价和收入质量均标为推断。
[CI001, CI003, CI004, CI005, CI006, CI007]| 产品 / 交易对手 | 公开定价披露 | 标价 vs. 实际成交价 | 可观察合同线索 | 主要折扣 / 未知项 | 来源支持的含义 |
|---|---|---|---|---|---|
| 面向 GI 诊所、医院和 ASC 的 SKOUT | 标价未知;实际成交价可能按账户、预算周期和支持范围谈判 | 产品页营销疗效和工作流收益,但未披露价格或报销代码 | 硬件 / 软件拆分、服务费、耗材和合同模式未知 | 若看不到独立报销,变现很可能依赖医疗机构资本开支或捆绑运营预算,而非透明的按次付费 | |
| 申办方 / CRO 网络接入与试验执行 | 可能按方案、地域和表现计入工作说明书或里程碑预算 | 官方材料向申办方和 CRO 销售速度、激活可靠性和入组提升 | 研究最低规模、取消保护、利润率结构或续约机制未知 | 这更像企业服务收入,但公开来源无法说明其复现性或待交付订单属性 | |
| 医疗机构网络研究伙伴关系 | 可能是定制化伙伴经济,包含运营支持和附属服务整合 | GI Alliance、One GI 和 USHV 都强调运营骨干和共同扩大研究,而非简单转售软件 | 排他费、收入分成、人员负担和集成成本未知 | 关系质量可能较强,但实际经济无法用公开数据建模 | |
| AI 赋能的临床服务 / 算法 | 看不到稳定的 Medicare 标价;支付路径仍处在过渡、演进或假设阶段 | 法律和政策来源讨论拟议 APC、捆绑、基于价值定价和周期性重定价,而非已落定报销 | 未来报销归设备商、医疗机构、医院还是三方共享,仍未知 | AI 变现具备战略吸引力,但政策脆弱,尤其是叠加在已付费流程上的工作流工具 |
空值表示已审阅公开来源没有披露标价。变现观察区分了明确公开事实、推断的合同结构和政策语境。
[CI012, CI013, CI014, CI015, CI020, CI021]梳理医疗提供方关系、站点运营、AI 工具和 SKOUT 如何把活动转成可变现场景,同时价格和利润率大多仍未披露。
[CI003, CI005, CI006, CI007, CI008, CI009]4.2 成本结构、报销摩擦与单位经济性代理指标
这个模式运营上看起来有力量,但财务上比纯软件更重。Iterative 自身语言强调集中化运营、专家人员、业务开发、研究中心支持、监管职能和主动研究中心管理。NextStage 收购新闻稿尤其有信息量,因为它描述了一个横跨多个治疗领域、覆盖研究中心人员和集中职能的 250 人组织。这意味着,即便不考虑产品研发或未来收购,公司也会承担显著人工成本、入驻成本和企业管理开销。 公开报销背景又增加了一层摩擦。Medicare 确实覆盖筛查性结肠镜,但已审阅来源没有显示类似 SKOUT 的结肠镜 AI 有稳定的独立支付路径。与此同时,AGA 和 Becker's 描述了恶化的 GI 报销压力,尤其是 ASC 和医院内镜;Boston Scientific 指南也强调,不同服务地点和谈判后的商业保险合同会让经济性差异极大。这些事实很重要,因为它们会限制医疗服务客户愿意为附加技术或辅助服务支付的价格。 单位经济性披露仍不完整。Iterative 最好的公开代理指标是研究中心数量、合作伙伴数量、启动速度和招募表现。这些指标说明产品-服务组合有运营价值。但没有 CAC、LTV、回本周期、申办方复购率或单研究中心收入,投资人无法把这些运营成果转化为可靠承销模型。[CI015, CI016, CI017, CI018, CI019, CI020]
| 指标 | 数值 / 状态 | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| 研究站点覆盖 | 全球 >100 个研究站点 | 高 | 网络密度越高,申办方销售和患者触达越好,规模因此重要 | 拆分全资管理 vs. 伙伴运营站点、活跃 vs. 非活跃站点,以及每个站点的平均研究负载 |
| 申办方 / 伙伴密度 | >40 家制药 / 生物技术 / 器械 / CRO 伙伴 | 高 | 交易对手基础广,是商业需求和潜在复购的正向代理指标 | 披露申办方集中度、复购率和前 10 大收入敞口 |
| 入组效率 | IBD 中 0.33 名患者 / 站点 / 月 vs. 0.10 基准 | 高 | 入组速度是公司公开证明运营价值和潜在定价权的最清晰证据 | 说明该指标在 IBD 之外是否成立,以及如何转化为每项试验的实际申办方收入 |
| 激活速度 | 激活速度快 2 倍,启动时间最多缩短 3 个月 | 高 | 若可重复,更快激活可支撑溢价或更高申办方胜率 | 披露按治疗领域划分的激活时间中位数,并说明经济口径是否包含转付站点成本 |
| 员工数 / 站点代理指标 | 公司总部开销前约 1.7-2.5 名员工 / 站点(166-250 名员工 / 100+ 个站点) | 低 | 这一粗略比率表明,模型依赖集中化杠杆和伙伴劳动力,而不是完全由公司自有站点配满人员 | 按站点队列提供站点层人员配置、总部开销分摊和全负荷人员成本 |
| 获客成本(CAC) | 低 | 没有 CAC,无法承保销售效率和 GTM 资本强度 | 分别披露申办方 CAC 和医疗机构伙伴 CAC | |
| 客户终身价值(LTV) | 低 | 没有 LTV 或留存数据,投资人无法判断粘性关系能否抵消高实施成本 | 按申办方和医疗机构伙伴队列提供毛利润 LTV | |
| 回本周期 | 低 | 服务占比较重的业务运营上可能好看,但若回本周期长,仍会烧钱 | 披露新伙伴上线和公司收购站点的回本周期 | |
| 单站点 / 单伙伴收入 | 低 | 这是从运营规模通向财务产出的最简单桥梁,但仍缺失 | 提供每个活跃站点、每个申办方和每个医疗机构网络的平均年化收入 | |
| 毛利率 | 低 | 需要毛利率,才能把高价值软件 / 工作流贡献与人力密集型服务交付拆开 | 按收入流披露直接成本和毛利率:申办方服务、伙伴运营和 SKOUT |
各行混合了已披露运营代理指标和故意留空的财务指标。空值表示该指标是承保必需项,但在已审阅公开来源中未披露。
[CI004, CI015, CI016, CI017, CI018, CI042]展示从站点足迹到申办方价值的最佳公开路径,同时明确标出 CAC / LTV / 回本周期仍缺失。
激活速度和入组等运营指标是真实的公开代理变量,但 CAC、LTV、单站点收入和毛利率仍未披露,因此无法转换成客户层面的单位经济性。
[CI004, CI007, CI026, CI042, CI043, CI045]展示融资和合同回款可能如何先被站点运营、收购和产品投入消耗,然后才能判断公开现金跑道。
[CI024, CI025, CI026, CI041, CI046, CI050]4.3 资本充足性、融资依赖与公开可比公司背景
Iterative 并不明显缺钱,但公开信息仍不透明。唯一硬的一手资本事实是 $77M Series C,管理层明确把这笔钱与进入心脏病学、肥胖、地域扩张和深化医疗服务网络合作绑定。二手来源对投后估值约 $1.3B-$1.4B、累计融资约 $270M-$273M 的判断趋同,但这些数字仍是交叉推断,而非公司披露。官方来源仍未公布当前现金、烧钱速度、债务或现金跑道。 这个缺口更重要,因为公司是在运营层面扩张,而不只是数字化扩张。研究中心收购、多专科搭建和服务较重的交付模式,通常会先消耗现金,然后投资人才能观察到有吸引力的利润率结构。公开 CRO 可比公司展示了模式做大后的好状态。IQVIA、ICON 和 Medpace 都报告数十亿美元收入、明确流动性,以及利润率/积压订单数据;它们披露的经营利润率或 EBITDA 利润率集中在低双位数到二十出头。但这些公开可比公司也说明,研究服务会计很复杂,存在营运资本波动;ICON 还出现过收入确认重述和重大缺陷。 所以资本问题与其说是「Iterative 还能不能融资?」,不如说是「Iterative 是否已经搭出一台自我供血机器?」公开信息下,答案仍未知。[CI025, CI026, CI027, CI028, CI029, CI030]
| 项目 | 数值 / 状态 | 置信度 | 重要性 | 尽调要求 |
|---|---|---|---|---|
| 最新一级市场融资事件 | $77M Series C 轮 | 高 | 这是唯一明确支撑未来扩张计划的一手资本事实 | 确认扣除费用后的到账现金,以及是否包含任何老股流动性 |
| 迄今估计累计融资 | $270M-$273M 二手来源区间 | 中 | 这勾勒出历史融资能力,但并非公司披露,不应视为已审计 | 核对成立以来的准确股权融资和任何债务融资 |
| 最新估值估计 | $1.3B-$1.4B 二手来源区间 | 中 | 估值影响稀释预期和融资灵活性,但当前数字来自追踪器 | 提供 Series C 实际投后估值、优先股堆叠和任何结构性条款 |
| 计划资金用途 | 扩张心脏病学、肥胖、地域和医疗机构伙伴关系 | 高 | 这表明资本仍投向增长和网络建设,而非已经公开收割利润 | 披露站点扩张、M&A、产品研发和总部开销之间的预算拆分 |
| 公开在手现金 | 低 | 当前流动性是判断现金跑道和下一轮时间点的核心输入 | 提供非受限现金、受限现金和任何最低现金契约义务 | |
| 公开烧钱速度 / 现金跑道 | 低 | 没有烧钱速度,投资人无法判断 Series C 是过桥资本还是多年增长资本 | 提供月度或季度经营现金消耗,以及基准 / 下行情景现金跑道 | |
| 公开债务 / 租赁义务 | 低 | 即便股权融资看起来充足,债务、租赁或 earnout 仍可能实质改变风险 | 提供债务到期表、租赁承诺、收购 earnout,以及任何申办方融资安排 | |
| 资本强度信号 | 250 人团队、100+ 个站点、活跃收购、多专科扩张 | 中 | 运营规模意味着固定成本和营运资本需求可观,即使软件杠杆随时间改善 | 拆分人工、站点支持、技术和收购整合成本 |
| 下一轮触发因素(推断) | 若多专科建设、站点收购和产品商业化在利润率被证明前跑得快于内部现金生成,可能触发 | 低 | 这是公开投资人无法从当前披露中回答的实际问题 | 提供董事会基准融资计划,以及避免再融资所需的运营阈值 |
| 成熟 CRO 参照 | 上市可比公司披露约 13%-22% 经营 / EBITDA 利润率,以及待交付订单和现金数据 | 高 | 这为规模化临床研究经济可能长什么样、投资人会期待多少披露提供现实校验 | 说明 Iterative 的直接成本结构与大型 CRO 有何不同,以及管理层认为长期利润率可落在哪里 |
一手事实和二手来源区间分开呈现。空值表示该指标对承保至关重要,但已审阅公开记录中没有。
[CI025, CI026, CI029, CI030, CI031, CI032]可由公开资料支撑的少数财务输入区间:融资、估值、员工数、成熟 CRO 利润率参照,以及报销波动。
Iterative 未公开披露收入、毛利率、烧钱速度或现金跑道。因此,本图只使用经来源支撑的区间,或可从已审阅记录中辩护的可比公司背景。
[CI016, CI027, CI028, CI029, CI030, CI031]4.4 财务尽调阻塞点与承销结论
公开记录足以勾勒业务模型,但不足以承销它。投资人可以合理得出结论:Iterative 有真实商业需求、真实申办方和医疗服务关系,以及至少一个 SKOUT 这样的受监管产品触点。投资人也可以判断,公司有足够融资通道继续扩张。无法得出的结论是:模式中有多少是经常性的,直接利润率长什么样,研究中心经济性多久回本,或近期扩张后还剩多少现金。 最重要的阻塞点很直接:没有收入或 ARR 披露;没有按收入流拆分;没有毛利率或直接成本拆分;没有现金、烧钱速度、现金跑道或债务表;没有公开价格卡;也没有申办方留存、积压订单或客户集中度披露。员工数和估值代理指标互相冲突,也强化了二手来源只能近似的事实。 因此,财务结论是谨慎但不负面。Iterative 看起来商业上可信、战略融资也到位,但行为仍像一家希望投资人相信运营动能、而不是披露经济性的私人公司。对尽调来说,这意味着第 4 章应把上行空间视为可能存在,但在管理层分享实际财务报表和合同指标之前,资本充足性和单位经济性信心都仍未证实。[CI014, CI024, CI026, CI028, CI032, CI045]
| 缺失指标 | 对承保的影响 | 当前公开代理指标 | 缺失时置信度 | 具体尽调路径 |
|---|---|---|---|---|
| 收入 / ARR 及按收入流拆分 | 无法区分软件式复现经济与项目或服务收入 | 仅有运营规模指标:站点、伙伴、激活和入组 | 低 | 要求提供按收入流拆分的月度或季度收入桥,并说明 Series B 以来的历史结构演变。 |
| 毛利率和直接成本结构 | 无法评估模型最终能否接近软件利润率,还是仍偏人力密集 | 仅有成熟 CRO 可比利润率 | 低 | 要求按收入流拆分毛利率、直接人工和站点支持成本 |
| 现金、烧钱速度、现金跑道和债务 | 无法判断资本充足性或下一轮时间点 | Series C 金额及二手融资区间 | 低 | 要求提供当前现金、季度经营现金流、债务 / 租赁到期表和基准情景现金跑道 |
| 合同定价和收入确认条款 | 无法判断收入是类似待交付订单、基于里程碑、或有,还是大量转付 | 医疗机构伙伴关系和申办方服务仅作定性描述 | 低 | 要求提供样本合同经济、付款节奏、取消条款,以及待交付订单 / 递延收入披露 |
| 申办方留存、集中度和待交付订单 | 无法衡量需求质量或未来收入可见性 | 仅 >40 个伙伴 | 低 | 要求提供过去 12 个月申办方留存、前 10 大收入集中度、待交付订单和 book-to-bill |
| 站点层盈利能力和上线回本 | 无法判断快速扩张是在创造价值,还是只是在吸收资本 | 激活速度和入组表现指标 | 低 | 要求提供伙伴站点上线回本、收购站点 EBITDA 和每个活跃站点收入 |
| 经核对的员工数和经营计划假设 | 员工数代理指标冲突,使成本结构建模不稳定 | 官方 250+ vs. Tracxn 166 | 低 | 要求提供当前 FTE 数量,按职能和地域拆分,并给出计划招聘 / 生产率假设 |
这些是仍会阻碍投资级财务承保的最低公开数据缺口。若公开证据只能提供方向性而非决策级答案,表中列为代理指标。
[CI014, CI026, CI027, CI028, CI032, CI045]4.5 图表
05产品与技术
5.1 产品栈与客户流程
Iterative Health 并不把自己呈现为单一产品医疗技术供应商。横跨官网、关于页面和当前招聘描述,公司都在描述一个混合医疗技术与服务平台:它把嵌入社区的研究基础设施、AI 驱动的疾病评估工具,以及至少一个受监管的现场护理器械结合起来。按客户流程看,这套栈服务两个相连但不同的用户群。申办方和 CRO 用 Iterative 更快启动研究中心、降低运营负担,并改善 GI 周边试验的招募。医疗服务机构和内镜医生在筛查与监测性结肠镜中使用 SKOUT,把它作为检查室内的辅助器械。因此,公司的商业承诺不只是算法表现,而是把患者识别、研究中心运营、内镜评估和申办方报告连接起来的更宽运营模式。 公开可见模块强化了这一判断。SKOUT 是最具体的产品触点,因为它有 FDA 许可适应证、GUDID 记录和已发布的流程描述。但更宽平台也能从公司反复强调的内容中看见:嵌入社区护理的全球研究中心网络,带运营支持的启动和预筛流程,用于溃疡性结肠炎的 AI 驱动内镜与组织学分析,以及面向试验终点的 AI-人工混合中央阅片范式。这种广度具有战略意义。它说明 Iterative 试图拥有更多 GI 研究流程,而不是只吃到单个检测算法能捕获的价值;相对主要销售一个 CADe 模块的品类同行,这是重要差异点。[CE001, CE002, CE003, CE004, CE005, CE018]
| 模块 / 资产 | 主要用户 | 状态 / 成熟度 | 差异化 | 尽调缺口 |
|---|---|---|---|---|
| SKOUT 实时 CADe | GI 内镜医生 | 获 FDA 许可并已商业分销 | 流程内直接叠加,带工作流控制和已发表疗效主张 | 装机量、定价模式和互操作覆盖未公开 |
| 站点网络 | 申办方、CRO、医疗机构集团 | 横跨 GI 相邻专科的规模化运营网络 | 嵌入社区、以站点为中心的试验执行,而非独立软件 | 确切站点数、留存和经济情况未完全披露 |
| 试验激活 / 预筛运营 | 研究站点协调员和申办方 | 已在公开网络分析中证明运营成效 | 瞄准文档、培训、预算和预筛瓶颈 | 未披露独立产品包装或定价 |
| IBD 内镜 AI 评分 | 试验申办方和中央阅片方 | 可从会议和期刊摘要中公开看到 | 把 Iterative 从息肉检测推进到终点量化 | 没有全文论文时,方法细节仅部分公开 |
| AI 赋能的组织学关联 | 药物开发者和 GI 试验团队 | 处于研究阶段,但具备临床相关性 | 把 UC 的宏观内镜表现与微观炎症连接起来 | 商业包装、验证范围和客户采用未披露 |
| AI + 人工混合中央阅片 | 中央阅片方、申办方、监管方 | 已在 BMJ 公开摘要中验证 | 结合两个独立模型,并由医生定向裁决 | 运营定价、模型治理细节和生产使用仍未披露 |
| 申办方表现 / 对标分析 | 申办方和网络运营方 | 可从 DDW 和网络相关发表中看到 | 在规模化层面把站点表现与激活、随机化结果挂钩 | 仪表盘、API 和导出工作流未公开描述 |
各行反映公开可见的模块和运营资产,而非完整内部产品目录。
[CE004, CE006, CE018, CE025, CE028, CE040]| 用户任务 | 当前工作流 | Iterative 方案 | 可衡量收益 | 局限 |
|---|---|---|---|---|
| 在筛查或监测结肠镜检查中检测息肉 | 人工实时查看标准内镜视频 | SKOUT 在实时检查中叠加疑似息肉标记,并支持旁路 / SKOUT 画面切换 | 官方材料称,检查时间不会增加,病灶检出更强 | 公开资料只列明受支持的处理器环境 |
| 启动新的 IBD 试验中心 | 手工推进合同、培训、预算和文档流程 | Iterative 的研究中心网络运营把启动工作流集中起来并提供支持 | 已发表的网络分析显示,从中心筛选到启动的中位时间为 74 天 | 该运营层背后的具体软件工具没有单独说明 |
| 更快筛选并随机化试验患者 | 研究中心手工预筛,还要争夺有限的合格患者 | 网络模式叠加运营支持,缩短首例筛选和首例随机化时间 | 引用分析显示,首例患者筛选用时 45 天,首例患者随机化用时 83 天 | 已发表材料证明其跑赢基准,但客户层面的可重复性未披露 |
| 为溃疡性结肠炎内镜终点评分 | 人工中央阅片采用 2+1 范式,材料差异较大 | Iterative 在 2M+1H 混合框架中贡献一个模型 | BMJ 摘要显示,每段视频所需人工阅片减少 81%,一致性仍不劣 | 仍依赖独立模型和医生裁定 |
| 连接内镜与组织学,用于疾病评估 | 肉眼炎症和显微炎症通常分开审阅 | AI 驱动的内镜和组织学分析能暴露细微炎症模式 | 常规相关性较弱时,可帮助解读轻中度 UC | 商业部署细节和客户背书未公开 |
| 为申办方做网络表现基准 | 申办方团队需要在碎片化 CRO / 研究中心数据中比较中心表现 | Iterative 发布其研究中心网络的启动和随机化基准 | DDW 2026 更新称,在规模大约翻倍后,>3x 随机化优势仍在 | 没有公开客户仪表盘截图,也没有 API / 导出细节 |
收益和限制结合了官方出版物,以及可获得的独立佐证;流程细节留空通常反映实现方式未披露,而不是能力缺失。
[CE008, CE012, CE018, CE019, CE020, CE021]公开工作流把申办方 / 站点激活步骤,与术中 AI 辅助和试验评估输出连接起来。
[CE003, CE008, CE018, CE024, CE028, CE029]5.2 架构、集成与运营模式
SKOUT 的技术架构比其余技术栈有更具体的文档。FDA 许可和 GUDID 记录描述了一个实时软件层:它通过 SDI 接收内镜视频,在结肠镜过程中分析视频流,并在检测到潜在结直肠息肉时叠加蓝色矩形框。同一批来源还显示,产品设计直接适配检查流程,而不是迫使用户进入单独审阅环境:用户可以在旁路内镜视频流和 SKOUT 视频流之间切换;当工具进入视野或照明不足时,系统会暂停检测。兼容性目前受特定处理器生态限制,尤其是 Olympus EVIS EXERA II/III 和 Fujifilm Eluxeo VP-7000 输出,这让集成质量和已安装设备组合成为直接采用依赖。 更宽的运营架构更偏服务。Iterative 的研究中心网络论文把启动、预算、预筛、培训和监管文件描述为它试图集中化或支持的运营瓶颈。IBD 终点工具则叠在这套流程之上:AI 驱动的内镜和组织学分析增加了结构化疾病评估输出;BMJ 描述的 2M+1H 框架则把两个独立 ML 模型和一名胃肠病学专家裁决者放进中央阅片,而不是直接替代人工审阅。同一套架构也让分销和文档合作变得重要。Iterative 的 FDA 许可公告把 SKOUT 商业化与 Provation 绑定,后者的 GI 流程足迹可以降低进入市场摩擦。简单说,Iterative 的技术护城河不是单个模型,而是器械集成、运营赋能、人工监督,以及横跨研究中心和内镜室的流程邻接组合。[CE006, CE007, CE008, CE009, CE010, CE018]
| 层级 / 流程 / 组件 | 作用 | 依赖 | 风险 |
|---|---|---|---|
| 视频处理器接口 | 将 SDI 内镜视频流接入 SKOUT | 兼容 Olympus EVIS EXERA II/III 与 Fujifilm Eluxeo VP-7000 的 HD 输出 | 不支持的处理器或工作流变体会卡住部署 |
| SKOUT 检测 / 叠加引擎 | 处理实时视频,并用蓝色矩形框标出疑似息肉 | 实时推理要可靠,视觉叠加要稳定 | 误报或叠加体验差会侵蚀临床医生信任 |
| 模式选择和设备状态控制 | 用户可在旁路与 SKOUT 画面之间切换,并查看激活 / 错误状态 | 清晰的 GUI 行为和操作员培训 | 设备只是辅助,不是自主决策,人因风险很关键 |
| 研究中心网络运营层 | 负责启动、监管文档、培训、预算和预筛支持 | 中央运营团队加参与研究中心 | 研究中心人员配置或申办方协同变弱,表现就会下滑 |
| AI 疾病评估模型 | 为 IBD 项目量化内镜和组织学炎症 | 经筛选的多模态数据集和持续模型维护 | 泛化能力和商业化路径仍只披露了一部分 |
| 混合中央阅片裁定 | 终点评分先经过两个 ML 模型,意见不一致时交由人工裁定 | 独立模型、裁定医生可用性和治理流程 | 分歧率或监督负担上升,运营节省就会消失 |
| 分销 / 文档合作伙伴层 | Provation 帮助商业化触达既有 GI 工作流环境 | 合作伙伴持续同频,渠道执行要跟上 | 依赖渠道会压缩直接商业控制力,也会降低利润率可见度 |
本表强调可见的工作流架构,而不是公司未公开披露的隐藏模型服务基础设施。
[CE007, CE008, CE009, CE010, CE024, CE028]Iterative 的公开架构,把受监管的内镜室设备叠在站点运营、疾病评估模型和面向申办方的工作流层之上。
[CE004, CE007, CE018, CE025, CE040, CE041]Iterative 的产品栈依赖硬件兼容性、站点采用、人类专家、申办方数据流和渠道伙伴。
[CE041, CE042, CE043, CE044, CE048]5.3 信任、证据与竞争背景
Iterative 的信任故事起点不是纯 AI 新颖性,而是监管和流程护栏。SKOUT 现在拥有作为 II 类胃肠道病变软件检测系统的 510(k) 许可,并在 AccessGUDID 中列为已商业分销。器械标签明确写明它辅助而非替代受训胃肠病学专家,用户界面也包含状态信号;当工具进入视野时,系统进入暂停检测状态。这些控制很重要,因为它们把 SKOUT 定位为受监管检查中的增强层,而不是自主诊断决策者。公开指引方向上支持:Yale 对 AGA 指南草案的总结称,CADe 在成人结肠镜中获得有条件推荐,且不会实质增加检查时间,这大体支撑了 SKOUT 所在品类的采用逻辑。 尽管如此,证据基础仍有细节。Iterative 官方材料和合作伙伴简报强调了强疗效信号,例如检查时间不增加、每 4.5 名患者多切除一个腺瘤,以及在无蒂或中等大小病变上表现更好;更早的 SKOUT 研究和更宽的 CADe 文献也支持更高 ADR/APC。但文献同样警告假阳性、自动化偏见、操作者技能退化、泛化能力和长期结果。Medtronic、Olympus 和 Fujifilm 的竞争材料显示,实时 AI 检测、无蒂病变强调和流程保持主张已经不再是 Iterative 独有。因此,对 Iterative 技术栈的信任,不只取决于品类层面的 AI 疗效,还取决于其分销、集成、研究中心运营和 human-in-the-loop 试验流程组合,能否扛住其他 CADe 供应商和运营替代方案的竞争。[CE012, CE013, CE014, CE015, CE016, CE032]
| 控制 / 质量信号 | 状态 | 范围 | 公开缺口 |
|---|---|---|---|
| FDA 510(k) 准入 | 已于 May 9, 2025 获批 | SKOUT 用于成人结直肠癌筛查或监测性结肠镜检查 | 未披露装机量、报销情况或商业采用 |
| AccessGUDID 商业分销 | 有效 | 列明 SKOUT210 已进入商业分销,并附工作流描述 | 没有公开使用量或账户名单 |
| 人在回路标注 | 准入文件和伙伴简报均明确提到 | SKOUT 辅助受训胃肠科医生,并在使用工具时暂停 | 未公开误报率或操作员培训完成数据 |
| 品类指南支持 | 有条件支持 | Yale 摘要的 AGA 指南草案支持成人结肠镜使用 CADe,且检查时间几乎不增加或没有实质增加 | 指南是品类层面的支持,不是 SKOUT 特定采用证据 |
| 已发表的检测效果 | 正向,但来源混合 | 官方材料和 SKOUT 研究支持 ADR / APC 提升及无蒂病灶价值 | 完整独立复现和长期结局证据仍有限 |
| 已知 CADe 风险 | 持续存在 | 文献提到误报、自动化偏差、操作员技能退化、集成和泛化问题 | 没有公开的 SKOUT 特定现场表现仪表盘来回应这些担忧 |
| 竞争压力 | 高 | Medtronic、Olympus 和 Fujifilm 都在销售带工作流主张的 AI 结肠镜辅助产品 | 公开数据还不能说明,除覆盖面和分销外,SKOUT 为什么能在整个品类胜出 |
状态只反映已审阅的公开来源。缺口描述的是尽调中仍存在重大未披露的信息,不一定是产品缺陷。
[CE006, CE011, CE032, CE033, CE034, CE036]公开证据显示,SKOUT 和站点网络是最成熟的界面;更广泛的 IBD 工具前景不错,但商业透明度较低。
成熟度为定性判断,反映公开证据深度,而不是私有产品遥测或客户数量。
[CE011, CE023, CE040, CE045, CE050]5.4 路线图、成熟度与关键依赖
公开路线图主要通过里程碑证据显现,而不是通过前瞻产品计划显现。器械一侧,Iterative 从早期 SKOUT 疗效数据,推进到其价值分析材料引用的更大随机研究,再到 2025 年 FDA 许可和商业分销状态。试验技术一侧,2025-2026 年的发布节奏很重要:研究中心网络论文和 DDW 2026 更新强调了规模化运营表现;溃疡性结肠炎内镜加组织学工作显示 Iterative 正从程序辅助扩展到终点分析;BMJ 中央阅片论文则展示了一个更具监管意识、用于高风险试验评分的人工加模型框架。这些都是有意义的成熟度信号,因为它们指向的路线图,是沿 GI 研究流程扩张,而不仅是给 SKOUT 添加功能。 主要依赖同样清楚。SKOUT 采用取决于兼容处理器足迹和 Provation 帮助下的商业触达。研究中心网络论点取决于社区研究中心持续采用、申办方愿意把研究导向以研究中心为核心的模式,以及有足够训练有素的运营人员来管理启动、预筛和激活。中央阅片和疾病评估工具依赖经过整理的视频与组织学数据、可维护的模型表现,以及专家人工裁决者的接入。最后,公开记录仍留下若干业务关键空白:没有披露装机基础,没有公开定价,没有详尽互操作矩阵,也没有针对器械或 IBD 工具链的详细前瞻发布计划。因此,虽然 Iterative 的产品成熟度强于单功能初创公司,但在投资人以充分信心承销技术规模化之前,仍有重要披露缺口需要补齐。[CE011, CE017, CE022, CE023, CE041, CE042]
| 日期 / 阶段 | 功能 / 里程碑 | 状态 | 含义 | 来源 |
|---|---|---|---|---|
| 2021 研究发表 | 早期 SKOUT 前瞻性研究 | 已发表 | 在更广泛商业化前,建立初步检测效果信号 | SE013 |
| 当前简报引用的随机研究证据 | 1,359 例患者 SKOUT 数据集 | 已发表 / 已用于市场材料 | 相比只讲试点设备,支撑更强的成熟度叙事 | SE022 |
| 2025-05-09 | SKOUT 获得 FDA 510(k) 准入 | 已完成 | 让 SKOUT 变成受监管的商业工作流产品 | SE012 |
| 2025 商业状态 | AccessGUDID 商业分销登记 | 有效 | 表明设备已经越过准入,进入商业分销 | SE023 |
| 2025-2026 发表周期 | 混合 AI-人工中央阅片,以及 UC 内镜加组织学成果 | 推进中 | 显示公司从操作辅助扩展到终点和疾病评估工具 | SE009 / SE010 |
| 2026 DDW 更新 | 研究中心网络在翻倍规模下的扩展表现 | 推进中 | 说明运营模型可能已经越过初始基准队列,具备扩展性 | SE007 |
| June 2026 招聘信号 | startup.jobs 公司页显示 52 个开放岗位 | 持续扩建 | 意味着公司仍在为网络和产品运营投入组织资源 | SE024 |
路线图条目反映外部可见的里程碑,而不是详尽的未来功能计划。来源列使用本地来源 ID 以便追踪。
[CE011, CE017, CE022, CE023, CE045, CE047]5.5 图表
06客户情况
6.1 按买方、用户与付款方划分的客户分层
Iterative Health 销售进入的是一个相互连接但并不相同的客户栈。医疗服务一侧,公司争取医生集团、管理服务组织、主要研究者和研究中心运营商,这些客户希望把临床研究搭建或扩展为有利润的辅助服务。申办方一侧,Iterative 直接面向申办方、CRO 和生物技术公司销售,解决更快启动、招募预测和表现可管理的研究中心网络需求。围绕 SKOUT 还有第三类更不透明的买方:正在评估 AI 辅助结肠镜设备的 GI 内镜医生和医疗服务机构。公开证明在前两层最强,在第三层弱得多。 这种分层很重要,因为 Iterative 卖的不只是软件席位。它卖的是运营杠杆、患者识别,以及能够稳定执行的研究中心接入。医疗服务机构侧的主张强调集中化业务运营、人员配置、监管帮助、培训和申办方接入;申办方侧的主张强调高表现研究中心和加速试验执行。公司现在把网络呈现为多专科,而不是仅限 GI,覆盖 GI、肝病学、肥胖和心脏病学,并拥有 100+ 个研究中心和 40+ 个申办方/CRO/器械合作伙伴。但具名客户证据仍集中在 GI Alliance、One GI、US Heart & Vascular、Gastro Health 等大型美国医疗服务网络,以及与 Gastro One 的研究中心级案例;申办方侧具名证明大多限于 Takeda 和汇总数量。[CU001, CU002, CU003, CU006, CU007, CU008]
| 分层 | 买方 / 用户 / 付款方 | 主要用例 | 公开规模信号 | 收入 / 战略价值 | 尽调缺口 |
|---|---|---|---|---|---|
| 大型 GI 医疗服务网络 / MSO | 医疗服务高管、PI、研究中心负责人;服务方承担运营伙伴角色,申办方为研究付费 | 把研究打造成可规模化的附加服务,覆盖启动、人员、预算和申办方接入 | GI Alliance 21 个活跃研究中心、80+ 项试验;One GI 34 家诊所 / 13 个活跃研究中心 | 战略价值高:一个关系可打开多名研究者、多个中心和患者池 | 收入分成、排他性和合同期限未披露 |
| 社区 GI 与肝病研究中心 | 本地研究中心运营方和 CRC 团队;Iterative 充当运营骨干 | 降低监管和运营负担,同时改善启动、筛选和留存 | 已发表 IBD 队列含 27 个中心;Docwire 摘要称美国 21 个、欧洲 6 个 | 支撑申办方表现主张的核心交付层 | 具名中心列表和单中心留存数据未公开 |
| 申办方、CRO 和生物科技公司 | 临床运营、外包和研究设计团队 | 入组预测、中心选择、更快启动和更可预测的试验执行 | C 轮材料提到 40+ 家制药 / 生物科技 / 器械 / CRO 伙伴 | 试验执行收入和申办方扩张的主要付款群体 | 除 Takeda 外的具名申办方及各申办方确切研究数量未公开 |
| 具名申办方合作伙伴 | 医学事务和药物开发团队 | AI 驱动的患者选择、疾病评估和临床试验优化 | 公开资料中只明确找到 Takeda 这一具名申办方关系 | 信号强,因为它证明申办方可信度;但不足以衡量申办方集中度 | 商业范围、重复研究量和定价未披露 |
| 相邻专科网络(心脏病 / 肥胖扩张) | 大型多中心专科集团和医生推动者 | 将 Iterative 的网络模式从 GI 延伸到新的治疗领域 | US Heart & Vascular 伙伴关系在 2026 带来心脏病入口和 100+ 中心表述 | 同时撬动医疗服务方和申办方的先落地再扩张路径 | 具名肥胖客户未公开,心脏病铺开仍处早期 |
| SKOUT 设备用户 | 内镜医生、GI 诊所、门诊手术中心、采购团队 | 结肠镜检查中的实时 AI 辅助息肉检测 | 公开页面有产品演示入口和匿名证言,但没有具名账户 | 可能把收入从网络服务之外拓宽 | 装机量、具名部署和客户满意度数据缺失 |
分层结合了公司官方定位、具名伙伴公告和客户侧确认。公开材料未按分层披露收入结构。
[CU001, CU002, CU003, CU006, CU007, CU008]6.2 采用轨迹与部署规模
公开采用曲线不是靠按队列披露的客户数,而是通过网络里程碑显出来。2025 年 4 月,GI Alliance 称,纳入其 21 个活跃研究站点和 80 多项活跃 Phase II-IV 试验后,Iterative 网络扩展到美国和欧洲 70 多个地点。到 2025 年 9 月 One GI 合作时,Iterative 已成为一个 34 家诊所、13 个活跃研究站点网络的嵌入式运营伙伴。到 2026 年 4 月,US Heart & Vascular 合作把公司重新定位为进入心脏病学的多专科网络;同月 Series C 公告中,Iterative 引述的规模已是北美、欧洲、印度和澳大利亚 100 多个研究站点,以及 40 多家制药、生物技术、医疗器械和 CRO 伙伴。 最有力的落地证据不是账户数,而是试验表现指标。Iterative 的医疗服务方页面称,启动速度快 2x、患者入组高 3.4x,AI 预筛查将随机化提高 40%。JCC 摘要和后续 DDW 2026 发布给出更具体的运营数字:从站点选择到启动 74 天、到首位患者筛查 45 天、到首位随机化 83 天,每站点每月大约随机化 0.33-0.34 名患者;4 项完成试验在满额入组时仍交付约每站点每月 0.32 名患者。这些是高质量采用代理指标,因为它们显示,即使分母和研究数量上升,网络仍在运转;但 Iterative 仍未按赞助方、站点或产品线公布标准客户队列数据。[CU003, CU004, CU005, CU010, CU014, CU015]
| 期间 / 指标 | 数值 / 里程碑 | 日期 | 来源 | 置信度 | 含义 | 缺失分母 |
|---|---|---|---|---|---|---|
| Gastro Health 切入点 | AIR 部署进一个覆盖七州、拥有 380 名医生和 150 个地点的 GI 平台,并明确表示会继续扩展 | Pre-2025(页面未披露日期) | gastrohealth.com 新闻稿 | 中 | 说明 Iterative 可先靠技术进入大型医疗服务平台,再推进更广泛的运营铺开 | 未提供启动数、试验数或续约分母 |
| GI Alliance 网络扩张 | 21 个活跃 GIA 研究中心、80+ 项活跃 II-IV 期试验;Iterative 网络扩展到美国 / 欧洲 70+ 个地点 | 2025-04-23 | GI Alliance 与 BioSpace | 高 | 首次大规模公开证明:该网络能在单一服务方伙伴下聚合有意义的研究中心体量 | 未披露申办方组合或收入贡献 |
| Gastro One 中心执行 | 启动速度快 2x;>50% 入组来自 Iterative;~5 个月完成两名患者随机化 | 2025-06-23 | Iterative 案例研究 | 中 | 证明的是中心层面的运营杠杆,不只是新增一个标志客户 | 只描述了一个中心和一项试验 |
| One GI 正式合作 | 六个州 34 家诊所;13 个活跃研究中心成为网络中的排他性主力 | 2025-09-30 | Iterative 与 OneGI | 高 | 又一个规模化服务方网络加深了客户证明,也显示既有关系内可能继续多中心扩张 | 未披露收入分成,也未披露经 One GI 承接的申办方项目数量 |
| JCC 研究中心网络数据 | 74 天启动、45 天首例筛选、83 天首例随机化、0.34 名患者 / 中心 / 月 | 2026-02-19 | Iterative 对同行评审文章的摘要 | 高 | 提供跨 2/3 期 IBD 研究的采用和利用率基准指标 | 底层摘要页无法访问;样本量细节仍依赖间接信息 |
| USHV 心脏病扩张 | 研究中心加入网络;该网络现已覆盖三大洲 100+ 个中心 | 2026-04-02 | Iterative 与 USHV | 高 | 显示客户基础从 GI 扩展到心脏病和多专科医疗服务集团 | 新增 USHV 中心的确切数量未公开 |
| C 轮规模快照 | 全球 100+ 个中心、40+ 家制药 / 生物科技 / 器械 / CRO 伙伴 | 2026-04-30 | Iterative 与 Pulse 2.0 | 高 | 这是服务方和申办方合并基础最明确的公开规模点 | 未拆分活跃、已签约和试点阶段伙伴 |
| DDW 2026 扩容后表现 | 0.33 次随机化 / 中心 / 月,对比 0.10 基准;4 项试验以 0.32 中位速率完成入组 | 2026-05-04 | Iterative、BusinessWire、Morningstar、Yahoo | 高 | 显示在网络分母较上一公开读数大约翻倍时,使用质量仍维持高位 | 仍没有申办方层面的队列或续约分母 |
Iterative 不公布年度新增客户、申办方队列或设备装机量。因此,采用情况只能从里程碑披露和表现指标中推断。
[CU003, CU009, CU010, CU014, CU016, CU017]6.3 具名客户证据与证据质量
具名客户证据可信,但不均衡。医疗服务方一侧,GI Alliance、One GI、US Heart & Vascular、Gastro Health 和 Gastro One 都出现在公开材料中,并附有高管引述、站点数、试验数或量化工作流结果。GI Alliance 和 One GI 给出最干净的平台级证据,因为公司和客户侧组织都描述了这段关系,也都量化了挂靠的站点规模。Gastro One 给出最丰富的单站点运营案例,包括更快启动、随机化,以及通过 Iterative 获得的入组占比。US Heart & Vascular 将逻辑从 GI 拓宽到心脏病学;Gastro Health 则证明,Iterative 能先靠 AI 招募工具落地,再推进更广的网络铺开。 赞助方侧证据明显更薄。Takeda 是公开材料中发现的唯一具名赞助方关系,即便如此,证据也围绕 AI 赋能的试验设计和疾病评估展开,而不是明确合同范围或重复研究量。公司披露的 40 多家赞助方 / CRO / 器械伙伴数量有助于说明覆盖面,但不能替代具名部署、每个赞助方的研究数量或赞助方级留存数据。SKOUT 作为客户业务的公开证据也弱。产品页面有匿名推荐和功效主张,但没有具名医疗系统或门诊手术中心部署参考。因此,本章的证据质量最好概括为:医疗服务方网络客户标识 强,试验表现结果好,赞助方具名有限,具名产品安装弱。[CU006, CU007, CU008, CU009, CU010, CU011]
| 客户 / 伙伴 | 分层 | 部署 / 用例 | 生产 / 试点 | 结果 / 证据 | 限制 |
|---|---|---|---|---|---|
| GI Alliance | 医疗服务网络 / GI MSO | 战略伙伴关系,将 21 个活跃研究中心和 80+ 项 GI / 肝病试验接入 Iterative 网络 | 生产规模网络关系 | 客户侧和第三方新闻确认中心数量,并有 GI Alliance CMO 一年后的 DDW 跟进评论 | 未公开经济条款、排他性或续约安排 |
| One GI | 医疗服务网络 / GI MSO | 战略伙伴关系:34 家诊所网络中的 13 个活跃研究中心成为 Iterative 网络的排他性主力 | 生产规模网络关系 | Iterative 和 OneGI 都称这是多年关系,并引用中心负责人 Ziad Younes 对深度运营支持的评价 | 未公开申办方数量或财务贡献 |
| Gastro One | 单中心 / 医疗服务方案例研究 | 为一项罕见病 1b-A 期 GI 试验提供运营支持 | 特定试验执行案例研究 | 启动速度快 2x,>50% 入组来自 Iterative,~5 个月完成两例随机化,并把一个数百万美元机会提前 4 个月 | 单中心证据,不一定代表全网络经济性 |
| US Heart & Vascular | 医疗服务网络 / 心脏病 | 社区型心血管研究合作与网络接入 | 早期生产扩张关系 | 客户侧和公司侧公告确认业务越过 GI,并使用 100+ 中心网络表述 | 心脏病中心和研究的确切数量未披露 |
| Gastro Health | 医疗服务平台 / AI 招募 | 部署 AIR 技术,改善试验招募和患者识别 | 以技术切入,并有扩张意图 | 客户侧新闻稿点名一个拥有 380 名医生、150 个地点、覆盖七州的平台,并表示有意在全网络扩展 | 没有与铺开绑定的公开启动或入组指标 |
| Takeda | 申办方 / 生物制药 | 为 IBD 项目提供 AI 驱动的内镜评估、患者选择和试验设计支持 | 具名申办方合作 | 公开资料中唯一明确具名的申办方关系;博客称相关工作已经用于指导药物开发和上市后策略 | 商业范围、研究数量和续约节奏未披露 |
| SKOUT 用户(未具名) | 设备买方 / 内镜诊所 | 结肠镜检查中的实时 AI 息肉检测 | 从公开客户证据看,生产状态不清楚 | 产品页面包含匿名临床医生好评和效果主张 | 没有公开的具名医疗系统、ASC 或 GI 诊所部署案例 |
各行覆盖公开证据最清晰的具名关系。本表有意区分医疗服务网络证据、申办方证据和设备产品证据。
[CU006, CU007, CU008, CU009, CU010, CU019]6.4 留存、重复使用和耐久性代理指标
Iterative 不披露 NRR、GRR、客户标识流失、续约率、合同期限或公开评测平台评分,所以留存判断只能依赖间接信号。最好的重复使用代理指标是纵向的。Ziad Younes 说 Iterative 已经作为可信伙伴服务其站点多年,这比一次性实施引述更有力。GI Alliance 首席医疗官将 DDW 2026 结果描述为战略合作一年后已经成为可能的证据。DDW 发布还称,2026 年展示是第二次大规模公开报告聚合网络表现,并强调 4 项研究以稳定随机化率完成满额入组。这些都是有用信号,因为它们说明 Iterative 不只是宣布客户标识,而是在入驻后继续执行。 但投资人不应把运营一致性误读为客户留存证明。临床试验具有阶段性,赞助方常常同时使用多个站点网络,医疗服务集团也可能同时保留多条研究或技术关系。公开证据强调启动、预筛查、病历审阅减负和试验内患者留存;它没有说明多少赞助方带着新方案回来、多少医疗服务集团在第一年后扩大合同范围,或多少收入会自动续约。被拦截的 OUP 摘要页面也意味着,直接会议记录不如周边公司摘要和联合新闻稿覆盖容易取得。简言之,耐久性方向上偏正面,但硬留存数据仍未披露。[CU023, CU024, CU025, CU027, CU028, CU029]
| 指标 / 代理指标 | 数值 / 观察 | 分层 | 置信度 | 尽调问题 |
|---|---|---|---|---|
| NRR / GRR | 未公开披露 | 全部客户 | 高 | 索取 2024-2026 年按申办方和医疗机构队列拆分的毛 / 净收入留存率 |
| Logo 流失 | 未公开披露 | 所有客户 | 高 | 索取按申办方、医疗机构网络和单个站点拆分的 logo 流失率 |
| 合同期限 / 续约节奏 | 未公开披露 | 医疗机构与申办方合同 | 高 | 索取按头部客户拆分的平均期限、自动续约机制和续约管线 |
| 评价平台 / NPS 信号 | 未找到网络业务的公开 G2、Gartner Peer Insights 或 NPS 数据 | 所有客户 | 高 | 索取可背书的满意度调查,或按队列拆分的申办方 / 站点 NPS |
| 医疗机构多年合作背书 | Ziad Younes 称,Iterative 多年来一直是其站点可信赖的伙伴 | 医疗机构站点 | 高 | 厘清开始日期、合同范围演变,以及该关系对收入的影响 |
| 后续合作证据 | GI Alliance CMO 称,DDW 2026 数据体现了战略合作推进一年后能做到什么 | 大型医疗机构网络 | 高 | 索取 GI Alliance 关系内赢得的研究、续约率和新增站点数 |
| 重复公开绩效披露 | DDW 2026 被定位为聚合网络绩效的第二次重要公开披露,规模约翻倍 | 申办方与网络站点 | 高 | 索取申办方再次授标率和每个申办方的重复方案数量 |
| 已完成试验的持久性代理指标 | 四项试验完成全量入组,同时中位随机化仍维持在约 0.32 名患者 / 站点 / 月 | IBD 申办方项目 | 高 | 索取完成全量入组的研究是否转化为后续授标或新治疗领域工作 |
Iterative 不公布传统留存指标,因此多数行只能用重复公开证据或多年伙伴关系表述来代理。
[CU023, CU024, CU027, CU028, CU029, CU030]6.5 扩张路径与集中度风险
Iterative 的扩张逻辑清晰:在现有医疗服务网络内增加站点,把同一网络模型从 GI 推向肥胖和心脏病学,随着试验表现数据累积深化赞助方关系,并可能把 AIR、疾病评估工作流等 AI 工具与站点运营一起交叉销售。2026 年与 US Heart & Vascular 一起进入心脏病学,以及 Series C 信息中围绕肥胖和更广多专科增长的表述,都支持这一判断。如果公司在进入相邻治疗领域时仍能守住启动和入组优势,现有赞助方就可能有理由把更多试验工作集中到该网络内。 风险在于,公开证据集中在少数大型伙伴生态中,而且仍高度依赖伙伴撰写的公告。GI Alliance、One GI、US Heart & Vascular、Gastro Health 和 Gastro One 主导具名医疗服务方记录;赞助方侧具名稀少;SKOUT 安装情况不透明。这留下了几个未回答的问题:头部账户收入集中度、赞助方续约频率,以及医疗服务集团是独家合作,还是只是几条并行研究渠道之一。采购摩擦也是实际风险,因为 Iterative 的价值主张依赖人员配置、监管工作、预算和病历审阅支持;如果站点或赞助方把这些能力内化,可能会更强硬压价,或把工作拆给多个供应商。器械侧竞争会放大这个问题:Medtronic 的 GI Genius 公开营销强检测主张,所以 SKOUT 采用很可能取决于工作流契合和网络关系,而不是一个无人竞争的产品品类。[CU018, CU034, CU035, CU036, CU037, CU038]
| 扩张驱动 / 风险 | 公开证据显示什么 | 可能影响 | 尽调路径 |
|---|---|---|---|
| GI 内大型网络扩张 | GI Alliance、One GI、Gastro Health 和 Gastro One 显示,Iterative 能通过医疗机构组织落地并扩张 | 若一份合同能打开多个站点是利好;若少数网络主导具名证据和收入,则风险上升 | 索取前十大医疗机构网络收入占比和已签活跃合同下的站点数 |
| 申办方组合扩张 | 40+ 家申办方 / CRO / 器械合作伙伴和 Takeda 关系,暗示其广度不止一个申办方 | 若申办方带来重复方案,利好交叉销售;没有申办方队列数据则仍不透明 | 索取按活跃研究拆分的申办方数量、重复研究率和头部申办方收入贡献 |
| 多专科扩张 | USHV 加上 Series C 信息,把模型延伸到心脏病学和肥胖领域 | 可能扩大 TAM,并降低 GI 集中度 | 索取首批心脏病学申办方、已接入心脏病学站点数和早期启动指标 |
| 医疗机构集中度 | 大多数公开具名证据集中在 GI Alliance、One GI、USHV、Gastro Health 和 Gastro One | 若少数网络贡献大部分收入或患者触达,集中度风险高 | 索取按医疗机构伙伴拆分的收入和研究量敞口 |
| 申办方多栖合作 | 申办方通常同时使用多个站点网络和 CRO;公开证据显示绩效,但不显示排他性 | 风险中高:强劲的单项研究结果未必自动转化为重复授标 | 索取申办方留存,以及对 CRO / 竞争站点网络的赢输分析 |
| 采购 / 运营替代风险 | Iterative 把人员配置、监管、预算、病历审查和启动支持作为核心价值出售 | 如果客户自建这些能力,或选择其他供应商,定价权可能被压缩 | 索取服务相对纯技术产品的附加率,以及流失原因 |
| SKOUT 产品可见度缺口 | SKOUT 页面缺少具名客户,而竞争对手 GI Genius 对外打出强公开主张 | 器械侧集中度和采用风险难以量化;品类竞争已很清楚 | 索取装机基数、具名客户、复购 / 续约数据,以及按器械拆分的竞争赢输 |
| 国际多元化缺口 | Series C 提到印度和澳大利亚,但具名伙伴证据几乎都集中在美国 | 全球机会存在,但美国以外的执行质量和留存仍未被充分证明 | 索取美国以外活跃站点数、申办方组合和按地域拆分的绩效 |
公开证据在医疗机构网络扩张上最强,在申办方集中度和 SKOUT 客户集中度上最弱。
[CU018, CU034, CU035, CU036, CU037, CU038]6.6 图表与表格
07风险
7.1 监管、法律和报销风险
Iterative 的监管和法律风险,并非主要来自当下可见的执法事件,而是来自尚未稳定的运营边界。SKOUT 所在品类中,临床证据方向上为正,但在可推广性、假阳性、假阴性、技能退化和长期结果上仍有争议。BMJ 2025 年综述更进一步,把 AI 内镜界定为责任问题,责任如今横跨临床医生、机构、开发者和监管者。这一点很重要,因为主要机构对常规使用 CADe 仍有分歧,说明该品类尚未沉淀成无争议的护理标准。如果出现不良事件或令人失望的真实世界证据,问责不会整齐落在单一主体上。 报销侧同样重要。Medicare 明确覆盖筛查性结肠镜,但审阅过的政策和法律来源没有显示内镜 AI 已经有稳定、器械特定的支付路径。相反,政策讨论仍停留在拟议立法和 AI 赋能医疗器械的通用支付框架层面。与此同时,GI 医疗服务经济性承压:AGA 和 Becker’s 描述了内镜报销的削减和较长期侵蚀,这会削弱买方对附加技术和运营开销的胃口。Iterative 确实有基础合规设施——隐私政策、DPO、服务提供商披露和一般安全表述——但其公开安全姿态仍停留在高层次。按实际尽调口径,监管问题不是 Iterative 是否完全没有合规流程;而是如果审查加强,这套流程能否足够强地支撑一个数据密集的 AI 和站点运营模型。[CR010, CR011, CR013, CR014, CR016, CR017]
| 风险 | 证据 / 触发因素 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 尽调诉求 |
|---|---|---|---|---|---|---|
| AI 报销路径不确定性 | Medicare 尚无可预测的 AI 专属支付路径;拟议立法此前未推进 | 高 | 高 | 早期 | 高 | 索取当前支付方结构、任何 AI 专属编码策略,以及按站点类型拆分的 SKOUT 报销假设 |
| 品类指南分裂 | 2025 年主要机构对常规使用 CADe 分歧明显;证据确定性仍有争议 | 中 | 高 | 早期 | 高 | 索取内部医学事务对指南风险的判断、PMCF 计划和不良事件升级机制 |
| AI 责任 / 共同问责 | BMJ 综述称,一旦发生伤害,临床医生、机构、开发者和监管方共同承担责任 | 中 | 高 | 部分 | 中高 | 索取产品责任保险、赔偿结构和投诉处理流程 |
| 数据隐私 / 安全审查 | 隐私政策显示基线控制,但公开安全细节偏高层,且依赖供应商 | 中 | 中高 | 部分 | 中高 | 索取 SOC 2 / ISO 证据、DPA 组合、BAAs、供应商清单和事件历史 |
各行按当前剩余严重性排序,而不是按法律类别排序。
[CR010, CR016, CR017, CR018, CR035, CR036]7.2 运营、质量和安全风险
核心运营押注在于,Iterative 能否在把高接触网络扩到 100 多个站点并进入新专科时,仍保持试验表现优势。公开数据支撑上行情景:公司称启动快 2x、入组高 3.4x,DDW 2026 结果显示,在更大规模下仍大约维持 3x 基准随机化表现。这是有意义的缓释,反驳了该模型只是一组试点胜利的说法。但同一批来源也说明了模型为何脆弱。Iterative 不只是交付一台器械或 SaaS 仪表盘;它还提供预算、法律和监管支持、培训、赞助方互动、研究启动和人员管理。这意味着,即便底层软件保持稳定,质量漂移仍可能来自劳动力、流程或站点治理失灵。 AI 侧的运营风险,并不只是 SKOUT 或相近 CADe 工具能否在研究中奏效;关键是它们在日常使用中、跨不同用户、设备、病例组合和临床文化时,表现能否保持可靠。综述引用了假阳性、假阴性、数据集偏差、操作者过度依赖,以及部分分析中较弱的真实世界可复现性。公司自有医疗服务方材料强调持续的 PI 和 CRC 培训,外部访谈报道则称网络花了多年才建成,而且仍需要不断改进。两者合在一起,支持一种平衡看法:运营证据优于许多初创公司,但该模型仍重执行,因此对人员质量、培训纪律和流程控制高度敏感。数据安全属于同一类风险,因为研究运营、赞助方协调和 AI 工作流都会增加数据处理复杂度,而公开安全披露仍偏薄。[CR001, CR002, CR006, CR011, CR013, CR014]
| 失效模式 | 证据 | 可能性 | 严重性 | 缓释成熟度 | 剩余敞口 | 未解决缺口 |
|---|---|---|---|---|---|---|
| 100+ 个站点的质量漂移 | 模型依靠反复培训、人员管理和标准化启动,覆盖一个大型分布式网络 | 中 | 高 | 部分 | 中高 | 需要站点级方差、方案偏离数据,以及重新激活 / 补救率 |
| AI 假阳性 / 假阴性 / 泛化性 | 综述指出,尽管 ADR 提升,假阳性、假阴性、偏差和泛化性担忧仍在 | 中 | 高 | 部分 | 高 | 需要按病变类型、场景、设备栈和用户群体拆分的上市后表现 |
| 操作者过度依赖 / 技能退化 | 非 AI 检查在 AI 使用后 ADR 下降,显示治理薄弱时存在依赖风险 | 中 | 中高 | 早期 | 中高 | 需要培训方案、持续胜任力检查和使用护栏 |
| 数据处理 / 供应商安全复杂性 | 隐私政策提到多个服务提供商,安全表述仅停留在一般层面 | 中 | 中高 | 早期 | 中高 | 需要架构、供应商控制、审计结果和泄露历史 |
该登记表把网络运营风险与 AI 质量、安全风险合并呈现,因为这些失效在实践中会相互放大。
[CR002, CR010, CR011, CR013, CR014, CR028]7.3 伙伴、依赖和人员风险
Iterative 的公开进展真实存在,但集中度高。GI Alliance、One GI 和 US Heart & Vascular 主导具名医疗服务方记录,而且每段关系在战略上都足够重要,任何一方都可能对增长观感、赞助方进入和管线 密度产生超比例影响。GI Alliance 带来 21 个活跃站点和 80 多项试验;One GI 带来 34 家诊所和 13 个活跃站点;USHV 打开心脏病学,并强化多专科野心。这些是优势,也是依赖。如果主要伙伴表现不佳、重新定价、延迟扩张,或其经济重要性低于标题里的站点数暗示,外部叙事可能迅速恶化,因为太多证据都落在少数生态里。 产品和工作流层面还有第二层依赖。Medtronic 和 Olympus 不只是替代 AI 供应商;它们还是大型在位厂商,可以捆绑培训、硬件兼容性和装机基础杠杆。Olympus 明确把 CADDIE 绑定到其设备中枢和白光工作流,Medtronic 则围绕敏感性和低假阳性营销 GI Genius。因此,Iterative 不能只靠模型质量取胜,还要赢在渠道进入和工作流契合。人员风险叠加在上述因素之上。管理层梯队显然在改善——现在有总裁 / CFO、COO、医疗领导和法律领导——但这支梯队看起来仍是近期组建,而不是已经在当前规模下长期经历战斗检验。对创始人 Jonathan Ng 的依赖仍然明显,因为战略仍要求 AI、医疗服务方合作、赞助方商业化和多专科扩张同步执行。[CR003, CR004, CR005, CR007, CR008, CR021]
| 依赖 | 交易对手 / 系统 | 角色 | 集中度 | 失效情景 | 严重性 | 缓释 | 剩余敞口 |
|---|---|---|---|---|---|---|---|
| 大型医疗机构网络伙伴 | GI Alliance / One GI / USHV 合作伙伴 | 提供站点、患者和证据点 | 高 | 一个或多个伙伴扩张停滞、重新定价,或经济表现不及预期 | 高 | 拓宽多专科网络,并提高具名站点多样性 | 高 |
| 申办方 / CRO 需求 | 已披露的 40+ 家伙伴群体 | 供给试验库存,并支撑溢价定价逻辑 | 中高 | 申办方需求转弱,或多栖合作压低钱包份额 | 高 | 用绩效数据和预测深化重复授标 | 中高 |
| AI 内镜工作流栈 | Medtronic / Olympus / 站点设备选择 | 围绕集成、培训和硬件兼容性竞争 | 高 | 站点向既有厂商生态标准化,SKOUT 失去分发杠杆 | 高 | 在具名部署中证明更优结果和工作流适配 | 高 |
| 第三方服务提供商 | 托管 / 安全 / 分析供应商 | 支撑网站和运营工作流 | 中 | 供应商宕机或安全失误造成合规或服务中断 | 中高 | 合同控制与监测 | 中高 |
各行按该依赖对收入、证据质量或执行可信度的直接影响排序。
[CR021, CR022, CR023, CR025, CR026, CR027]| 角色 / 职能 | 依赖或缺口 | 可能性 | 严重性 | 缓释 | 剩余敞口 | 尽调路径 |
|---|---|---|---|---|---|---|
| 创始人 / CEO | Jonathan Ng 仍是 AI、医疗机构和申办方战略最清晰的单一整合者 | 中 | 高 | 更厚的高管梯队和董事会支持 | 中高 | 索取继任计划、运营节奏和决策权地图 |
| 财务 / 规模化纪律 | 总裁 / CFO 岗位直到 2025 年末才增设 | 中 | 中高 | Bill Kayser 带来 GI 与财务经验 | 中 | 索取 2026 年经营计划、KPI 包和资本配置框架 |
| 站点运营人员 | 模型依赖在众多站点招聘、培训并留住 CRC、PI 和支持人员 | 高 | 高 | 培训营、模块和集中化管理 | 中高 | 索取人员配比、空缺数据、流失率和站点支持 SLA |
| 多专科扩张 | 心脏病学和肥胖增加 TAM,也要求新的专科打法 | 中 | 高 | 以合作伙伴带动切入,并复用现有运营骨干 | 中高 | 索取专科级 KPI、试点结果,以及按垂直领域划分的领导归属 |
人员登记表强调跨越多个价值驱动因素的角色,而不是组织图上的每个高管头衔。
[CR007, CR008, CR024, CR028, CR040]7.4 财务 / 模型风险、缓释因素和投资逻辑破裂触发项
财务模型风险高,主要因为运营故事远比经济性更可见。公开材料支持网络规模、伙伴数量和表现指标,第三方报道也暗示 Series C 后资本充足。但仍没有公开收入结构、毛利率、烧钱速度、续约率、头部账户集中度或 SKOUT 装机基础披露。这让投资人只能在直接可见度有限的情况下,承保一个复杂的服务加 AI 模型,并判断价值主要流向赞助方服务、医疗服务方网络运营、SKOUT 销售,还是若干部分的组合。据报道的估值背景只会让这个问题更尖锐:一家公司可以资金充足,但如果关键变现杠杆不透明,经济性仍难承保。 因此,实际投资问题不是 Iterative 有没有风险——每家临床基础设施公司都有——而是这些风险能否监控,以及现有缓释是否足够强,足以支撑剩余风险暴露。最好的缓释因素很具体:真实伙伴网络、表现能在更大规模下保持的证据、更宽的高管梯队,以及至少基线的隐私 / 合规基础设施。主要投资逻辑破裂触发项同样具体:运营指标回落到行业常态、主要医疗服务方生态停滞或流失,或 AI 报销和客户经济性始终无法支撑持久定价。如果其中任何一项发生,今天让 Iterative 有吸引力的同一批特征——一体化模型和对伙伴网络的依赖——会把风险迅速传导到增长、毛利率和估值上。[CR016, CR018, CR019, CR020, CR031, CR032]
| 风险 | 指标 | 阈值 / 事件 | 行动含义 |
|---|---|---|---|
| 网络绩效漂移 | 随机化和启动指标 | 持续下滑,逼近已发布行业基准 | 除非能用刻意的组合结构调整解释,否则视为投资假设破裂 |
| 伙伴集中度 | 头部医疗机构生态的收入或研究占比 | 前 1-3 个网络对经济性的主导程度超过公开证据已暗示的水平 | 提高集中度折价,并要求合同保护 |
| 报销风险 | AI 赋能内镜的支付方路径 | GI 医疗机构预算收紧时,可变现报销路径没有进展 | 下调 SKOUT 上行空间,重新测算器械贡献 |
| 商业防御性 | SKOUT 对 Medtronic / Olympus 的具名赢单 | 持续缺少具名部署,或在既有厂商账户中持续失利 | 假设器械护城河更弱,交叉销售价值更低 |
| 安全 / 合规 | 审计级控制和事件历史 | 重大控制缺口、不良事件或未披露事件 | 上调法律 / 合规风险,并重新审视信任假设 |
| 管理层厚度 | 领导节奏和运营 KPI | 高管高流失,或新专科归属薄弱 | 提高执行折价,并缩短对融资 / 现金续航的信心 |
| 财务韧性 | 毛利率、现金消耗和续约画像 | 低毛利服务组合,或申办方重复经济性弱 | 降低估值支撑,并要求更严的下行情形 |
| 规模化证据 | 站点级方差和方案遵从 | 离散度大,且缺少强补救 | 把中央骨干优势视为不如其声称的持久 |
这些指标设计成尽调中或未来组合监控中都能检查。
[CR002, CR016, CR019, CR021, CR025, CR034]7.5 图表与表格
08估值
8.1 投资逻辑、反向逻辑和建议判断
Iterative Health 已经值得认真讨论估值,因为其运营证据优于平均 AI 医疗初创公司。官方材料显示,研究网络超过 100 个站点,赞助方和 CRO 关系超过 40 家,并反复声称启动速度比行业基准快 2x、入组表现高 3x。DDW 2026 更新尤其重要,因为它暗示这一逻辑在更大规模下仍成立,而不只是存在于试点队列内。上述组合支撑正面情景:Iterative 似乎已经建成一个有差异化、以站点为中心的临床研究网络,具备真实工作流和赞助方相关性,而不只是一个狭窄的内镜算法故事。 反向逻辑在于,公开记录给上行定价很激进,却没有同样解释经济性。第三方来源集中在 $1.3 billion 到 $1.4 billion 标记附近,但公司在最新融资或表现更新中没有披露收入、ARR、毛利率或留存。因此,公开证据更能支持公司质量,而不是支持精确承保。在当下市场,这一区别很重要:AI 和数据资产仍有溢价倍数,但前提是收入质量、耐久性和变现清晰度可见。因为 Iterative 的价格只有在强执行情景下才显得合理,而在更接近 CRO 的情景下显得偏高,最能被公开证据支撑的建议是继续研究,置信度中等,风险高,估值立场偏高。[CV003, CV004, CV005, CV014, CV017, CV018]
| 维度 | 评估 | 公开证据依据 | 决策含义 |
|---|---|---|---|
| 建议 | 继续研究 | 公司质量看起来真实,但价格已是后期阶段,经济性仍不透明 | 不要仅凭公开资料为入场定价 |
| 信心 | 中 | 融资和运营证据能较好互相印证;收入和利润率不能 | 用尽调验证或推翻投资假设,不是微调它 |
| 风险评级 | 高 | 执行、报销、伙伴集中度和财务不透明风险同时存在 | 如果尽调不达预期,先假设下行空间不小 |
| 估值立场 | 偏高 | 当前估值只有在高溢价执行情形下才显得合理 | 宁可平价或更低入场,不为新一轮加价买单 |
| 入场纪律 | 没有数据室,就不支付新溢价 | 公开证据不足以支撑高于据称 $1.3B-$1.4B 区间的价格 | 任何 IC 建议上调前,要求完整 KPI 包和条款 |
| 上调条件 | 收入、利润率和条款必须越过溢价区间门槛 | 需要证明经济性更接近 AI/data 溢价,而不是 CRO 底线 | 只有披露指标支撑价格,才转为跟踪 / 买入 |
判断仅反映公开证据,且有意对价格敏感,而不是泛化的公司质量评分。
[CV042, CV043, CV049, CV050, CV051, CV052]| 主题 | 正向假设 | 反向假设 | 什么会改变判断 |
|---|---|---|---|
| 运营验证 | 100+ 个站点、40+ 个合作伙伴、激活速度快 2x、入组 >3x,说明对申办方和站点有真实价值 | 现有验证更多指向吞吐量,而非经济性;变现质量仍被遮住 | 按申办方 / 站点队列披露收入,并给出规模化项目利润率 |
| 平台范围 | 拓展到心脏病学和肥胖症,可能把网络价值拉出 GI 之外 | 相较 GI,扩科仍处早期;收入兑现前,可能先拖慢执行 | 证明多专科预订、留存和贡献利润率 |
| 融资质量 | 投资人阵容强,估值较前轮上调,说明投资需求仍在 | 公开股权结构细节不足;优先股堆叠和稀释会显著改写回报 | 提供完整公司章程、投资人权利和股权结构堆叠 |
| 估值语境 | 如果 Iterative 已接近优质 AI / 数据平台,$1.3B-$1.4B 可能合理 | 如果经济性仍更像 CRO 服务,同一估值就显得偏高 | 披露当前收入、毛利率、续约 / 留存 |
| 报销 / AI 经济性 | AI 辅助工作流可能带来差异化,也给未来变现加杠杆 | Medicare 路径仍未定,GI 支付压力会拖慢采用或削弱定价权 | 披露装机基础、定价、报销假设和附着率 |
各行只拆出最可能改变投资决策的少数变量,而不是复述完整报告。
[CV003, CV004, CV005, CV014, CV015, CV017]建议维持「继续研究」:运营证据确实存在,但按当前价格看,经济性和变现披露仍不足。
[CV003, CV004, CV005, CV014, CV017, CV019]公司在证据和可选性上得分较高,但经济性可见度和估值支撑偏弱。
[CV017, CV023, CV035, CV043, CV050, CV051]8.2 当前估值语境和可比公司组
当前估值语境足以建立纪律框架,但不足以支撑激进确信。Forge 显示 2026 年 4 月 Series C 估值为 $1.3 billion,TechCrunch 在 2026 独角兽追踪器中将 Iterative 列为 $1.4 billion;ai2.work 也报告约 $1.3 billion 标记。公司官方材料确认 $77 million 融资,但不确认投后 估值。Forge 只补充了部分融资条款线索,包括 1.0x 非参与清算优先权和 8% 股息率,因此公开记录中的下行保护和稀释仍不完整。 公开可比公司提示谨慎。IQVIA 和 ICON 交易在大约 1x-2x 收入附近,Medpace 则接近大约 5x 收入,并且在这组公司中呈现最佳的公开可见增长和利润率组合。与此同时,当前 2026 年 M&A 评论显示,优质医疗科技 标的成交约 4x-6x 收入,只有高端 AI / 数据资产能推到 6x-8x 以上区间。这意味着,只有当 Iterative 的业务已经明显更接近高端 AI / 数据经济性,而不是服务密集型 CRO 画像时,当前标记才站得住。由于收入未披露,单凭公开证据无法确认。[CV009, CV010, CV011, CV014, CV015, CV016]
| 可比公司 / 参照 | 2026 指标快照 | 隐含倍数 / 估值 | 如何映射 Iterative | 局限 |
|---|---|---|---|---|
| IQVIA | $30.14B 市值 / $16.63B TTM 收入 | ~1.8x 收入 | 为规模化医疗研究基础设施提供多元化 CRO / 数据估值底线 | 规模更大、已上市,也比 Iterative 更多元 |
| ICON | $10.53B 市值 / $8.10B TTM 收入 | ~1.3x 收入 | 服务执行占比较高时,全球 CRO 是基准 | AI / 数据可选性弱于 Iterative 所声称的空间 |
| Medpace | $12.72B 市值 / $2.67B TTM 收入;Q1 收入 +26.5%,EBITDA 利润率 21.1% | ~4.8x 收入 | 最适合映射高质量研究执行和增长的上市样本 | 仍是上市盈利公司,成熟度远高于 Iterative |
| 2026 年优质医疗科技并购区间 | 普通医疗科技 ~4x-6x;优质 AI / 数据 ~6x-8x+ | 市场区间,非公司特定倍数 | 给优质混合模式可能拿到的倍数划定范围 | 交易区间高度取决于质量、监管和持续性 |
| Iterative 当前私募估值 | 第三方估值语境约 $1.3B-$1.4B;累计融资约 $271M-$273M+ | 后期私募独角兽估值 | 当前讨论的实际入场价 | 未披露收入、毛利率或留存,倍数缺少锚 |
| 2026 年独角兽重估语境 | PitchBook 称许多独角兽估值已过期;AI 公司生成速度仍异常高 | 不是直接倍数,而是折价框架 | 支持谨慎看待未被公开基本面刷新过的私募估值 | 属于市场语境,不是直接经营可比 |
样本是部分而非穷尽,混合了上市可比、市场区间和私募估值语境,因为 Iterative 同时带有 CRO、AI 和私募独角兽属性。
[CV019, CV020, CV021, CV022, CV023, CV024]只有收入进入高溢价 AI / 数据区间上沿后,当前估值标记才更容易站住脚。
条形图用来自公开可比公司和 2026 年医疗科技并购评论的多档倍数,展示企业价值结果;并非公司指引。
[CV023, CV024, CV036, CV037, CV038]8.3 乐观、基准和悲观情景区间
这里的所有情景工作都明确基于假设,因为当前收入和利润率未披露。乐观情景假设 Iterative 把网络证据转化为多专科收入,表现继续显著高于基准,并获得上沿 AI / 数据倍数。在这条路径下,约 $1.8 billion 到 $2.6 billion 的退出有可能实现;但即便如此,从 $1.3 billion 到 $1.4 billion 起始标记出发,也不会形成经典后期风险投资回报。基准情景更保守:Iterative 作为强 CRO / 数据混合体扩张,收入落在约 $160 million 到 $220 million,并以 5x-7x 倍数退出,价值约 $0.9 billion 到 $1.5 billion。 悲观情景重要,因为今天的私有市场环境比 2021 年更不宽容。如果报销摩擦、伙伴集中或执行滑坡把 Iterative 推向类 CRO 经济性,标记可能压缩到约 $0.4 billion 到 $0.9 billion。这一下行不是遥远幻想;PitchBook 和 Finro 都描述了一个市场:当变现单薄或陈旧标记终于被检验时,公司会被重新定价。因此,概率加权图景更接近“基本已经定价”,而不是“明显便宜”,这也是本章落在继续研究和偏高,而不是合理和买入的原因。[CV023, CV024, CV025, CV027, CV036, CV037]
| 情景 | 核心假设 | 退出倍数逻辑 | 退出估值区间 | 以 $1.35B 入场价测算的回报含义 | 概率信号 |
|---|---|---|---|---|---|
| 乐观 | 收入大约达到 $250M-$320M;多专科扩张开始变现;执行效率持续超过基准 >2x | 7x-8x 优质 AI / 数据区间 | $1.8B-$2.6B | ~1.2x-1.9x MOIC,未计优先权影响 | 需要证明经济性已接近优质 AI / 数据资产 |
| 基准 | 收入大约达到 $160M-$220M;公司表现为强 CRO / 数据混合体 | 5x-7x 混合区间 | $0.9B-$1.5B | ~0.7x-1.1x MOIC(回报倍数) | 与当前公开证据最一致 |
| 悲观 | 收入大约停在 $110M-$170M;报销拖累或执行失手导致估值压缩 | 3x-4.5x CRO / 重估区间 | $0.4B-$0.9B | ~0.3x-0.7x MOIC(回报倍数) | 如果眼下增长好看、后续变现跟不上,就会触发 |
情景测算基于假设,因为当前收入和利润率未披露;MOIC 仅作示意,不包括精确稀释 / 优先权影响。
[CV036, CV037, CV038, CV039, CV040, CV041]按当前披露估值标记,基准情景回报有限,下行空间仍然显著。
区间基于情景推演,未纳入精确稀释或优先证券清偿瀑布,因为股权结构条款并未完全公开。
[CV014, CV046, CV047, CV048, CV049]8.4 最终尽调问题和投资逻辑破裂触发项
剩余阻碍不是表面问题。若要从继续研究转向观察或买入,投资人需要当前收入、毛利率、留存、集中度和优先股堆叠数据,而不是另一个伙伴 标识或吞吐量轶事。报销也比公司叙事暗示的更重要。本报告审阅的法律和政策来源仍把 AI 赋能器械的 Medicare 报销描述为一项演进中的政策工程,而 GI 支付来源则描述了内镜经济性的真实压力。这不会扼杀 SKOUT 或更广平台逻辑,但确实意味着,在管理层证明相反情况之前,估值情景应假设存在变现摩擦。 监控框架很直接。如果实际经济性远弱于当前标记要求、网络表现回落到基准附近,或多专科扩张未能转化为真实收入,投资逻辑就会破裂。反过来,如果尽调显示当前收入已经支撑高端 AI / 数据倍数区间,赞助方和医疗服务方集中度可接受,且股权结构表没有异常苛刻条款,判断可能改善。在那之前,IC 的正确姿态是有纪律的好奇,而不是确信。[CV030, CV031, CV032, CV033, CV034, CV035]
| 触发器 | 阈值 | 对投资逻辑的传导 | 行动含义 |
|---|---|---|---|
| 经济性不达标 | 当前收入 / 毛利率显著低于 $1.3B-$1.4B 所需水平 | 当前估值不再像优质执行,开始像定价过高 | 叫停,或要求入场价大幅下调 |
| 业绩回落 | 网络 KPI 在激活、入组或随机化上回落到基准附近 | 核心运营差异化削弱 | 立即下调乐观 / 基准假设 |
| 报销无法变现 | AI / 内镜经济性仍被打包收费,或不能报销,且缺少清晰定价权 | SKOUT 式上行仍只是可选性,不是现金流 | 从模型中剔除软件溢价假设 |
| 扩张失手 | 心脏病学 / 肥胖症无法转化为已预订收入或复购申办方 | 多专科可选性消失 | 基准情景回到仅 GI、低倍数结果 |
| 股权结构拖累 | 优先股堆叠或附函条款让新资金显著处于劣后位置 | 即使企业价值不变,回报曲线也恶化 | 暂停,直到条款重谈或价格重置 |
触发器聚焦会迫使立即重写情景的具体事件,而不是泛泛的业务风险。
[CV015, CV035, CV039, CV041, CV045]| 主题 | 缺失证据 | 重要性 | 负责人 / 尽调路径 |
|---|---|---|---|
| 收入桥 | 按申办方服务、站点网络费用和 SKOUT / 设备收入拆分的 2025 年审计收入与 2026 YTD | 决定当前估值是收入的 6x、10x,还是 14x+ | CFO 资料室 / 月度 KPI 包 |
| 利润率质量 | 毛利率、按服务线拆分的贡献利润率,以及现金消耗 / 现金跑道 | 区分可扩展平台经济性和人力密集执行 | 财务尽调 + 董事会材料 |
| 留存 / 集中度 | 前 10 大申办方收入占比、合作伙伴集中度、队列续约和流失 | 检验网络投资逻辑的持续性和下行风险 | 客户分析 + 合同审阅 |
| 股权结构表 / 优先权 | 当前公司章程、清算优先序列、股息、按比例认购权和附函 | 把企业价值情景转成投资人回报情景时必需 | 法律尽调 + 资料室 |
| 多专科变现 | 心脏病学和肥胖症管线、预订、激活速度和贡献利润率 | 乐观情景需要 GI 之外的真实收入 | 商业尽调 + 专科运营审查 |
| SKOUT 变现 | 装机基础、定价、报销假设,以及与研究关系的附着率 | 缺少这组证据,AI 上行仍主要是叙事 | 商业 / 报销工作流 |
要求按其改变估值信心和价格纪律的直接程度排序。
[CV044]免责声明
本报告是基于公开证据的尽调快照,不构成投资建议。重要的财务、法律、技术和合同事实仍未公开;作出任何投资决定前,应直接向管理层和原始文件核验。
证据索引
| 编号 | 陈述 | 可信度 | 来源 |
|---|---|---|---|
| CO001 | Iterative Health was founded in 2017. | 高 | SO002, SO008 |
| CO002 | Iterative Health currently identifies Cambridge, Massachusetts, and New York, New York, as its headquarters locations. | 高 | SO002, SO005 |
| CO003 | Iterative Health describes itself as a healthcare technology and services company powering the acceleration of clinical research. | 高 | SO001, SO002, SO005 |
| CO004 | The company’s current operating focus spans GI, hepatology, obesity, and cardiology in community-based care settings. | 高 | SO001, SO012, SO015 |
| CO005 | Iterative Health continues to market SKOUT as a real-time AI product for polyp detection alongside its clinical-research network offering. | 高 | SO001, SO003, SO010 |
| CO006 | Jonathan Ng, MBBS, is the founder and CEO of Iterative Health. | 高 | SO002, SO011 |
| CO007 | Early company materials describe Iterative Scopes as a 2017 MIT spinout whose founding concept was developed while Jonathan Ng was at MIT and Harvard. | 高 | SO008, SO011 |
| CO008 | Public materials indicate the business broadened from the Iterative Scopes precision-gastroenterology identity into the wider Iterative Health multispecialty clinical-research platform. | 中 | SO002, SO005, SO007 |
| CO009 | Bill Kayser joined Iterative Health as President and Chief Financial Officer in October 2025. | 高 | SO006, SO019 |
| CO010 | Before joining Iterative Health, Bill Kayser had been CFO of GI Alliance and Prospero Health and had led corporate strategy and M&A work at McKesson. | 高 | SO006, SO019 |
| CO011 | Iterative Health publicly described itself as having 250+ employees worldwide in late 2025. | 高 | SO006, SO013 |
| CO012 | A May 2026 company release described Iterative Health as a 250-person team with more than 100 sites on three continents. | 高 | SO012, SO014 |
| CO013 | Iterative Health’s current team page lists Dana Feuchtbaum as Chief Operating Officer. | 高 | SO002, SO015 |
| CO014 | Iterative Health’s current team page lists Nadege Gunn, MD, as Chief Medical Officer for Hepatology and Obesity. | 中 | SO002, SO004 |
| CO015 | Iterative Health announced a $77 million Series C in 2026 led by Intrepid Growth Partners and GV, with EDBI joining and Insight Partners plus Obvious Ventures participating. | 中 | SO005, SO023 |
| CO016 | The 2026 Series C added Ajay Agrawal to the board and Anthony Philippakis as a board observer. | 中 | SO005, SO023 |
| CO017 | Official 2026 materials state that Iterative Health’s network includes more than 100 research sites and more than 40 pharmaceutical, biotech, medical device, and CRO partners. | 高 | SO005, SO012 |
| CO018 | Official company materials say the network delivers roughly 2x faster site activation and 3x higher patient enrollment than industry benchmarks in IBD trials. | 高 | SO005, SO015 |
| CO019 | Iterative Scopes raised a $150 million Series B in January 2022 co-led by Insight Partners and Clearlake. | 高 | SO007, SO017, SO018 |
| CO020 | The January 2022 Series B announcement said the company had raised a total of $182 million in 2021 including the $30 million Series A. | 高 | SO007, SO017 |
| CO021 | Iterative Scopes said its August 2021 Series A raised $30 million after a $13.5 million seed round and included investors such as Obvious Ventures, Eli Lilly, JJDC, Breyer Capital, and Seae Ventures. | 中 | SO008, SO021 |
| CO022 | The Series B round brought Insight Partners managing director Lonne Jaffe onto the board. | 高 | SO007, SO017 |
| CO023 | The Series A announcement said Eli Lilly’s Lotus Mallbris and Obvious Ventures’ Nan Li would join the board. | 中 | SO008 |
| CO024 | The April 2022 acquisition of CRSG and Precision Research brought Chris Fourment into the senior leadership team and deepened Iterative’s ties to IBD clinical-trial operations. | 中 | SO009, SO024 |
| CO025 | Iterative Scopes announced in September 2022 that SKOUT had received FDA clearance for adults undergoing colorectal cancer screening or surveillance colonoscopy. | 高 | SO010, SO020 |
| CO026 | FDA materials describe SKOUT as a real-time computer-aided detection aid that highlights suspected colorectal polyps but does not replace physician judgment or primary interpretation. | 高 | SO020, SO010 |
| CO027 | Official SKOUT materials report a 27% relative increase in adenomas per colonoscopy and a 44% relative increase in 5-9 mm proximal-colon polyp detection. | 高 | SO010, SO020 |
| CO028 | GI Alliance said its April 2025 partnership with Iterative Health added 21 active research sites conducting more than 80 Phase II-IV trials and extended Iterative’s network to more than 70 locations in the U.S. and Europe. | 中 | SO016 |
| CO029 | Iterative Health said One GI’s network includes 34 clinics across six states and 13 active research sites that became part of Iterative’s global site network. | 中 | SO013 |
| CO030 | The March 2026 US Heart & Vascular partnership marked Iterative Health’s expansion into cardiology and pushed the company’s network above 100 sites across three continents. | 高 | SO012, SO014 |
| CO031 | The May 2026 NextStage transaction added three cardiology research sites in Beaumont, Port Arthur, and Waco, Texas. | 中 | SO014 |
| CO032 | At DDW 2026, Iterative Health said its network sustained more than 3x industry-benchmark enrollment performance at expanded scale, with a per-study randomization rate of 0.33 patients per site per month across U.S. and European sites over two years. | 中 | SO015 |
| CO033 | TechCrunch listed Iterative Health among 2026 unicorns at approximately a $1.4 billion valuation with more than $270 million in total funding to date. | 中 | SO022 |
| CO034 | Forge’s secondary-market profile listed Iterative Health at a roughly $1.3 billion Series C valuation and $272.32 million of total funding, but the same page misdescribed the company’s operations and should be treated as indicative rather than definitive. | 低 | SO021 |
| CO035 | AI2.work summarized the 2026 Series C as implying roughly a $1.3 billion valuation and more than $273 million of total funding. | 低 | SO023, SO005 |
| CO036 | Current headcount disclosures conflict: Tracxn estimated 166 employees in April 2026, versus company statements of 250+ employees in late 2025 and a 250-person team in May 2026. | 中 | SO024, SO006, SO014 |
| CO037 | Tracxn also listed Iterative Health as a Series C company with about $271 million of total funding and Cambridge headquarters, broadly consistent with other secondary trackers but not definitive. | 中 | SO024, SO022 |
| CO038 | Iterative Health’s team page lists both a Head of Ex-US Markets and an EU Site Network Lead, supporting an international operating footprint beyond a purely U.S. office base. | 中 | SO002, SO005 |
| CO039 | Official 2025-2026 materials identify GI Alliance, One GI, and US Heart & Vascular as anchor provider partners as Iterative broadens from GI and hepatology into obesity and cardiology. | 高 | SO005, SO012, SO013, SO016 |
| CO040 | An independent 2025 CADe literature review concluded that systems such as SKOUT can improve detection but still face false positives, automation bias, workflow-integration, generalizability, and long-term-outcome questions. | 高 | SO025, SO020 |
| CO041 | The same 2025 review listed SKOUT among FDA-approved CADe products commercially available by 2025. | 高 | SO025, SO020 |
| CO042 | In the reviewed public materials, Iterative Health did not disclose current revenue or ARR. | 中 | SO001, SO005, SO021, SO024 |
| CO043 | In the reviewed public materials, Iterative Health did not disclose debt facilities, secondary-liquidity amounts, or a fully reconciled current cap table. | 中 | SO005, SO021, SO024 |
| CM001 | Iterative Health participates in two related but distinct markets: AI-assisted colonoscopy quality tooling and GI or community clinical-trial site-network services. | 中 | SM018, SM019, SM020, SM025 |
| CM002 | For this chapter, included spend is procedure-room detection and quality tooling plus sponsor- or provider-funded research-site infrastructure, while broad non-invasive screening, generic GI software, and full-service CRO spend sit outside the core market boundary. | 中 | SM001, SM015, SM018, SM019, SM020 |
| CM003 | The American Cancer Society said in 2026 that more than 20 million Americans eligible for colorectal cancer screening are not tested as recommended, or about one in three adults. | 中 | SM001 |
| CM004 | Among U.S. adults aged 45 to 49, up-to-date colorectal cancer screening prevalence reached 33.7% in 2023 after sitting near 20% in 2019 and 2021. | 中 | SM021 |
| CM005 | Within that 45 to 49 cohort, 2023 colonoscopy prevalence was 27.7% while stool-based testing prevalence was 7.1%. | 中 | SM021 |
| CM006 | Screening gains for adults aged 45 to 49 were concentrated among higher-educated and privately insured people, while less-educated and uninsured groups showed little change. | 中 | SM021 |
| CM007 | The 2024 ACG/ASGE quality update expanded ADR measurement to screening, surveillance, and diagnostic colonoscopy in patients aged 45 and older while excluding positive non-colonoscopy tests, IBD surveillance, and known-neoplasia therapy. | 高 | SM002, SM004, SM006 |
| CM008 | The same update set ADR-A benchmarks at 35% overall, 40% in men, and 30% in women and added a 6% sessile serrated lesion detection target. | 高 | SM004, SM006 |
| CM009 | GI commentary says broader ADR-A measurement reduces indication misclassification and increases denominator size, but practices with more surveillance exams may also report higher overall ADRs. | 中 | SM005, SM006 |
| CM010 | Updated quality programs also expect about 90% adequate bowel preparation and 90% interval adherence, extending buyer pressure beyond lesion detection alone. | 中 | SM004, SM006 |
| CM011 | Yale's summary of a 44-study meta-analysis said CADe increases adenoma detection but only slightly increases advanced colorectal neoplasia detection. | 中 | SM007 |
| CM012 | Yale also said draft AGA guidance made only a conditional recommendation for CADe use while reporting little to no adverse events and no clinically meaningful increase in procedure time. | 中 | SM007 |
| CM013 | Olympus' EAGLE trial reported a 7.3 percentage-point ADR increase in screening and surveillance colonoscopies plus substantial relative increases in large adenoma, non-polyploid adenoma, and SSL detection without workflow disruption. | 中 | SM008 |
| CM014 | The FDA's May 2025 510(k) letter cleared the SKOUT system as a Class II gastrointestinal lesion software detection system for adult screening or surveillance colonoscopy. | 中 | SM025 |
| CM015 | FDA documentation shows SKOUT is an informational visual aid that pauses detection when tools are present or lighting is inadequate, reinforcing that CADe is adjunctive rather than autonomous. | 中 | SM025 |
| CM016 | Iterative markets SKOUT as real-time AI for polyp detection aligned with rising colonoscopy quality expectations and claims no increase in procedure time. | 中 | SM018 |
| CM017 | The 2026 CADe narrative review said more than a dozen CADe systems were commercially available as of 2025, placing Iterative in a crowded vendor landscape rather than a greenfield category. | 中 | SM022 |
| CM018 | The same review said CADe commercialization has outpaced real-world evidence, leaving open questions on false positives, automation bias, operator deskilling, generalizability, and long-term outcomes. | 中 | SM022 |
| CM019 | An earlier AI review likewise said CADe improves ADR, PDR, and consistency but lacks long-term effectiveness data. | 中 | SM009 |
| CM020 | Iterative's JCC blog said its network achieved median site selection-to-activation in 74 days versus published 122 to 171 day benchmarks and 83 days to first patient randomized versus a 140 day benchmark. | 中 | SM010 |
| CM021 | DocWire's interview with the lead study author independently repeated the 74-day activation figure and said the model saves about three months versus benchmark. | 中 | SM011 |
| CM022 | Iterative's JCC blog said the network's weighted average randomization rate was 0.34 patients per site per month across phase 2 and phase 3 UC and CD trials, more than three times the 2020 benchmark. | 中 | SM010 |
| CM023 | DocWire independently repeated the 0.34 patients per site per month result versus an industry benchmark around 0.10. | 中 | SM011 |
| CM024 | Iterative's DDW 2026 release reported similar performance at larger scale, with 0.33 patients per site per month overall and 0.32 median among fully enrolled trials against the same 0.10 benchmark. | 中 | SM023 |
| CM025 | The DDW 2026 release said those later results reflected roughly two-fold expansion in trials and sites versus the earlier ECCO analysis. | 中 | SM023 |
| CM026 | Iterative's 2026 Series C release said its global network includes more than 100 research sites and more than 40 pharma, biotech, medical-device, and CRO partners. | 中 | SM019 |
| CM027 | Clinical Trials Arena independently reported the same greater-than-100-site and greater-than-40-partner network scale. | 高 | SM024, SM019 |
| CM028 | GI Alliance's 2025 partnership added 21 active research sites and over 80 phase II to IV trials, bringing the combined network to more than 70 locations in the U.S. and Europe at that time. | 中 | SM020 |
| CM029 | Iterative's Series C release said more than half of research sites enroll one or fewer patients per study and nearly 90% of U.S.-based trials fail to meet enrollment targets on time. | 中 | SM019 |
| CM030 | Clinical Trials Arena likewise framed site recruitment as one of sponsors' biggest bottlenecks and noted regulator pressure for broader patient diversity. | 中 | SM024 |
| CM031 | Medicare Part B covers screening colonoscopy with no minimum age and covers follow-up colonoscopy after positive stool- or blood-based tests as a screening test. | 高 | SM015, SM001 |
| CM032 | ACS' 2026 guideline broadened screening options to include blood-based and new at-home stool tests, meaning CADe vendors compete in a screening ecosystem where colonoscopy is important but not the only entry point. | 中 | SM001 |
| CM033 | The reviewed payment sources did not surface a dedicated Medicare add-on reimbursement pathway for AI-assisted colonoscopy, so purchase decisions depend on provider-side ROI rather than payer-created incremental revenue. | 中 | SM012, SM013, SM016 |
| CM034 | AGA, ACG, and ASGE said the 2026 CMS proposed rule would cut physician payments for ASC and hospital-based GI endoscopy and E/M by $58 million, including average 8% endoscopy cuts for ASC and HOPD practices. | 中 | SM012 |
| CM035 | Becker's said Medicare paid about $220 per colonoscopy in 2025, while GI leaders described long-term reimbursement erosion and only a 2.9% ASC update as insufficient versus cost growth. | 中 | SM013 |
| CM036 | ACG's 2026 Medicare reform blog said diagnostic colonoscopy reimbursement for CPT 45378 fell 15% from 2019 while the Medicare Economic Index rose 24%. | 中 | SM014 |
| CM037 | Boston Scientific's 2026 guide said Medicare reimbursement differs materially by site of service, with ASCs paid less than hospital outpatient departments for comparable endoscopy codes. | 中 | SM016 |
| CM038 | The MACRA cost-measure paper said site of service is the biggest driver of colonoscopy cost variation and that hospital outpatient procedures cost more than the same colonoscopy in an ASC. | 中 | SM017 |
| CM039 | The same paper said episode-based cost measures also make providers accountable for anesthesia, pathology, and post-procedure utilization inside the colonoscopy episode. | 中 | SM017 |
| CM040 | Yale reported its VA implementation treats CADe as a small software and hardware addition activated with a button press, implying adoption can be incremental at the procedure-room level. | 中 | SM007 |
| CM041 | Olympus pitched cloud deployment as reducing hardware dependence and enabling subscription procurement, suggesting vendors are adapting go-to-market to budget friction rather than waiting for new reimbursement codes. | 中 | SM008 |
| CM042 | For site-network services, the economic buyer is primarily sponsor clinical operations or CRO outsourcing teams seeking centralized access to diverse, higher-performing community sites. | 中 | SM019, SM024 |
| CM043 | GI Alliance said most care for chronic GI diseases like IBD and MASH occurs in community practice, supporting the idea that community GI groups are strategically important research sites rather than peripheral channels. | 中 | SM020 |
| CM044 | Iterative's site-network materials attributed performance gains to centralized regulatory and contracting support, study-specific training, and technology-enabled patient identification embedded directly in routine care. | 中 | SM010, SM023 |
| CM045 | DocWire said the model reduces administrative burden by keeping sites inside a persistent operating relationship rather than restarting processes from scratch for every study. | 中 | SM011 |
| CM046 | The public proof for Iterative's research-network business is strongest in GI and IBD; expansion into hepatology, obesity, and cardiology is a growth plan rather than an equally evidenced operating line. | 中 | SM019, SM023, SM024 |
| CM047 | Public sources do not disclose Iterative's CADe installed base, pricing, or procedure penetration, so a precise SAM or SOM for AI-assisted colonoscopy cannot be defended from the reviewed record. | 中 | SM018, SM019, SM024 |
| CM048 | Public sources likewise do not disclose sponsor contract values, backlog, revenue per site, or take rates, so a precise SAM or SOM for the GI trial-network services business remains unresolved. | 中 | SM019, SM020, SM024 |
| CM049 | Community GI practices receive a co-benefit from the site-network model because patients can access trials without leaving everyday care settings even when sponsor budgets fund the core service. | 中 | SM010, SM020 |
| CM050 | The JAMA disparity findings and ACS emphasis on coverage and affordability imply that screening demand will not automatically translate into uniform colonoscopy or AI-tool adoption across subpopulations. | 中 | SM001, SM021 |
| CP001 | Iterative Health's direct procedure-room offering is SKOUT, a real-time AI product for polyp detection during colonoscopy. | 中 | SP001 |
| CP002 | Iterative says SKOUT is trained to spot clinically significant lesions and is enriched for sessile polyp detection. | 中 | SP001 |
| CP003 | Iterative says endoscopists using SKOUT resected one additional adenoma for every 4.5 patients examined without increasing resection of hyperplastic polyps. | 中 | SP001 |
| CP004 | Medtronic markets GI Genius as an intelligent endoscopy module that helps physicians detect colorectal polyps of different sizes, shapes, and morphologies. | 中 | SP003 |
| CP005 | Medtronic says GI Genius has 99.7% sensitivity and less than 1% false positives. | 中 | SP003 |
| CP006 | Olympus positions CADDIE as a cloud-based CADe add-on inside the OLYSENSE hub that connects to Olympus imaging equipment. | 高 | SP004, SP005 |
| CP007 | Olympus says the CADDIE algorithm was trained on 162,207 image frames containing 1,711 polyps from 906 patients. | 中 | SP004 |
| CP008 | Olympus says OLYSENSE is part of the EVIS X1 intelligent endoscopy ecosystem. | 中 | SP005 |
| CP009 | Olympus says the 2025 OLYSENSE CAD/AI launch included multiple cloud applications in Europe but only detection-focused CADDIE in the United States. | 中 | SP008 |
| CP010 | Olympus says ENDO-AID is fully compatible with several EVIS X1, EXERA III, and LUCERA ELITE endoscopes. | 中 | SP006 |
| CP011 | Olympus says ENDO-AID offers two detection modes that differ in sensitivity and specificity. | 中 | SP006 |
| CP012 | Olympus says the EAGLE trial found a 7.3 percentage-point ADR increase for CADDIE-assisted colonoscopy versus standard colonoscopy in screening and surveillance patients. | 高 | SP007, SP008 |
| CP013 | Olympus says the EAGLE trial improved detection of large adenomas, flat adenomas, and sessile serrated lesions. | 中 | SP007 |
| CP014 | Fujifilm markets CAD EYE as its AI-based computer-aided endoscopy offering. | 中 | SP009, SP010 |
| CP015 | Fujifilm says CAD EYE runs through the EX-1 expansion unit and compatible Fujifilm processors and colonoscopes. | 中 | SP010 |
| CP016 | Odin Vision markets CADDIE as cloud-based AI for endoscopy that supports detection and characterization of polyps during colonoscopy. | 中 | SP011 |
| CP017 | Odin Vision says CADDIE requires minimal training and supports hands-free interaction during procedures. | 中 | SP011 |
| CP018 | Yale's summary of a 44-study meta-analysis says CADe systems increase adenoma detection but only slightly increase advanced colorectal neoplasia detection. | 中 | SP012 |
| CP019 | AGA makes no recommendation for or against CADe-assisted colonoscopy because evidence on CRC incidence, CRC mortality, and PCCRC remains very low certainty. | 高 | SP015, SP012 |
| CP020 | AGA highlights unresolved evidence gaps on cost-effectiveness, access, diverse settings, advanced lesions, and long-term outcomes for CADe. | 中 | SP015 |
| CP021 | GI & Hepatology News reported that a TriNetX analysis of more than 1.5 million matched patients linked AI-assisted colonoscopy to a 47% lower interval colorectal cancer rate. | 中 | SP013 |
| CP022 | A US multicenter pilot study found significantly more non-neoplastic polyps with real-time CAD and no difference in 10 mm or larger polyps versus historical controls. | 中 | SP014 |
| CP023 | Standard colonoscopy without CADe remains a viable status-quo substitute because independent guidelines and literature do not treat universal AI adoption as settled. | 中 | SP012, SP014, SP015 |
| CP024 | Iterative's Series C materials describe a model that combines centralized operations, expert staffing, proprietary AI technology, and deep clinical-trial expertise. | 高 | SP026, SP027 |
| CP025 | Iterative says sponsors and CROs gain centralized access to industry-leading sites and diverse patient populations through its network model. | 中 | SP026 |
| CP026 | IQVIA says it has an 88% higher Phase II and 25% higher Phase III IBD trial success rate than industry rates. | 中 | SP016 |
| CP027 | IQVIA says more than 25% of outsourced industry-sponsored phase II and III IBD trials launched since 2015 were conducted by IQVIA. | 中 | SP016 |
| CP028 | IQVIA says its strategic site networks include more than 1,100 high-performing sites across 100 countries. | 高 | SP016, SP017 |
| CP029 | ICON says it conducted more than 114 gastrointestinal studies involving 28,400-plus patients and 10,000-plus sites in the last five years. | 中 | SP018 |
| CP030 | ICON says its Site & Patient Solutions aim to increase predictability in enrollment and retention through upfront site and participant management. | 中 | SP019 |
| CP031 | ICON says its Accellacare network provides access to 8.1 million-plus patients at more than 50 sites across five countries. | 高 | SP020, SP019 |
| CP032 | ICON says Accellacare sites are 58% faster to the site-initiation visit than non-Accellacare sites. | 中 | SP020 |
| CP033 | Alimentiv markets itself as a specialized GI CRO and says it supports more than 70% of IBD compounds in development. | 高 | SP021, SP022 |
| CP034 | Alimentiv says it has access to 5,000-plus sites in 60-plus countries and more than 30 years of GI research experience. | 高 | SP021, SP022, SP023 |
| CP035 | Alimentiv says its site network is built around GI-specific site centricity and research teams averaging 10 years of GI experience. | 中 | SP023 |
| CP036 | GI Alliance says its partnership with Iterative added 21 active research sites running more than 80 Phase II-IV GI and hepatology trials and took the combined network above 70 locations in the US and Europe. | 中 | SP024 |
| CP037 | One GI says its 34 clinics across six states and 13 active research sites are now an exclusive mainstay in Iterative's global site network. | 中 | SP025 |
| CP038 | Pulse 2.0 reports that Iterative's global network now includes more than 100 research sites and more than 40 pharmaceutical, biotech, medical-device, and CRO partners. | 中 | SP027 |
| CP039 | Yahoo Finance's paid DDW 2026 release says Iterative's enrollment rates remained more than 3 times industry benchmarks at roughly double the prior scale. | 中 | SP028 |
| CP040 | Docwire's interview summary says the cited IBD study covered 27 Iterative network sites and reported faster activation and randomization than published benchmarks. | 中 | SP029 |
| CP041 | The reviewed public materials show a split competitive landscape in which OEM AI vendors dominate procedure-room distribution while CRO and site-network rivals dominate sponsor relationships and site operations. | 中 | SP003, SP004, SP010, SP016, SP018, SP021 |
| CP042 | Reviewed official pages for Iterative, Medtronic, Olympus, Fujifilm, IQVIA, ICON, and Alimentiv emphasize demos or contact-led sales rather than public list prices. | 中 | SP001, SP002, SP003, SP004, SP005, SP010, SP016, SP018, SP021 |
| CP043 | Olympus and Fujifilm present stronger channel lock-in risk than Iterative because their AI offerings are embedded in branded endoscopy ecosystems and compatible hardware stacks. | 中 | SP005, SP006, SP008, SP010 |
| CP044 | Medtronic presents a more modular AI threat than Olympus or Fujifilm because GI Genius is sold as a stand-alone intelligent endoscopy module. | 中 | SP003 |
| CP045 | IQVIA and ICON can out-scale Iterative in sponsor coverage because both market global site footprints and multispecialty patient-access infrastructure beyond GI. | 中 | SP016, SP017, SP018, SP020 |
| CP046 | Alimentiv can pressure Iterative from the specialty side because it markets GI-only depth, centralized endoscopy heritage, and 5,000-plus site relationships. | 中 | SP021, SP022, SP023 |
| CP047 | Iterative's clearest public differentiation is that it pairs AI tooling with embedded community-site operations while most reviewed rivals specialize in only one side of that stack. | 中 | SP001, SP026, SP016, SP018, SP021 |
| CP048 | A buyer can reproduce much of Iterative's value proposition by pairing a CADe vendor with a CRO or site-network operator instead of buying one integrated platform. | 中 | SP003, SP005, SP010, SP016, SP020, SP021 |
| CP049 | Provider-controlled research ancillaries are a structural competitive risk because large GI groups can keep research access exclusive or redirect it to alternative operators. | 中 | SP024, SP025 |
| CP050 | Public sources do not disclose Iterative's installed SKOUT base, sponsor win rates versus named CRO rivals, or direct cross-vendor price comparisons across either arena. | 中 | SP001, SP016, SP018, SP021, SP026 |
| CI001 | Iterative publicly describes itself in 2025-2026 as a healthcare technology and services company powering the acceleration of clinical research, not merely as a standalone AI device vendor. | 高 | SI003, SI004, SI008 |
| CI002 | The company’s current model combines centralized operations, expert staffing, proprietary AI technology, and clinical trial expertise to support trial execution. | 高 | SI003, SI010 |
| CI003 | Iterative markets centralized access to its site network and patient populations for sponsors and CROs, indicating a sponsor-facing services revenue stream. | 高 | SI003, SI008 |
| CI004 | Official 2026 materials say Iterative’s network includes more than 100 research sites and over 40 pharmaceutical, biotech, medical-device, and CRO partners. | 高 | SI003, SI008 |
| CI005 | The GI Alliance partnership added 21 active sites and 80-plus trials, and states that Iterative will lead trial management while supplying operational infrastructure and trial pipeline. | 中 | SI007 |
| CI006 | The One GI partnership put 13 active research sites into Iterative’s network and describes a proven site-serving operational model rather than a pure software license. | 中 | SI006 |
| CI007 | The US Heart & Vascular partnership says participating sites gain end-to-end support across clinical operations, services, business development, technology innovation, and corporate functions. | 中 | SI008 |
| CI008 | Iterative’s acquisition of three NextStage cardiology sites adds active cardiovascular trial operations and expands the company’s multispecialty site-footprint directly on balance sheet. | 中 | SI009 |
| CI009 | The 2022 acquisition of CRSG/Precision deepened Iterative’s end-to-end clinical-research capabilities and specifically tied AI Recruitment services to provider and pharmaceutical stakeholders. | 中 | SI005 |
| CI010 | Taken together, the reviewed public evidence indicates Iterative’s primary commercial engine is sponsor/CRO and provider-network services, with AI embedded inside those workflows rather than sold mainly as standalone SaaS. | 中 | SI003, SI005, SI006, SI007, SI008 |
| CI011 | SKOUT is marketed as a real-time AI polyp-detection product for colonoscopy, giving Iterative at least one concrete device/product revenue surface alongside services. | 中 | SI001 |
| CI012 | The SKOUT product page emphasizes efficacy and workflow claims but does not publish a list price, reimbursement schedule, or obvious per-procedure fee. | 中 | SI001 |
| CI013 | Iterative’s public resources flow is dominated by partnership announcements, site-network performance data, and clinical-research content rather than pricing or customer-tier packaging. | 中 | SI002, SI003, SI010 |
| CI014 | No reviewed official or secondary source disclosed public list pricing for SKOUT, sponsor services, provider-network partnerships, or AI-enabled site tooling. | 高 | SI001, SI002, SI003, SI006, SI007, SI008 |
| CI015 | Medicare Part B covers screening colonoscopies, but the public Medicare coverage page does not identify a separate payment pathway for Iterative’s colonoscopy AI. | 中 | SI017 |
| CI016 | AGA warned that CMS’s proposed 2026 rules would cut GI ASC and hospital-based endoscopy and E/M payments by $58 million, including an average 8% cut in facility-based physician endoscopy payments. | 中 | SI015 |
| CI017 | Becker’s reported that 2025 Medicare pays about $220 per colonoscopy, GI procedure reimbursement has fallen 33% inflation-adjusted since 2007-2022, and ASC reimbursement still lags inflation. | 中 | SI016 |
| CI018 | Boston Scientific’s 2026 reimbursement guide says physician and facility payments vary materially by site of service and notes the ASC conversion factor is approximately 60% of OPPS, while private-payer reimbursement depends on negotiated contracts. | 中 | SI018 |
| CI019 | The MACRA/MIPS colonoscopy cost-measure literature shows GI providers are held responsible for some costs occurring within 14 days after colonoscopy and that site of service, anesthesia, pathology, and complications drive cost variation. | 中 | SI019 |
| CI020 | A peer-reviewed reimbursement review says CMS reimburses at least eight AI devices and warns that per-use reimbursement may encourage overuse of AI if payment design is not carefully controlled. | 中 | SI020 |
| CI021 | Sidley wrote that the proposed Health Tech Investment Act would create a transitional Medicare payment pathway for algorithm-based healthcare services for at least five years, but the bill had not advanced as of that update. | 中 | SI021 |
| CI022 | Akin Gump says CMS payment policy for AI-enabled services is evolving and that there is no standard method for covering and paying every FDA-approved AI-enabled device across settings. | 中 | SI022 |
| CI023 | Harvard’s policy summary argues that bundling, value-based pricing, and regular price adjustments are plausible payment structures because fee-for-service fits AI-enabled clinical services poorly. | 中 | SI023 |
| CI024 | Given the absence of a durable standalone payment path in the reviewed sources, SKOUT monetization likely depends on bundled device/procedure budgets or negotiated enterprise purchasing rather than a mature Medicare add-on code. | 中 | SI017, SI020, SI021, SI022, SI023 |
| CI025 | The official Series C raised $77 million and management said the proceeds would support cardiology, obesity, geographic expansion, and broader provider partnerships. | 中 | SI003 |
| CI026 | Reviewed official sources do not disclose revenue, ARR, gross margin, burn, cash on hand, runway, or debt obligations for Iterative Health. | 高 | SI001, SI002, SI003, SI004, SI006, SI008, SI009, SI010 |
| CI027 | Official company materials described Iterative as having 250-plus employees worldwide in October 2025, and a May 2026 acquisition release described a 250-person team. | 高 | SI004, SI009 |
| CI028 | Tracxn estimated Iterative Health had 166 employees as of April 2026, which conflicts with the company’s 250-plus employee claim. | 低 | SI014 |
| CI029 | Forge lists Iterative’s April 2026 Series C post-money valuation at $1.3 billion and total funding at $272.32 million. | 中 | SI011 |
| CI030 | TechCrunch described Iterative as a $1.4 billion unicorn with more than $270 million raised and a $75 million Series C according to PitchBook. | 中 | SI012 |
| CI031 | AI2.work described Iterative’s total funding as more than $273 million and its valuation as approximately $1.3 billion after the Series C. | 低 | SI013 |
| CI032 | Latest valuation and total-funding figures for Iterative should therefore be treated as a secondary-source range, not as a company-disclosed fact. | 中 | SI011, SI012, SI013 |
| CI033 | IQVIA’s March 2025 SEC filing reported $3.829 billion of quarterly revenue, $496 million of operating income, $1.740 billion of cash, and $1.940 billion of unearned income. | 中 | SI024 |
| CI034 | IQVIA’s March 2025 operating margin was approximately 13.0%, calculated as $496 million of operating income divided by $3.829 billion of revenue. | 中 | SI024 |
| CI035 | ICON’s FY2025 results showed $8.251 billion of revenue, $1.531 billion of adjusted EBITDA (18.6% margin), $21.8 billion of backlog, $647 million of cash, and $2.8 billion of net debt. | 中 | SI027 |
| CI036 | ICON’s 2025 20-F and Q4 2025 release disclosed revenue-recognition errors, prior-period revenue overstatements, and material weaknesses tied to clinical trial services accounting. | 高 | SI026, SI027 |
| CI037 | Medpace’s FY2025 Q4 release showed $2.530 billion of revenue, $557.7 million of EBITDA (22.0% margin), $3.027 billion of backlog, and $497 million of cash. | 中 | SI031 |
| CI038 | Medpace’s 2025 annual report and Q4 2025 release describe a 6,200-person, 46-country CRO with explicit liquidity, backlog, and margin disclosure. | 高 | SI030, SI031 |
| CI039 | Across IQVIA, ICON, and Medpace, mature CRO margin context is roughly 13.0% to 22.0% using reported operating or EBITDA margins, but those companies also publish backlog, cash, and accounting detail that Iterative does not. | 高 | SI024, SI027, SI031 |
| CI040 | Iterative’s reviewed company materials repeatedly emphasize site success, centralized operations, staffing, trial management, and clinical research operations, indicating a labor- and service-heavy model rather than a purely software gross-margin structure. | 高 | SI003, SI007, SI008, SI009 |
| CI041 | The NextStage acquisition release says a 250-person team spans site staff plus clinical, regulatory, financial, and operational functions, implying a meaningful fixed-cost base outside pure R&D. | 中 | SI009 |
| CI042 | DDW 2026 network data showed 0.33 randomized patients per site per month versus a 0.10 benchmark and more than 3x benchmark enrollment performance at roughly double prior scale. | 中 | SI010 |
| CI043 | GI Alliance plus the Series C release show that Iterative’s network scaled from more than 70 locations in April 2025 to more than 100 research sites in 2026, confirming fast operating expansion. | 高 | SI007, SI003 |
| CI044 | Using the public headcount range of 166 to 250 employees against a network of over 100 sites implies a rough 1.7 to 2.5 employees per site before corporate overhead, suggesting the model depends on partner labor and centralized leverage rather than dense on-balance-sheet staffing at each site. | 低 | SI003, SI004, SI009, SI014 |
| CI045 | No public sources disclose CAC, LTV, payback, sponsor retention, revenue per site, or margin per site, so activation and enrollment data are only operating-efficiency proxies, not proof of attractive unit economics. | 高 | SI002, SI003, SI010 |
| CI046 | Because Iterative is expanding into cardiology and obesity while also acquiring sites, capital needs likely rise before public revenue or margin visibility improves. | 中 | SI003, SI008, SI009 |
| CI047 | Public diligence risk is not lack of financing access but lack of public P&L, cash-flow, pricing, and contract-economics disclosure. | 中 | SI003, SI011, SI012, SI013, SI026 |
| CI048 | Revenue quality should be viewed as mixed: provider-network and workflow relationships can be sticky, but the observable model appears materially services- and trial-budget-driven rather than clean recurring software ARR. | 中 | SI003, SI005, SI006, SI007, SI008, SI010 |
| CI049 | Iterative’s GTM appears enterprise and B2B-direct, with negotiated relationships to sponsors, CROs, and provider networks rather than self-serve adoption. | 高 | SI003, SI006, SI007, SI008 |
| CI050 | Mature CRO filings show meaningful working-capital and revenue-recognition complexity, which makes Iterative’s non-disclosure on backlog, unearned revenue, and cash conversion especially material for diligence. | 中 | SI024, SI026, SI027, SI030, SI031 |
| CI051 | Capital adequacy cannot be confidently underwritten from public data because current cash, burn, debt, and runway are undisclosed; the only durable public capital facts are the Series C amount and a secondary-source funding and valuation range. | 高 | SI003, SI011, SI012, SI013, SI026 |
| CI052 | The financial verdict is that Iterative has credible monetization surfaces and financing access, but public evidence is still insufficient to prove revenue quality, margin trajectory, or self-funding capacity. | 中 | SI003, SI010, SI011, SI024, SI027, SI031 |
| CE001 | Iterative Health describes itself as a healthcare technology and services company that combines clinical-trials expertise with cutting-edge AI to accelerate research. | 高 | SE002, SE024 |
| CE002 | The company’s homepage presents a site network embedded into community care across GI, hepatology, obesity, and cardiology rather than a single-product workflow. | 中 | SE001 |
| CE003 | Iterative publicly says it helps investigators and sponsors run trials more efficiently by accelerating enrollment and reducing operational burden. | 高 | SE001, SE002 |
| CE004 | The public product stack spans SKOUT, a site-centric research network, operational startup support, and AI-enabled disease-assessment workflows for IBD studies. | 中 | SE001, SE004, SE008, SE009, SE010, SE024 |
| CE005 | A June 2026 startup.jobs profile says Iterative serves clinical research sites across the United States and Europe with AI and machine-learning tools plus advanced disease-assessment tools. | 低 | SE024 |
| CE006 | SKOUT has FDA 510(k) clearance as a Class II gastrointestinal lesion software detection system for adult colorectal-cancer screening or surveillance colonoscopy. | 高 | SE012, SE006 |
| CE007 | SKOUT detects potential colorectal polyps in real time and marks them with a blue rectangular outline on the endoscopic video feed. | 高 | SE012, SE023 |
| CE008 | The GUDID description says SKOUT integrates directly into colonoscopy workflow and lets users switch between bypass endoscopic feed and SKOUT feed with a mode-selection button. | 中 | SE023 |
| CE009 | The FDA clearance documents compatibility with Olympus EVIS EXERA II, Olympus EVIS EXERA III, and Fujifilm Eluxeo VP-7000 HD-output environments via SDI input. | 中 | SE012 |
| CE010 | SKOUT pauses its detection marker when endoscopic tools are present or lighting conditions are inadequate so the overlay does not interfere with intervention. | 中 | SE012 |
| CE011 | AccessGUDID lists SKOUT model SKOUT210 as in commercial distribution, indicating the device progressed beyond clearance into active commercial status. | 中 | SE023 |
| CE012 | Official SKOUT materials say the device does not increase procedure time during colonoscopy. | 高 | SE004, SE022, SE016 |
| CE013 | Iterative’s product page says high-performing endoscopists resected one additional adenoma for every 4.5 patients examined with SKOUT without increasing hyperplastic polyp resection. | 中 | SE004, SE013 |
| CE014 | The Provation value brief says SKOUT drove a 27% relative increase in adenomas per colonoscopy overall, a 29% increase in the proximal colon, and a 44% relative increase in 5-9 mm polyp detection while detecting sessile serrated polyps. | 中 | SE022 |
| CE015 | An early SKOUT study reported overall adenoma detection rates of 54.2% with SKOUT versus 40.6% in the comparator cohort. | 中 | SE013 |
| CE016 | The same study reported screening-exam adenoma detection rates of 53.6% with SKOUT versus 30.8% without it. | 中 | SE013 |
| CE017 | Provation’s current SKOUT brief cites a randomized U.S. dataset of 1,359 patients across five academic and community centers and 22 high-performing endoscopists. | 中 | SE022 |
| CE018 | Iterative’s site-network materials describe a coordinated, operationally supported network intended to accelerate enrollment while maintaining data quality. | 高 | SE008, SE019 |
| CE019 | The published site-network summary reports a median 74 days from site selection to activation versus 122-171 day industry benchmarks. | 高 | SE008, SE019 |
| CE020 | The same site-network summary reports a median 45 days from activation to first patient screened. | 中 | SE008 |
| CE021 | The site-network summary reports a median 83 days to first patient randomized versus a 140-day benchmark. | 高 | SE008, SE019 |
| CE022 | Iterative says its network achieved a weighted average randomization rate of 0.34 patients per site per month, more than three times 2020 benchmarks. | 高 | SE008, SE007 |
| CE023 | The DDW 2026 update says the more-than-3x randomization performance persisted at approximately double the underlying trial-and-site scale. | 中 | SE007 |
| CE024 | Iterative frames the site-network problem as one of documentation, training, budgeting, and prescreening burden that operational support can remove from local sites. | 中 | SE008, SE019 |
| CE025 | Iterative’s ulcerative-colitis publication summary says AI-enabled endoscopy and histology can reveal subtle inflammation patterns in mild-to-moderate UC that conventional correlations may miss. | 中 | SE009 |
| CE026 | Iterative’s ACG 2024 summary says traditional endoscopic assessment in IBD is time-consuming, prone to inter-observer variability, and often not reflective of real-world patient populations. | 中 | SE011 |
| CE027 | The BMJ central-reading summary says experienced readers disagree roughly 40% of the time when scoring the same endoscopic video with Mayo Endoscopic Score. | 中 | SE010 |
| CE028 | Iterative’s hybrid central-reading framework uses two independently developed machine-learning models plus a board-certified gastroenterologist who adjudicates only when the models disagree. | 中 | SE010 |
| CE029 | The BMJ summary reports statistical non-inferiority to traditional central reading with kappa 0.78 and an 81% reduction in required human reads per video. | 中 | SE010 |
| CE030 | The same BMJ summary says 16% of human-only cases produced different final scores depending on reader assignment. | 中 | SE010 |
| CE031 | Across the BMJ summary and SKOUT value brief, Iterative frames AI as augmentation with human oversight rather than physician replacement. | 中 | SE010, SE022, SE012 |
| CE032 | The combination of a 510(k) record and commercial-distribution listing gives SKOUT more public regulatory trust anchors than the broader trial-tech modules possess. | 高 | SE012, SE023 |
| CE033 | Yale’s guideline summary says AGA draft guidance conditionally recommends CADe in adult colonoscopy and sees little to no adverse events or clinically meaningful time increase. | 高 | SE016, SE015 |
| CE034 | The broader CADe literature still highlights false positives, automation bias, operator deskilling, system integration, generalizability, and long-term outcome uncertainty as active limitations. | 高 | SE015, SE014 |
| CE035 | A DDW 2026 real-world analysis across more than 1.5 million matched patients found AI-assisted colonoscopy associated with a 47% reduction in interval colorectal cancer. | 中 | SE018 |
| CE036 | Because SKOUT only assists trained gastroenterologists and pauses during tool use, public safety design assumes human confirmation rather than autonomous device action. | 高 | SE012, SE022 |
| CE037 | Medtronic publicly advertises GI Genius with 99.7% sensitivity and less than 1% false positives for colorectal-polyp detection. | 中 | SE021 |
| CE038 | Olympus positions CADDIE as a cloud-based CADe solution that aids detection of large, flat, and sessile serrated lesions without disrupting workflow. | 中 | SE020 |
| CE039 | Fujifilm commercialized CAD EYE in 2024 as a real-time detection system for colonic mucosal lesions during colonoscopy. | 中 | SE025 |
| CE040 | Iterative’s most visible differentiation is category breadth: it combines a regulated in-procedure device with site-network operations and IBD assessment workflows rather than selling only a standalone CADe box. | 中 | SE001, SE004, SE008, SE009, SE010, SE024 |
| CE041 | SKOUT commercialization depends materially on the Provation relationship, which Iterative describes as an exclusive distribution arrangement into existing GI workflow environments. | 高 | SE006, SE017 |
| CE042 | The documented processor list means SKOUT deployment is constrained by hardware compatibility with Olympus and Fujifilm environments. | 中 | SE012 |
| CE043 | The site-network model depends on continued participation by community research sites plus ongoing centralized operational support to keep activation and enrollment metrics high. | 中 | SE001, SE008, SE019 |
| CE044 | The hybrid central-reading model depends on access to curated video data, independently developed models, and board-certified gastroenterologist adjudicators. | 中 | SE010 |
| CE045 | Public roadmap visibility is milestone-based: the clearest evidence is FDA clearance, publication cadence, and performance updates rather than a detailed forward feature roadmap. | 中 | SE005, SE006, SE007, SE008, SE009, SE010, SE023 |
| CE046 | Iterative’s resources hub currently foregrounds site-network, SKOUT, and GI-clinical-research outputs rather than a long list of independent software SKUs. | 中 | SE005 |
| CE047 | A June 2026 startup.jobs page listing 52 open roles suggests the operating model still depends on active organizational build-out rather than a fully self-serve product footprint. | 低 | SE024 |
| CE048 | Public sources still do not disclose SKOUT pricing, installed base, exhaustive interoperability, or a forward release calendar for the broader IBD toolchain. | 中 | SE004, SE005, SE023 |
| CE049 | Iterative markets SKOUT as improving clinician diagnostics while the broader company stack is framed as speeding therapeutic development for sponsors and sites. | 中 | SE001, SE004 |
| CE050 | Because competitors also market real-time detection, sessile-lesion support, and workflow preservation, Iterative likely has to win on workflow adjacency, distribution, and operating-model breadth rather than on CADe uniqueness alone. | 中 | SE020, SE021, SE025, SE004, SE008, SE010 |
| CU001 | Iterative Health’s public customer stack spans provider networks and research sites, sponsors / CROs / biotech teams, and a third buyer group around SKOUT device deployment. | 高 | SU001, SU002, SU003, SU023 |
| CU002 | The provider pitch is about turning clinical research into a scalable ancillary service, while the sponsor pitch is about enrollment forecasts, activation speed, and operational execution. | 高 | SU002, SU003 |
| CU003 | Series C materials say Iterative’s global network includes 100+ research sites across North America, Europe, India, and Australia plus 40+ pharmaceutical, biotech, medical-device, and CRO partners. | 高 | SU005, SU019 |
| CU004 | Iterative’s provider page cites 2x faster site activation and a 3.4x increase in patient enrollments for partner sites. | 中 | SU002 |
| CU005 | The same provider page says AI-powered prescreening lifts patient randomizations by 40% and relieves thousands of hours of chart-review work. | 中 | SU002 |
| CU006 | GI Alliance publicly said 21 active research sites and 80+ active Phase II-IV trials joined Iterative’s network, extending it to 70+ locations in the U.S. and Europe in April 2025. | 高 | SU012, SU021 |
| CU007 | One GI publicly said its 34-clinic network and 13 active research sites became an exclusive mainstay in Iterative’s global site network. | 高 | SU010, SU013 |
| CU008 | US Heart & Vascular’s customer-side announcement says its sites joined a network that now spans 100+ sites across three continents, providing Iterative’s clearest public cardiology customer proof. | 高 | SU011, SU014 |
| CU009 | Gastro Health described deploying Iterative’s AIR technology into a 380-physician, 150-location, seven-state platform and framed the relationship as a starting point for broader expansion. | 中 | SU015 |
| CU010 | The Gastro One case study reports 2x faster activation, two patient randomizations in about five months, more than 50% of enrollments sourced through Iterative, and a four-month acceleration toward a multi-million-dollar opportunity. | 中 | SU008 |
| CU011 | Iterative’s Takeda article says the collaboration draws on de-identified endoscopy videos and EHR data from over 100 GI community sites across seven states to improve IBD trial assessment and patient selection. | 中 | SU009 |
| CU012 | Iterative explicitly markets enrollment forecasts and partnership support to sponsors, CROs, and biotech companies through its sponsor inquiry page. | 中 | SU003 |
| CU013 | Iterative’s resources hub shows a steady 2025-2026 cadence of customer-proof content around GI Alliance, Gastro One, One GI, Takeda, and DDW 2026 rather than a one-time burst of announcements. | 中 | SU004 |
| CU014 | Iterative’s JCC summary says median site-selection-to-activation time was 74 days, activation-to-first-screened was 45 days, activation-to-first-randomized was 83 days, and weighted average randomization was 0.34 patients per site per month. | 中 | SU007 |
| CU015 | Docwire’s transcript says the study covered 27 Iterative network sites (21 U.S., 6 Europe), that the company then had more than 80 sites overall, and that the observed randomization rate was 0.34 patients per site per month versus about 0.1 in published benchmarks. | 中 | SU020 |
| CU016 | DDW 2026 materials report 0.33 patients randomized per site per month versus a 0.10 benchmark and say four completed trials still delivered a 0.32 median randomization rate. | 高 | SU006, SU016, SU018 |
| CU017 | The DDW 2026 release says the new analysis reflects an approximately two-fold increase in network scale across trials and sites versus the prior ECCO-era public dataset. | 高 | SU006, SU017, SU018 |
| CU018 | Taken together, the 2025-2026 announcements show a shift from GI-focused provider partnerships toward a broader multispecialty network spanning cardiology and obesity as well as GI and hepatology. | 中 | SU004, SU005, SU011 |
| CU019 | Provider-side named customer proof is strongest for GI Alliance, One GI, Gastro One, USHV, and Gastro Health because those relationships include site counts, executive quotes, or quantified workflow outcomes. | 高 | SU008, SU010, SU011, SU012, SU013, SU014, SU015 |
| CU020 | Sponsor-side named proof is much thinner: Takeda is the only clearly named sponsor found publicly, while the other 40+ sponsor / CRO / device partners remain unnamed. | 高 | SU005, SU009, SU002 |
| CU021 | The SKOUT product page contains anonymous clinician testimonials and efficacy claims but no named provider account, installed-base number, or deployment list. | 中 | SU023 |
| CU022 | Iterative’s public customer proof skews toward partner announcements and company-authored case studies rather than independent satisfaction surveys or broad customer-review surfaces. | 中 | SU004, SU008, SU010, SU011, SU012, SU015 |
| CU023 | Morningstar and BusinessWire syndications quote GI Alliance CMO Casey Chapman saying the DDW 2026 data show what became possible one year into the strategic partnership. | 高 | SU017, SU018 |
| CU024 | Ziad Younes says Iterative has been a trusted partner to his site for several years, giving the chapter its clearest public continuity signal for a named provider relationship. | 高 | SU008, SU010, SU013 |
| CU025 | One GI’s description of its 13 active sites as an exclusive mainstay in Iterative’s network implies deeper operating integration than a casual referral or one-study collaboration. | 高 | SU010, SU013 |
| CU026 | Gastro Health is best read as a technology-deployment proof point with future expansion intent, not yet as a fully quantified network-performance customer story. | 中 | SU015 |
| CU027 | No public NRR, GRR, logo churn, NPS, review-platform score, or average contract length was found across Iterative’s homepage, provider/sponsor pages, resources hub, Gastro Health release, or SKOUT page. | 高 | SU001, SU002, SU003, SU004, SU015, SU023 |
| CU028 | Repeat-usage proxies are operational rather than SaaS-like: several-year site relationship language, a one-year GI Alliance follow-on quote, and a second major public performance readout at larger scale. | 高 | SU008, SU017, SU018 |
| CU029 | Those proxies are encouraging but still do not show sponsor re-award rates, provider renewals, or revenue retention by cohort. | 高 | SU017, SU018, SU020 |
| CU030 | Docwire’s interview says the model took several years of work to build and can now be used across sites globally, supporting the idea that Iterative has a repeatable operating backbone rather than a one-off pilot process. | 中 | SU020 |
| CU031 | Public durability proof focuses on activation, screening, randomization, and within-trial patient retention rather than portfolio-level satisfaction or long-term contract outcomes. | 高 | SU007, SU008, SU020 |
| CU032 | Because clinical-research customers often multi-home across CROs and site networks, strong single-study execution is only a proxy for retention and not proof of sponsor exclusivity. | 中 | SU003, SU020, SU024 |
| CU033 | The direct OUP ECCO/JCC abstract page was blocked in this run, so public verification of the underlying abstract remains indirect through Iterative summaries, BusinessWire syndication, and Docwire’s transcript. | 中 | SU006, SU018, SU020, SU025 |
| CU034 | Customer concentration risk looks high on the provider side because GI Alliance, One GI, USHV, Gastro Health, and Gastro One account for most of the named public proof while revenue contribution by network is undisclosed. | 中 | SU005, SU010, SU011, SU012, SU015 |
| CU035 | Sponsor concentration is difficult to size publicly: Iterative discloses 40+ sponsor / CRO / device partners but not top-customer revenue share, study count per sponsor, or sponsor renewal cohorts. | 高 | SU002, SU005, SU019 |
| CU036 | Procurement friction remains real because Iterative’s value proposition depends on staffing, regulatory help, budgeting, chart reviews, and operational coordination that some large sites or sponsors may already perform internally. | 中 | SU007, SU008, SU020 |
| CU037 | Medtronic’s GI Genius page advertises 99.7% sensitivity and less than 1% false positives, underscoring that SKOUT buyers face credible CADe alternatives. | 中 | SU024 |
| CU038 | Because SKOUT’s public customer proof is unnamed while category competitors are explicit, Iterative’s moat likely rests more on sponsor/site network relationships and operating-model breadth than on uncontested device demand. | 中 | SU023, SU024, SU005 |
| CU039 | The clearest expansion path beyond original GI roots is to use the same site-serving model across cardiology and obesity while broadening sponsor wallet share in adjacent therapeutic areas. | 中 | SU001, SU005, SU011 |
| CU040 | Despite global network claims that include India and Australia, named customer proof remains heavily U.S.-centric, leaving international execution quality and concentration less proven than the headline scale numbers suggest. | 中 | SU005, SU012, SU014 |
| CU041 | Iterative does not publicly disclose average contract length, renewal cohorts, provider revenue-share economics, or sponsor re-award rates, leaving both durability and concentration impossible to quantify precisely. | 高 | SU001, SU002, SU003, SU015 |
| CR001 | Iterative says its global network now includes more than 100 research sites across North America, Europe, India, and Australia plus 40+ pharmaceutical, biotech, device, and CRO partners. | 高 | SR007, SR023 |
| CR002 | Public company and syndication materials say the network delivers 2x faster activation, 3.4x higher enrollment, 40% prescreening uplift, and roughly 0.33-0.34 patients randomized per site per month at scale. | 高 | SR003, SR008, SR022, SR029 |
| CR003 | GI Alliance added 21 active research sites and 80+ Phase II-IV trials, extending Iterative’s network to 70+ US and Europe locations in 2025. | 高 | SR024, SR031 |
| CR004 | One GI contributes 34 clinics and 13 active research sites, and public materials describe Iterative as a trusted partner for several years. | 高 | SR025, SR003 |
| CR005 | US Heart & Vascular extends the model into cardiology and describes a network spanning more than 100 sites across three continents. | 高 | SR026, SR007 |
| CR006 | The provider and sponsor materials show that Iterative’s value proposition depends on supplying operational backbone functions—budgets, legal/regulatory support, sponsor engagement, study startup, and staff management—rather than just software seats. | 高 | SR003, SR004 |
| CR007 | Public leadership pages show a broader bench than a pure founder-led startup, including a COO, President/CFO, partnerships, finance, product, medical, people, and legal leaders. | 高 | SR002, SR009 |
| CR008 | Bill Kayser joined in 2025 from GI Alliance, strengthening finance and partner-network experience but also underscoring that executive bench depth has been expanded recently rather than long-established. | 高 | SR009, SR030 |
| CR009 | The public SKOUT page provides efficacy claims and anonymous testimonials, but no named provider deployments or installed-base disclosures were found in the reviewed materials. | 高 | SR005, SR006 |
| CR010 | Iterative’s privacy policy discloses PI collection, service-provider sharing, legal rights handling, and only general “industry standard” security language; reviewed public materials did not surface audit-grade security disclosures. | 中 | SR010, SR006 |
| CR011 | Recent CADe literature consistently says AI can improve adenoma detection, but unresolved issues remain around false positives, automation bias, deskilling, integration, generalizability, and long-term outcomes. | 高 | SR011, SR014 |
| CR012 | The CADe category is crowded enough that SKOUT is not competing in an empty space; recent literature and competitor pages place it alongside Medtronic GI Genius and Olympus CADDIE. | 高 | SR011, SR027, SR028 |
| CR013 | The BMJ review reports false positives of 1.44%-3.40%, false negatives up to 1.03%, and cites non-randomized real-world meta-analysis showing no significant ADR improvement with CADe. | 中 | SR014 |
| CR014 | An ACG evidence summary reported non-AI ADR falling from 28.4% to 22.4% after centers introduced AI exposure, supporting the possibility of operator deskilling or over-reliance if governance is weak. | 高 | SR015, SR014 |
| CR015 | A 2026 real-world news summary linked AI-era colonoscopy to higher adenoma detection and a 47% lower interval CRC rate, which is an important mitigation against the most bearish efficacy view, but it remains observational rather than causal proof. | 中 | SR016, SR014 |
| CR016 | Reviewed legal and policy sources say there is still no consistent or predictable Medicare payment pathway for AI-enabled devices, and the Health Tech Investment Act had not progressed. | 高 | SR017, SR013 |
| CR017 | Multiple reimbursement analyses warn that fee-for-service and per-use payment structures fit AI poorly and can either overpay or underfit rapidly scaling software-driven devices. | 高 | SR013, SR018 |
| CR018 | Medicare Part B covers screening colonoscopy, but the reviewed materials do not show a settled separate AI payment benefit for endoscopy CADe systems such as SKOUT. | 高 | SR021, SR017 |
| CR019 | AGA and Becker’s describe worsening reimbursement pressure for GI providers, including average 8% cuts for ASC/HOPD endoscopy in a 2026 proposal and multiyear inflation-adjusted pay erosion. | 高 | SR019, SR020 |
| CR020 | Because Iterative sells into community GI and ASC-heavy workflows, procedure-level reimbursement compression can reduce customers’ willingness to pay for adjunct AI and research overhead. | 中 | SR019, SR020, SR021 |
| CR021 | Iterative’s strongest public proof clusters around GI Alliance, One GI, and US Heart & Vascular, making partner-network concentration a meaningful strategic dependency. | 高 | SR024, SR025, SR026 |
| CR022 | GI Alliance’s scale—more than 1,000 physicians and 400+ locations—shows why one partner can be strategically valuable, but also why account concentration could matter if economics are concentrated. | 中 | SR024 |
| CR023 | One GI’s 34 clinics and 13 active sites make it a meaningful contributor to network density, and the “exclusive mainstay” language implies deeper embedding than a loose pilot. | 中 | SR025 |
| CR024 | USHV expands the thesis into cardiology, increasing TAM while also increasing execution complexity beyond Iterative’s legacy GI and hepatology core. | 中 | SR026, SR007, SR022 |
| CR025 | Medtronic markets GI Genius on 99.7% sensitivity and less than 1% false positives, while Olympus markets CADDIE as a cloud-based add-on with automatic updates and platform integration. | 高 | SR027, SR028 |
| CR026 | The Medtronic and Olympus pages imply that incumbents can compete on workflow integration, training ecosystems, and equipment distribution—not just on core model accuracy. | 中 | SR027, SR028 |
| CR027 | Olympus explicitly ties CADDIE to the OLYSENSE hub and standard white-light imaging, illustrating that AI-endoscopy adoption depends on hardware and video-workflow compatibility. | 中 | SR027 |
| CR028 | Iterative’s provider model requires repeated PI/CRC training, hiring support, and ongoing staff management, which means labor availability and quality governance become harder as the network grows. | 高 | SR003, SR022 |
| CR029 | Docwire’s interview says the network took several years to build and still requires continuous improvement, reinforcing that multispecialty scaling is operationally heavy rather than frictionless. | 中 | SR022 |
| CR030 | DDW 2026 materials show network performance holding at roughly double prior scale, which is the company’s strongest mitigation against pure quality-drift skepticism. | 高 | SR008, SR029 |
| CR031 | Iterative clearly markets to sponsors, CROs, and biotechs, but public sponsor naming remains much thinner than provider-network proof, increasing uncertainty around sponsor concentration and renewal quality. | 中 | SR004, SR007 |
| CR032 | The company appears well capitalized after a $77M Series C and >$273M total funding, but the reported ~$1.3B valuation raises expectations relative to still-undisclosed revenue and margin data. | 中 | SR007, SR023 |
| CR033 | Clinical-trial delay economics are severe enough to support Iterative’s value proposition, but that same enterprise buying context implies long proof cycles and procurement scrutiny before revenue scales cleanly. | 中 | SR023, SR004, SR007 |
| CR034 | Reviewed public materials do not disclose revenue mix, gross margin, burn, renewal rates, or top-account concentration, so the financial model cannot be fully underwritten from public sources. | 高 | SR007, SR023, SR003 |
| CR035 | Iterative’s privacy policy, DPO appointment, and rights handling show baseline compliance infrastructure, but public security posture remains high-level rather than trust-center deep. | 中 | SR010 |
| CR036 | The BMJ review argues that responsibility for AI-enabled endoscopy is shared across clinicians, institutions, developers, and regulators, which increases accountability complexity if harm occurs. | 中 | SR014 |
| CR037 | The BMJ review says major 2025 bodies diverged on routine CADe use, with one weakly against, ESGE weakly in favor, and AGA making no recommendation due to low certainty and surveillance concerns. | 中 | SR014 |
| CR038 | SKOUT’s public page claims one extra adenoma for every 4.5 patients among high-performing endoscopists without more hyperplastic resections, but it still omits named customer, installed-base, and reimbursement disclosures. | 中 | SR005 |
| CR039 | Iterative’s premium-rate-card and direct-sponsor model depends on continuing to outperform benchmarks; if activation or randomization slips, premium pricing and partner stickiness could compress quickly. | 中 | SR003, SR007, SR022 |
| CR040 | The addition of a President/CFO and broader management team suggests improving depth, but the organization still looks mid-build rather than fully mature for a global multispecialty network. | 中 | SR002, SR009 |
| CR041 | The cleanest thesis-break events would be losing operational outperformance, suffering partner-concentration shocks, or failing to convert AI value into durable reimbursement and pricing under GI budget pressure. | 高 | SR008, SR019, SR020, SR024, SR025, SR026 |
| CR042 | Iterative’s public resources and marketing surfaces are rich in partnership and performance content but thin on trust, security, and reimbursement documentation, forcing investors to triangulate major risks from external sources. | 中 | SR006, SR010 |
| CR043 | Iterative maintains a public Greenhouse careers board with open roles across site operations, cardiology business development, talent acquisition, software, data, and international site management, underscoring ongoing hiring complexity. | 中 | SR032 |
| CR044 | Iterative’s ASC article explicitly frames limited staff bandwidth, regulatory complexity, recruitment difficulty, and milestone slippage as core trial-execution risks in ambulatory settings, while positioning end-to-end operational support as the mitigation. | 中 | SR033, SR003 |
| CR045 | Iterative U provides role-based, CE-eligible, end-to-end operational training and support, which is a real mitigation against quality drift but also evidence that the model requires continuous enablement investment. | 高 | SR034, SR003 |
| CR046 | The community-cardiology article argues that faster activation and broader patient access depend on centralized operational support, but also highlights upfront staffing, system, and turnover costs when community practices build research programs. | 中 | SR035, SR026 |
| CV001 | Iterative Health officially announced a $77 million Series C financing round in April 2026. | 高 | SV001, SV005 |
| CV002 | The Series C was led by Intrepid Growth Partners and GV, with Intrepid receiving a board seat and GV a board-observer role. | 中 | SV001 |
| CV003 | Iterative said its global research network exceeds 100 sites and more than 40 pharmaceutical, biotech, device, and CRO partners. | 高 | SV001, SV004 |
| CV004 | Iterative’s Series C release said the network delivered 2x faster activation and 3x higher patient enrollment than published IBD benchmarks. | 高 | SV001, SV004 |
| CV005 | Iterative’s DDW 2026 update said the network randomized 0.33 patients per site per month, more than 3x the industry benchmark, while operating at double the prior scale. | 中 | SV009 |
| CV006 | Iterative Scopes officially raised $150 million in Series B financing in January 2022. | 高 | SV007, SV002 |
| CV007 | Iterative Scopes officially raised $30 million in Series A financing in August 2021 after a previously disclosed $13.5 million seed round. | 高 | SV008, SV002 |
| CV008 | Using the company’s disclosed seed, Series A, Series B, and Series C amounts implies at least about $270.5 million of cumulative capital raised. | 中 | SV001, SV007, SV008 |
| CV009 | Forge placed Iterative Health’s Series C valuation at $1.3 billion in April 2026. | 中 | SV002 |
| CV010 | TechCrunch listed Iterative Health at a $1.4 billion valuation and more than $270 million of total funding on its 2026 unicorn tracker. | 中 | SV003 |
| CV011 | ai2.work reported that Iterative’s Series C brought total funding to more than $273 million and valued the company at roughly $1.3 billion. | 低 | SV004 |
| CV012 | Founder Lodge recorded the May 2026 round as a $77 million Series C financing. | 低 | SV005 |
| CV013 | Tracxn reported total funding of about $271 million and employee count of 166 as of late April 2026. | 低 | SV006 |
| CV014 | Taken together, the public valuation context clusters around roughly $1.3 billion to $1.4 billion, while the official company press release does not disclose a post-money valuation. | 中 | SV001, SV002, SV003, SV004 |
| CV015 | Forge shows partial Series C preference clues including first preference order, 1.0x non-participating liquidation preference, and an 8.0% dividend rate. | 低 | SV002 |
| CV016 | Forge’s historical marks imply Iterative stepped up from about $704 million at Series B and about $133 million at Series A to about $1.3 billion at Series C, or roughly a 1.8x-to-2.0x increase from B to C. | 中 | SV002 |
| CV017 | Neither Iterative’s Series C release nor its DDW 2026 performance release discloses revenue, ARR, gross margin, or retention metrics. | 中 | SV001, SV009 |
| CV018 | The public proof set is therefore stronger on operating throughput than on monetization quality or capital efficiency. | 中 | SV001, SV009, SV017 |
| CV019 | IQVIA’s June 2026 market cap of $30.14 billion and TTM revenue of $16.63 billion imply about a 1.8x revenue multiple. | 中 | SV011, SV012 |
| CV020 | ICON’s June 2026 market cap of $10.53 billion and TTM revenue of $8.10 billion imply about a 1.3x revenue multiple. | 中 | SV013, SV014 |
| CV021 | Medpace’s June 2026 market cap of $12.72 billion and TTM revenue of $2.67 billion imply about a 4.8x revenue multiple. | 中 | SV015, SV016 |
| CV022 | Medpace’s Q1 2026 revenue grew 26.5% year over year and its EBITDA margin reached 21.1%, illustrating the economics of a scaled, high-quality public CRO. | 中 | SV010 |
| CV023 | 2026 healthtech M&A benchmarks center around roughly 4x-6x revenue for general healthtech and 6x-8x-plus for premium AI and data assets. | 中 | SV018, SV019 |
| CV024 | Both Nelson Advisors and Healthcare Digital describe 2021-era 8x-10x revenue pricing as exceptional rather than normal in 2026. | 中 | SV018, SV019 |
| CV025 | Finro says Q1 2026 AI multiples no longer move in one clean direction and that narrative-first companies face sharper discounts. | 中 | SV017 |
| CV026 | Finro says contracted, repeatable revenue tied to specific workflows and efficient scaling increasingly determine who earns premium AI multiples. | 中 | SV017 |
| CV027 | PitchBook says nearly half of U.S. unicorns have not been priced in the past three years and many 2021-2022 unicorns no longer support billion-dollar values. | 中 | SV020 |
| CV028 | Prime Unicorn Index explicitly uses both primary financing rounds and secondary trading data to benchmark U.S. private unicorns. | 中 | SV021 |
| CV029 | VNTR says AI unicorn formation accelerated dramatically in Q1 2026 and openly asks whether the pace reflects late-cycle valuation inflation. | 中 | SV022 |
| CV030 | Sidley says Congress proposed a new Medicare payment classification for algorithm-based healthcare services because a stable reimbursement pathway for AI-enabled devices does not yet exist. | 中 | SV023 |
| CV031 | Health Affairs says CMS still lacks formal guidance for AI coverage and that per-use reimbursement models can mismatch scalable AI economics. | 中 | SV024 |
| CV032 | AGA says CMS’ proposed 2026 rules would cut physician payments for ASC and hospital-based GI endoscopy by an average of about 8% versus 2025 rates. | 中 | SV025 |
| CV033 | Becker’s describes continued GI reimbursement pressure from conversion-factor cuts, site-neutral pressure, and weak ASC updates. | 中 | SV026 |
| CV034 | Recent literature says CADe improves adenoma detection but still faces unresolved questions on generalizability, cost-effectiveness, and long-term outcomes. | 中 | SV027 |
| CV035 | Because reimbursement and clinical-economics clarity remain incomplete, SKOUT-like AI upside should not yet be valued as if it already had proven software-style monetization. | 中 | SV023, SV024, SV025, SV026, SV027 |
| CV036 | At a $1.3 billion-to-$1.4 billion valuation, Iterative would need roughly $270 million to $350 million of revenue to fit a 4x-5x CRO-like multiple band. | 中 | SV002, SV018, SV019 |
| CV037 | At the same valuation, Iterative would need roughly $165 million to $233 million of revenue to fit a 6x-8x premium AI/data multiple band. | 中 | SV002, SV018, SV019 |
| CV038 | If current revenue were still below $100 million, a $1.3 billion-to-$1.4 billion valuation would imply about 13x-14x-plus revenue. | 中 | SV002, SV017, SV018, SV019 |
| CV039 | The bull case requires Iterative to preserve 2x-3x operating outperformance while converting cardiology and obesity expansion into real revenue, not just narrative optionality. | 中 | SV001, SV004, SV009 |
| CV040 | The base case should treat Iterative as a high-quality CRO/data hybrid rather than as a pure software company until revenue mix and margins are disclosed. | 中 | SV010, SV018, SV019 |
| CV041 | The bear case centers on valuation compression, reimbursement drag, or network-performance slippage pushing Iterative toward lower CRO-like multiples. | 中 | SV017, SV020, SV023, SV025, SV026 |
| CV042 | Public evidence supports real company quality and momentum, but it does not yet establish a strong margin of safety at the current reported mark. | 中 | SV001, SV002, SV003, SV017, SV020 |
| CV043 | At the current reported price, the prudent public-evidence recommendation is research-more rather than buy. | 中 | SV001, SV002, SV017, SV020 |
| CV044 | The highest-priority diligence asks are current revenue, gross margin, retention, top-customer concentration, and full preference-stack terms. | 中 | SV001, SV002, SV017 |
| CV045 | The thesis breaks if disclosed revenue or margin are materially below what the current mark requires, if network performance reverts toward benchmark, or if reimbursement stays non-monetizable. | 中 | SV002, SV009, SV023, SV025, SV026 |
| CV046 | A supportable bull exit range is about $1.8 billion to $2.6 billion if Iterative reaches roughly $250 million to $320 million of revenue and still earns a 7x-8x premium multiple. | 低 | SV018, SV019, SV001, SV009 |
| CV047 | A supportable base exit range is about $0.9 billion to $1.5 billion if Iterative reaches roughly $160 million to $220 million of revenue and earns a 5x-7x hybrid multiple. | 低 | SV018, SV019, SV010 |
| CV048 | A supportable bear exit range is about $0.4 billion to $0.9 billion if revenue lands closer to roughly $110 million to $170 million and the market applies a 3x-4.5x multiple. | 低 | SV017, SV018, SV020 |
| CV049 | Probability-weighted value appears close to the current reported mark, leaving limited upside on public evidence and explaining why the call does not upgrade beyond research-more. | 低 | SV017, SV018, SV019, SV020 |
| CV050 | Risk rating should be high because the company must scale opaque economics while also navigating reimbursement and execution risk. | 中 | SV017, SV023, SV025, SV026 |
| CV051 | Valuation stance is stretched rather than clearly attractive because the current mark only looks fair if Iterative is already near premium AI/data economics. | 中 | SV002, SV018, SV019 |
| CV052 | Confidence in the recommendation is medium because sources triangulate funding and operating proof reasonably well, but financial disclosure is still too thin for a higher-confidence call. | 中 | SV001, SV002, SV003, SV017, SV020 |
| 编号 | 出版方 | 标题 | 引文 |
|---|---|---|---|
| SO001 | Iterative Health | Homepage – Iterative Health | |
| SO002 | Iterative Health | About Us – Iterative Health | |
| SO003 | Iterative Health | SKOUT | |
| SO004 | Iterative Health | Resources | |
| SO005 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | Today, Iterative Health’s global network includes more than 100 research sites across North America, Europe, India, and Australia, as well as partnerships with over 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SO006 | Iterative Health | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | |
| SO007 | Iterative Health | Iterative Scopes Announces $150 Million Series B to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SO008 | Iterative Health | Iterative Scopes Raises $30 Million Series A Financing to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SO009 | Iterative Health | Iterative Scopes Announces the Acquisition of Clinical Research Strategy Group and Precision Research | |
| SO010 | Iterative Health | Iterative Scopes Receives FDA Clearance for AI-Assisted Polyp Detection Device SKOUT™ | SKOUT demonstrated a 27% relative increase in the detection of adenomas per colonoscopy, with an average of one additional adenoma resected for every 4.5 patients examined. |
| SO011 | Iterative Health | Jonathan Ng, MBBS, Founder and CEO of Iterative Scopes, Named to PharmaVoice 100 | |
| SO012 | Iterative Health | Iterative Health and US Heart & Vascular Enter Strategic Partnership to Advance Community-Based Cardiovascular Research | |
| SO013 | Iterative Health | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation | |
| SO014 | Iterative Health | Iterative Health Acquires Cardiology Sites from NextStage Clinical Research | |
| SO015 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | |
| SO016 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | |
| SO017 | Insight Partners | Iterative Scopes Announces $150 Million Series B to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SO018 | Clearlake Capital | Iterative Scopes Announces $150 Million Series B to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SO019 | BioSpace | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | |
| SO020 | U.S. Food & Drug Administration | K251126: SKOUT system 510(k) summary | The SKOUT system is a software device designed to detect potential colorectal polyps in real time during colonoscopy examinations. |
| SO021 | Forge Global | Iterative Health IPO: Investment Opportunities & Pre-IPO Valuations | |
| SO022 | TechCrunch | Almost 40 new unicorns have been minted so far this year — here they are | |
| SO023 | AI2.work | Iterative Health Raises $77M to Fix Clinical Trials With AI | |
| SO024 | Tracxn | Iterative Health company profile | |
| SO025 | Translational Gastroenterology and Hepatology / PubMed Central | Computer-assisted detection of colorectal polyps: a narrative review of clinical utility, ongoing limitations, and opportunities for advancement | However, concerns about false positive rates, automation bias, operator deskilling, system integration, generalizability, and long-term outcomes persist. |
| SM001 | American Cancer Society | American Cancer Society Updates Colorectal Cancer Screening Guideline: Major Changes Emphasize Blood-Based and At-Home Stool Testing | |
| SM002 | Newswise | ACG/ASGE Release Updated Quality Indicators for Colonoscopy | |
| SM003 | AJMC | New Recommendations, Quality Indicators for Colonoscopy Released | |
| SM004 | Healio | New ACG, ASGE quality indicators for colonoscopy include 3 key updates | |
| SM005 | GI & Hepatology News | Expert weighs in on ADR-A as a colonoscopy quality metric | |
| SM006 | American College of Gastroenterology | Quality Indicators for Colonoscopy: New Targets… But Will They Be Measured? | |
| SM007 | Yale School of Medicine | AI-Assisted Colonoscopy: New Research and Guidelines for Clinical Use | |
| SM008 | Olympus Corporation | EAGLE Trial Shows Olympus CADDIE AI Solution Aids in the Detection of High-Risk and Hard-to-Detect Colorectal Lesions | |
| SM009 | PubMed Central | Current and future implications of artificial intelligence in colonoscopy | |
| SM010 | Iterative Health | New Journal Article in the Journal of Crohn’s and Colitis: Accelerating Enrollment in Phase 2 and 3 IBD Trials Through the Iterative Health Site Network | |
| SM011 | DocWire News | Improving Enrollment for Inflammatory Bowel Disease Trials Through Site Network | |
| SM012 | American Gastroenterological Association | Significant impacts to GI in CMS' proposed payment rules | |
| SM013 | Becker's ASC | 5 policy changes that could bite GI pay | |
| SM014 | American College of Gastroenterology | As Congress Eyes Medicare Reform, ACG Is Representing the Clinical GI Perspective | |
| SM015 | Medicare.gov | Colonoscopies (screening) | |
| SM016 | Boston Scientific | Endoscopy 2026 Procedural Payment Guide | |
| SM017 | PubMed Central | Holding gastroenterologists accountable for colonoscopy through MACRA episode-based cost measure | |
| SM018 | Iterative Health | SKOUT | |
| SM019 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | |
| SM020 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | |
| SM021 | PubMed Central | Trends in Colorectal Cancer Screening in US Adults Aged 45 to 49 Years | |
| SM022 | PubMed Central | Computer-assisted detection of colorectal polyps: a narrative review of clinical utility, ongoing limitations, and opportunities for advancement | |
| SM023 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | |
| SM024 | Clinical Trials Arena | Iterative Health banks $77m to level up AI-driven clinical research tech | |
| SM025 | U.S. Food and Drug Administration | K251126 SKOUT system 510(k) summary | |
| SP001 | Iterative Health | SKOUT | |
| SP002 | Iterative Health | Resources | |
| SP003 | Medtronic | GI Genius Intelligent Endoscopy Module | |
| SP004 | Olympus America | CADDIE Device Computer-assisted Detection | |
| SP005 | Olympus America | OLYSENSE | |
| SP006 | Olympus Europa | Endoscopy CAD system | ENDO-AID | Welcome to the AI Future in Endoscopy | |
| SP007 | Olympus Corporation | EAGLE Trial Shows Olympus CADDIE AI Solution Aids in the Detection of High-Risk and Hard-to-Detect Colorectal Lesions | |
| SP008 | Olympus Corporation | Olympus Launches OLYSENSE CAD/AI in the US and Europe | |
| SP009 | Fujifilm Healthcare Americas | AI Raises the Bar on Colonoscopy Performance with Fujifilm’s CAD EYE | |
| SP010 | Fujifilm | CAD EYE | Fujifilm [Singapore] | |
| SP011 | Odin Vision | Caddie - Odin Vision | |
| SP012 | Yale School of Medicine | AI-Assisted Colonoscopy: New Research and Guidelines for Clinical Use | |
| SP013 | GI & Hepatology News | AI-assisted colonoscopy linked to higher adenoma detection | |
| SP014 | PubMed Central | Clinical evaluation of a real-time artificial intelligence-based polyp detection system: a US multi-center pilot study | |
| SP015 | American Gastroenterological Association | Use of computer-aided detection systems (CADe) in colonoscopy | |
| SP016 | IQVIA | Gastroenterology & Hepatology | |
| SP017 | IQVIA | Site and Investigators | |
| SP018 | ICON plc | Gastrointestinal | ICON plc | |
| SP019 | ICON plc | Site & Patient Solutions | ICON plc | |
| SP020 | ICON plc | Global Site Network | ICON plc | |
| SP021 | Alimentiv | GI CRO | Gastroenterology Clinical Trials | Alimentiv | |
| SP022 | Alimentiv | Gastroenterology Clinical Trial Services | Alimentiv | |
| SP023 | Alimentiv | Site Network | Alimentiv | |
| SP024 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | |
| SP025 | Iterative Health | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation | |
| SP026 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | |
| SP027 | Pulse 2.0 | Iterative Health: $77 Million Raised To Scale Multispecialty Clinical Research Network Powered By AI | |
| SP028 | Yahoo Finance | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | |
| SP029 | Docwire News | Site Network for IBD Trials | |
| SI001 | Iterative Health | SKOUT | SKOUT is a real-time AI for polyp detection product page with evidence claims but no public list price. |
| SI002 | Iterative Health | Resources | |
| SI003 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | Today, Iterative Health’s global network includes more than 100 research sites ... as well as partnerships with over 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SI004 | Iterative Health | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | Today, Iterative Health is based in Cambridge, Massachusetts, and New York City with 250+ employees world-wide. |
| SI005 | Iterative Health | Iterative Scopes Announces the Acquisition of Clinical Research Strategy Group and Precision Research | CRSG/Precision work with a network of clinical research sites and community gastroenterologists nationwide ... enabling Iterative Scopes to enhance the value of its rapidly-growing AI Recruitment service. |
| SI006 | Iterative Health | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation | Its research ancillary, comprised of 13 active sites, is now an exclusive mainstay in Iterative Health’s global site network. |
| SI007 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | Iterative Health will lead trial management across current and future research sites, providing a proven operational infrastructure and supplying a robust pipeline of clinical trials. |
| SI008 | Iterative Health | Iterative Health and US Heart & Vascular Enter Strategic Partnership to Advance Community-Based Cardiovascular Research | USHV research sites will gain access to Iterative Health’s centralized operational backbone, providing end-to-end support across clinical research operations, services, business development, technology innovation, and corporate functions. |
| SI009 | Iterative Health | Iterative Health Acquires Cardiology Sites from NextStage Clinical Research | The company’s 250-person team, including site staff and centralized operations across multiple therapeutic areas, supports clinical, regulatory, financial, and operational functions. |
| SI010 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | The per-study randomization rate was 0.33 patients per site per month, representing more than 3x the industry benchmark of 0.10 patients per site per month. |
| SI011 | Forge | Iterative Health IPO: Investment Opportunities & Pre-IPO Valuations - Forge | Series C Valuation, Apr 2026: $1.3B ... Total Funding: $272.32MM. |
| SI012 | TechCrunch | Almost 40 new unicorns have been minted so far this year — here they are | Iterative Health — $1.4 billion ... last raised a $75 million Series C. It has raised more than $270 million in funding to date. |
| SI013 | AI2.work | Iterative Health Raises $77M to Fix Clinical Trials With AI | The round brings the company's total funding to over $273 million and reportedly values the company at approximately $1.3 billion. |
| SI014 | Tracxn | Iterative Health | Iterative Health has 166 employees as of Apr 26. |
| SI015 | American Gastroenterological Association | Significant impacts to GI in CMS' proposed payment rules | Together, these policies will cut payments to gastroenterologists for ASC and hospital-based endoscopy and E/M by $58 million. |
| SI016 | Becker's ASC Review | 5 policy changes that could bite GI pay | In 2025, Medicare pays about $220 per colonoscopy ... GI procedure pay dropped by 33% between 2007 and 2022, creating long-term instability for practices. |
| SI017 | Medicare.gov | Colonoscopies (screening) | Part B covers screening colonoscopies. |
| SI018 | Boston Scientific | GI Procedural Reimbursement Guide | The ASC payment system conversion factor is approximately 60% of that used in the OPPS. |
| SI019 | Clinical Gastroenterology and Hepatology / PMC | Holding gastroenterologists accountable for colonoscopy through MACRA episode-based cost measure | After the colonoscopy, providers are responsible for some costs that occur within 14 days of the procedure. |
| SI020 | NPJ Digital Medicine / PMC | Paying for artificial intelligence in medicine | Currently, CMS reimburses the use of at least 8 AI devices ... per-use reimbursement may result in overuse. |
| SI021 | Sidley Austin | Medicare Reimbursement Pathway for AI-Enabled Medical Devices Considered in Senate’s Health Tech Investment Act | The new technology APC is designed as a transitional reimbursement mechanism that would be assigned to a device for at least five years. |
| SI022 | Akin Gump | Artificial Intelligence in Clinical Decision-Making: Regulatory Roadmap and Reimbursement Strategies | There is no standard method for covering and paying every FDA-approved AI-enabled device. |
| SI023 | Harvard Medical School / Harvard Kennedy School | Transforming Healthcare with AI: Effective Reimbursement Can Lead to Better Care and Lower Costs | The authors propose ... Bundling, Value-based pricing, and Regular price adjustments. |
| SI024 | Securities and Exchange Commission / IQVIA | iqv-20250331 | Revenues $ 3,829 ... Income from operations 496 ... Cash and cash equivalents $ 1,740 ... Unearned income 1,940. |
| SI025 | ICON plc | Annual Reports | ICON plc 2025 20-F. |
| SI026 | ICON plc | Form 20-F for Icon PLC filed 05/27/2026 | A $92.7 million overstatement of revenue for the year ended December 31, 2024 and a $65.3 million overstatement of revenue for the year ended December 31, 2023. |
| SI027 | ICON plc | ICON Reports Fourth Quarter and Full Year 2025 Results | Full year revenue was $8,251.3 million ... Adjusted EBITDA was $1,530.7 million or 18.6% of revenue ... total backlog was $21.8 billion. |
| SI028 | Medpace | Annual Reports & Proxy | 2025 Annual Report. |
| SI029 | Medpace | Quarterly Results | Q4 2025 Earnings Press Release ... 10-K. |
| SI030 | Medpace | Form 10-K for Medpace Holdings INC filed 02/10/2026 | All statements other than statements of historical facts ... include statements regarding ... liquidity and our ability to fund our business operations and initiatives. |
| SI031 | Medpace | Medpace Holdings, Inc. Reports Fourth Quarter and Full Year 2025 Results | Revenue for the year ended December 31, 2025 increased 20.0% to $2,530.2 million ... EBITDA ... 22.0% of revenue ... Cash and cash equivalents were $497.0 million ... Backlog ... $3,027.2 million. |
| SE001 | Iterative Health | Homepage – Iterative Health | Partner with Iterative Health to streamline clinical trials, expand patient access, and leverage AI-powered tools that help sites and sponsors move research forward. |
| SE002 | Iterative Health | About Us – Iterative Health | By combining deep industry knowledge, operational excellence, and cutting-edge technology, we’re increasing access to research within the community care setting. |
| SE003 | Iterative Health | Careers – Iterative Health | We are a mission-driven, people-first team of builders who take ownership, work collaboratively, lead with transparency and accessibility, embrace resilience and innovation. |
| SE004 | Iterative Health | SKOUT | High-performing endoscopists resected 1 additional adenoma for every 4.5 patients examined with SKOUT, without increasing resection of hyperplastic polyps. |
| SE005 | Iterative Health | Resources | |
| SE006 | Iterative Health | Iterative Scopes Receives FDA Clearance for AI-Assisted Polyp Detection Device SKOUT™ | Provation is a market leader in gastrointestinal documentation... and the company will act as an exclusive distributor of SKOUT. |
| SE007 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | Updated analysis published confirms 3x patient randomization rates versus published industry benchmarks. |
| SE008 | Iterative Health | New Journal Article in the Journal of Crohn’s and Colitis: Accelerating Enrollment in Phase 2 and 3 IBD Trials Through the Iterative Health Site Network | The study demonstrated measurable improvements: Faster activation... Rapid first patient enrollment... Accelerated randomization. |
| SE009 | Iterative Health | New Journal Article in the Journal of Crohn’s and Colitis: AI-Enabled Endoscopy and Histology Reveal Novel Insights into Ulcerative Colitis | |
| SE010 | Iterative Health | BMJ Journals | Rethinking Central Reading in Ulcerative Colitis Trials: A Hybrid AI–Human Approach | The study evaluated a new central reading framework known as 2M+1H: 2M: Two independently developed machine learning models... 1H: A board-certified gastroenterologist adjudicates only when the models disagree. |
| SE011 | Iterative Health | Advancing GI Clinical Research with AI: Key Takeaways from ACG 2024 | |
| SE012 | U.S. Food & Drug Administration | K251126 SKOUT system 510(k) clearance letter | The polyp detection notification is a two-dimensional blue rectangular outline generated around any suspected polyps on the endoscopic video feed. |
| SE013 | PubMed Central | Improvement in adenoma detection using a novel artificial intelligence-aided polyp detection device | Overall, ADR with and without Skout was 54.2% and 40.6% respectively. |
| SE014 | PubMed Central | Current and future implications of artificial intelligence in colonoscopy | |
| SE015 | PubMed Central | Computer-assisted detection of colorectal polyps: a narrative review of clinical utility, ongoing limitations, and opportunities for advancement | Concerns about false positive rates, automation bias, operator deskilling, system integration, generalizability, and long-term outcomes persist. |
| SE016 | Yale School of Medicine | AI-Assisted Colonoscopy: New Research and Guidelines for Clinical Use | The guidelines provide evidence and make a conditional recommendation for using CADe to detect polyps in adults. |
| SE017 | The Healthcare Technology Report | FDA Clears SKOUT Polyp Detection Device From Iterative Scopes | |
| SE018 | GI & Hepatology News | AI-assisted colonoscopy linked to higher adenoma detection | A real-world analysis of more than 1.5 million matched patients links AI-assisted colonoscopy to a 47% reduction in interval colorectal cancer. |
| SE019 | Docwire News | Site Network for IBD Trials | |
| SE020 | Olympus | EAGLE Trial Shows Olympus® CADDIE™ AI Solution Aids in the Detection of High-Risk and Hard-to-Detect Colorectal Lesions | |
| SE021 | Medtronic | GI Genius™ Intelligent Endoscopy Module | |
| SE022 | Provation | SKOUT®: Real-Time AI for Polyp Detection | Artificial intelligence that augments but does not replace physician judgment. |
| SE023 | AccessGUDID | AccessGUDID - DEVICE: SKOUT system (00860010168417) | The SKOUT system integrates directly into the clinical workflow for colonoscopy procedures. |
| SE024 | startup.jobs | Iterative Health Jobs (June 2026) | Hiring for 52 positions. |
| SE025 | FUJIFILM Healthcare Americas Corporation | Fujifilm Announces the Commercialization of Two Novel Endoscopic Imaging Technologies, CAD EYE® and SCALE EYE® at the Digestive Disease Week (DDW®) 2024 Conference | |
| SU001 | Iterative Health | Homepage – Iterative Health | We bring GI, Hepatology, Obesity, and Cardiology research directly into the communities where care happens. |
| SU002 | Iterative Health | Providers – Iterative Health | 2X faster site activation; 3.4X increase in patient enrollments; AI-powered pre-screening technology demonstrates 40% uplift in patient randomizations. |
| SU003 | Iterative Health | Sponsor Inquiry – Iterative Health | Whether you’re a sponsor, CRO, or biotech company... request an enrollment forecast, and explore how Iterative Health can help. |
| SU004 | Iterative Health | Resources | Resources page lists GI Alliance, One GI, Gastro One, Takeda, and DDW 2026 materials across 2025-2026. |
| SU005 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | Today, Iterative Health’s global network includes more than 100 research sites... as well as partnerships with over 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SU006 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | The per-study randomization rate was 0.33 patients per site per month... Notably, among the four trials that completed full enrollment, the network maintained a global median randomization rate of 0.32. |
| SU007 | Iterative Health | New Journal Article in the Journal of Crohn’s and Colitis: Accelerating Enrollment in Phase 2 and 3 IBD Trials Through the Iterative Health Site Network | Faster activation: Median time from site selection to activation was 74 days... Accelerated randomization: Median time to first patient randomized was 83 days. |
| SU008 | Iterative Health | Case Study: How Gastro One Accelerated a Complex Phase 1b-A GI Trial with Iterative Health | Key Outcomes: 2x faster site activation; Higher patient retention; Increased randomizations; Over 50% of enrollments sourced through Iterative Health. |
| SU009 | Iterative Health | How Iterative Health and Takeda Are Using AI to Transform Inflammatory Bowel Disease Clinical Trials | Iterative Health processes... endoscopic videos and electronic health records from over 100 GI community sites across seven states. |
| SU010 | Iterative Health | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation | One GI’s network includes 34 clinics across six states... Its research ancillary, comprised of 13 active sites, is now an exclusive mainstay in Iterative Health’s global site network. |
| SU011 | Iterative Health | Iterative Health and US Heart & Vascular Enter Strategic Partnership to Advance Community-Based Cardiovascular Research | USHV research sites will join Iterative Health’s global site network, which now spans over 100 sites across three continents. |
| SU012 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | With the addition of 21 active GIA research sites—currently conducting over 80 Phase II–IV trials... the Iterative Health Site Network now extends to more than 70 locations in the US and Europe. |
| SU013 | OneGI | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation - OneGI | Iterative Health has been a trusted partner to our site for several years, and I’ve seen firsthand the depth of their support and expertise. |
| SU014 | U.S. Heart & Vascular | Iterative Health and US Heart & Vascular Enter Strategic Partnership | Under the agreement, USHV research sites will join Iterative Health’s global site network, which now spans over 100 sites across three continents. |
| SU015 | Gastro Health | Iterative Health Partners with Gastro Health to Accelerate Clinical Research in Gastroenterology Through Artificial Intelligence | Gastro Health... has a presence in seven states... with over 380 physicians and 150 locations... The partnership will allow Gastro Health physicians to leverage Iterative Health’s AI-based technology. |
| SU016 | Yahoo Finance | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | This is a paid press release... The per-study randomization rate was 0.33 patients per site per month, representing more than 3x the industry benchmark of 0.10. |
| SU017 | Morningstar | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | One year into our strategic partnership with Iterative Health, this data reflects what becomes possible when infrastructure, technology, and clinical expertise converge at the community level. |
| SU018 | Business Wire | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | The DDW 2026 presentation represents the second major public reporting of Iterative Health’s aggregated network performance. |
| SU019 | Pulse 2.0 | Iterative Health: $77 Million Raised To Scale Multispecialty Clinical Research Network Powered By AI | The company’s global network now includes more than 100 research sites... alongside partnerships with more than 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SU020 | Docwire News | Site Network for IBD Trials | There were 27 Iterative Health network sites in this study... 21 in the US and six in Europe... 0.34 patients per site per month compared to published industry, which is currently around 0.1. |
| SU021 | BioSpace | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | With the addition of 21 active GIA research sites—currently conducting over 80 Phase II–IV trials... the Iterative Health Site Network now extends to more than 70 locations in the US and Europe. |
| SU022 | BioSpace | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | Bill brings over two decades of experience... most recently serving as CFO of GI Alliance. |
| SU023 | Iterative Health | SKOUT | SKOUT is super accurate and reliable... It’s absolutely essential to have AI polyp detection... |
| SU024 | Medtronic | GI Genius™ Intelligent Endoscopy Module | GI Genius™ intelligent endoscopy module has a 99.7% sensitivity rate and less than 1% false positives. |
| SU025 | Oxford University Press | ECCO JCC abstract page for Accelerating Enrollment in Phase 2 and 3 Inflammatory Bowel Disease Trials Through an Innovative Site Network | |
| SR001 | Iterative Health | Iterative Health | By placing sites at the center, we power a global, performance-driven network with exceptional patient reach, trusted by Sponsors/CROs and Providers alike. |
| SR002 | Iterative Health | About Us - Iterative Health | Meet the people guiding our vision, culture, and impact. |
| SR003 | Iterative Health | Providers – Iterative Health | 2X faster site activation; 3.4X increase in patient enrollments. |
| SR004 | Iterative Health | Sponsor Inquiry – Iterative Health | Whether you’re a sponsor, CRO, or biotech company... request an enrollment forecast. |
| SR005 | Iterative Health | SKOUT - Iterative Health | High-performing endoscopists resected 1 additional adenoma for every 4.5 patients examined with SKOUT. |
| SR006 | Iterative Health | Resources - Iterative Health | Resources page lists customer and performance stories, but no trust-center style security materials. |
| SR007 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | Today, Iterative Health’s global network includes more than 100 research sites... and partnerships with over 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SR008 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | The per-study randomization rate was 0.33 patients per site per month... 3x the industry benchmark of 0.10. |
| SR009 | Iterative Health | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | Bill Kayser has joined the company as President and Chief Financial Officer. |
| SR010 | Iterative Health | Privacy Policy | Company employs industry standard organizational, technical, administrative and technological measures that are reasonably designed to help protect information. |
| SR011 | AME Publishing / PubMed Central | Computer-assisted detection of colorectal polyps: a narrative review of clinical utility, ongoing limitations, and opportunities for advancement | Concerns about false positive rates, automation bias, operator deskilling, system integration, generalizability, and long-term outcomes persist. |
| SR012 | Digestive Medicine Research / PubMed Central | Current and future implications of artificial intelligence in colonoscopy | The main ones being the fear of medico-legal implications, the lack of financial incentives to use optical diagnosis. |
| SR013 | Nature Medicine / PubMed Central | Paying for artificial intelligence in medicine | Per-use AI reimbursement may result in overuse—an undesirable outcome of AI reimbursement policy. |
| SR014 | Frontline Gastroenterology / BMJ | Artificial intelligence in endoscopy: navigating risk, responsibility and ethical challenges | False positive rates ranging from 1.44% to 3.40% may lead to unnecessary interventions and prolonged procedures. |
| SR015 | American College of Gastroenterology | Artificial intelligence in colonoscopy: Could it be making us worse? - American College of Gastroenterology | ADR before vs after AI exposure decreased significantly from 28.4% to 22.4%. |
| SR016 | GI & Hepatology News | AI-assisted colonoscopy linked to higher adenoma detection | Artificial intelligence–assisted colonoscopy was associated with higher adenoma detection and nearly half the rate of interval colorectal cancer. |
| SR017 | Sidley Austin LLP | Medicare Reimbursement Pathway for AI-Enabled Medical Devices Considered in Senate’s Health Tech Investment Act | Insights | Sidley Austin LLP | There still is no consistent and predictable payment pathway for the use of these devices in healthcare delivery under Medicare. |
| SR018 | Harvard Medical School – Health Care Policy | Transforming Healthcare with AI: Effective Reimbursement Can Lead to Better Care and Lower Costs | The current fee-for-service payment system... is not easily adapted to these AI-enabled clinical services. |
| SR019 | American Gastroenterological Association | Significant impacts to GI in CMS' proposed payment rules | Together, these policies will cut payments to gastroenterologists for ASC and hospital-based endoscopy and E/M by $58 million. |
| SR020 | Becker’s ASC Review | 5 policy changes that could bite GI pay - Becker’s ASC | Adjusted for inflation, GI procedure pay dropped by 33% between 2007 and 2022. |
| SR021 | Medicare.gov | Colonoscopies (screening) | Part B covers screening colonoscopies. |
| SR022 | Docwire News | Site Network for IBD Trials | Docwire News | There were 27 Iterative Health network sites in this study... 0.34 patients per site per month compared to published industry, which is currently around 0.1. |
| SR023 | ai2.work | Iterative Health Raises $77M to Fix Clinical Trials With AI | The round brings the company total funding to over $273 million and reportedly values the company at approximately $1.3 billion. |
| SR024 | GI Alliance | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community - GI Alliance | With the addition of 21 active GIA research sites... the Iterative Health Site Network now extends to more than 70 locations in the US and Europe. |
| SR025 | One GI | Iterative Health and One GI Announce Strategic Partnership, Advancing Research as a Driver of Care Innovation - OneGI | One GI’s network includes 34 clinics across six states... 13 active sites. |
| SR026 | US Heart & Vascular | Iterative Health and US Heart & Vascular Enter Strategic Partnership - US Heart & Vascular | USHV research sites will join Iterative Health’s global site network, which now spans over 100 sites across three continents. |
| SR027 | Olympus America | CADDIE™ Device Computer-assisted Detection | Olympus America | Medical | The CADDIE™ device is not intended to replace a full patient evaluation... and is limited for use with standard white-light endoscopy imaging only. |
| SR028 | Medtronic | GI Genius™ Intelligent Endoscopy Module | GI Genius™ intelligent endoscopy module has a 99.7% sensitivity rate and less than 1% false positives. |
| SR029 | BusinessWire | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | The abstract builds on findings previously published at ECCO, now reflecting an approximately two-fold increase in network scale across trials and sites. |
| SR030 | BioSpace | Iterative Health Appoints Bill Kayser as President and Chief Financial Officer | Bill brings over two decades of experience in healthcare finance, strategy, and leadership, most recently serving as CFO of GI Alliance. |
| SR031 | BioSpace | GI Alliance and Iterative Health Partner to Advance GI Clinical Research in the Community | With the addition of 21 active GIA research sites—currently conducting over 80 Phase II–IV trials... the network now extends to more than 70 locations. |
| SR032 | Greenhouse | Iterative Health | Iterative Health careers board. |
| SR033 | Iterative Health | Driving Operational Excellence in Ambulatory Surgery Centers: How Iterative Health Accelerates Clinical Trial Success | Iterative Health helps Ambulatory Surgery Centers accelerate clinical trials by streamlining operations, improving recruitment, and ensuring compliance. |
| SR034 | Iterative Health | Iterative U: Advancing Clinical Research Through On-Demand Learning | Iterative U provides on-demand, role-specific training for clinical research teams. |
| SR035 | Iterative Health | Beyond Academic Walls: The Case for Community-Based Cardiovascular Trials | Beyond Academic Walls: The Case for Community-Based Cardiovascular Trials |
| SV001 | Iterative Health | Iterative Health Closes $77 Million Series C to Accelerate the Future of Clinical Research | Today, Iterative Health’s global network includes more than 100 research sites ... as well as partnerships with over 40 pharmaceutical, biotech, medical device, and contract research organizations. |
| SV002 | Forge Global | Iterative Health IPO: Investment Opportunities & Pre-IPO Valuations - Forge | $1.3B Series C Valuation, Apr 2026 |
| SV003 | TechCrunch | Almost 40 new unicorns have been minted so far this year — here they are | Iterative Health — $1.4 billion |
| SV004 | ai2.work | Iterative Health Raises $77M to Fix Clinical Trials With AI | ...reportedly values the company at approximately $1.3 billion — making it the latest AI-powered healthtech to reach unicorn status. |
| SV005 | Founder Lodge | Iterative Health raises $77,000,000 at Series C on 2026-05-01 | |
| SV006 | Tracxn | Iterative Health | |
| SV007 | Iterative Scopes | Iterative Scopes Announces $150 Million Series B to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SV008 | Iterative Scopes | Iterative Scopes Raises $30 Million Series A Financing to Advance AI-Driven Precision Medicine for Gastroenterology | |
| SV009 | Iterative Health | Iterative Health Reports Expanded Trial Performance Data at Digestive Disease Week 2026, Demonstrating Consistent Outperformance of Industry Benchmarks at Double the Scale | 0.33 patients per site per month, representing more than 3x the industry benchmark |
| SV010 | Medpace | Medpace Holdings, Inc. Reports First Quarter 2026 Results | Revenue of $706.6 million in the first quarter of 2026 increased 26.5%... EBITDA margin of 21.1%. |
| SV011 | CompaniesMarketCap | IQVIA (IQV) - Market capitalization | |
| SV012 | CompaniesMarketCap | IQVIA (IQV) - Revenue | |
| SV013 | CompaniesMarketCap | ICON plc (ICLR) - Market capitalization | |
| SV014 | CompaniesMarketCap | ICON plc (ICLR) - Revenue | |
| SV015 | CompaniesMarketCap | Medpace (MEDP) - Market capitalization | |
| SV016 | CompaniesMarketCap | Medpace (MEDP) - Revenue | |
| SV017 | Finro | AI Valuation Multiples Q1 2026: Investors Reprice Quality | Companies still selling “growth now, business model later” faced sharper discounts as underwriting shifted toward durability rather than narrative momentum. |
| SV018 | Nelson Advisors | HealthTech M&A Multiples January 2026: Current Trends and Variables driving valuations | General HealthTech: ~4x–6x revenue remains the central band ... Premium AI, telehealth & analytics: 6x–8x+ |
| SV019 | Healthcare Digital | HealthTech and MedTech M&A 2026 Valuation Multipliers | ...high-quality assets with defensible moats command premium multiples, while those lacking operational leverage or regulatory readiness face significant compression. |
| SV020 | PitchBook | Q1 2026 PitchBook Analyst Note: Sizing the US Unicorn Herd | Nearly half have not been priced in the past three years, masking how much the market has changed since the 2021 venture boom. |
| SV021 | Prime Unicorn Index | Prime Unicorn Index--The Definitive Benchmark for U.S. Venture-Backed Unicorns | |
| SV022 | VNTR | Q1 2026: The AI Unicorn Factory Goes Into Overdrive | This leaves investors with a critical question: is this velocity a rational response ... or a familiar sign of late-cycle valuation inflation? |
| SV023 | Sidley Austin | Medicare Reimbursement Pathway for AI-Enabled Medical Devices Considered in Senate’s Health Tech Investment Act | The proposed bill would amend Title XVIII ... to create a Medicare payment system for algorithm-based healthcare services. |
| SV024 | Health Affairs | Paying for artificial intelligence in medicine | CMS can use the FDA’s regulatory standards ... however, no formal guidance regarding standards for CMS coverage of innovative technologies such as AI... |
| SV025 | American Gastroenterological Association | Significant impacts to GI in CMS proposed payment rules | Practices that provide ASC and hospital outpatient department endoscopy will experience an average 8% cut in physician payments for endoscopy compared to 2025 Medicare rates. |
| SV026 | Becker’s ASC Review | 5 policy changes that could bite GI pay | |
| SV027 | Frontiers in Medicine | Computer-assisted detection of colorectal polyps: a narrative review of clinical utility, ongoing limitations, and opportunities for advancement | Uncertainties remain regarding generalizability, cost-effectiveness, and long-term outcomes. |
| SV028 | U.S. Securities and Exchange Commission | Medpace Holdings 2025 Form 10-K | |
| SV029 | U.S. Securities and Exchange Commission | IQVIA Holdings 2025 Form 10-K | |
| SV030 | U.S. Securities and Exchange Commission | ICON plc Form 20-F |