XiangshuTech — Spatial AI for Surgery

从范德堡走出来的医疗 Spatial AI 公司A medical Spatial AI company from Vanderbilt

让外科医生看穿组织Let surgeons see through tissue

让手术室能"看见"、能"思考"、能"协作"。我们的平台建立在 IEEE TBME、IEEE RA-L 等顶级期刊发表、经亚表面靶点定位量化验证的研究之上 —— 把术前影像、术中变形与组织语义实时融合进医生的视野。An operating room that can see, think, and collaborate. Built on research published in IEEE TBME and IEEE RA-L, quantitatively validated by subsurface target localization — fusing preoperative imaging, intraoperative deformation, and tissue semantics directly into the surgeon's view.

探索技术平台Explore the Platform 投资合作Investor Inquiries

三个核心数字Three Key Numbers

临床的痛点 · 技术的突破 · 市场的体量The clinical pain. The technical breakthrough. The market.

45 万例

450K

中国每年软组织肿瘤手术量China — soft-tissue tumor surgeries / year

中国年新发肝癌约 37 万例（年切除 15 万例），年新发乳腺癌 42 万例（保乳 30 万例）。仅这两个术种合计 45 万例 —— 乳腺保乳再切除率高达 17.2%。 37k new liver cancer cases (15k resections) + 42k breast cancer cases (30k lumpectomies) per year. Breast lumpectomy re-excision rate: 17.2%.

10.4 → 4.2 mm

行业首次跨越临床可用线（5mm）First to cross the 5mm clinical usability line

基于范德堡原创 LIBR / LIBR+ 算法，肝脏活体亚表面误差 10.4→6.1mm，乳腺活体 10.4→4.2mm，仿体测试低至 1.7±0.8mm 。 Built on Vanderbilt's original LIBR / LIBR+: liver subsurface error 10.4→6.1mm, breast 10.4→4.2mm, phantom tests as low as 1.7±0.8mm.

$24.5B

2030 年全球手术导航市场Global surgical navigation market by 2030

软组织子赛道增长最快，且至今 尚未形成主导者。我们的 SAM（中国+高端国际）约 170 亿元人民币；仅需 5-10% 份额即可支撑 IPO 体量。 Soft-tissue segment grows fastest — and no dominant player yet. Our SAM (China + premium international) is ~¥17B by 2030. Even 5-10% market share supports IPO-scale revenue.

为什么是现在Why Now

2026 年是唯一的奇点窗口2026 is the singular window

五个独立变量首次同时到位 —— Spatial AI for Surgery 在过去十年都做不成。Five independent variables align for the first time. This couldn't be built any earlier.

01 · HARDWARE

医疗级 MR 设备进入手术室Medical-grade MR enters the OR

Apple Vision Pro 2025 年完成全球首例临床手术，至 2026 累计数百例；国产 MR（灵伴/Rokid）医疗专用化方案成熟。Apple Vision Pro completed the first clinical procedures in 2025, hundreds by 2026. Chinese MR vendors (Rokid, Lingban) maturing for medical use.

02 · AI PARADIGM

Spatial Intelligence 成主流Spatial Intelligence goes mainstream

Fei-Fei Li 创立的 World Labs 2026 年 2 月以 $50 亿估值融资 $10 亿 —— 资本市场首次明确把空间智能与 LLM 并列。Fei-Fei Li's World Labs raised $1B at $5B valuation in Feb 2026 — markets now treat Spatial AI as a peer to LLMs.

03 · ROBOTICS

协作机械臂成本拐点Cobot cost inflection

KUKA LBR Med、KINOVA Gen3 Med 等医疗合规协作机械臂价格下探至 30-50 万元区间，首次商业化可行。Medical-compliant cobots (KUKA LBR Med, KINOVA Gen3) dropped to ¥300k-500k — first time commercially viable.

04 · MARKET PROOF

国产手术导航被验证可放量Chinese surgical navigation proven scalable

天智航 2025 H1 营收同比 +114.89%，骨科手术导航定位系统销售 +641.83% —— 国产导航品类的真金白银信号。Tinavi (688277) H1 2025: revenue +114.89% YoY, orthopedic nav system sales +641.83%. Real hospital procurement at scale.

05 · POLICY

医保独立支付通道开通Independent reimbursement channel

国家医保局 2026 年发布"导航 / 参与执行 / 精准执行"三档收费政策，导航类产品首次获得清晰的医院收入回收路径。China's NHSA 2026 policy creates three tiers (Navigate / Assist / Execute) — surgical navigation has a clear reimbursement path for the first time.

Surgical Spatial AI 平台Surgical Spatial AI Platform

三层架构 · 围绕医生构建Three-layer architecture, surgeon-centered

所有感知、计算、协作都围绕一个目标：让医生在最自然的状态下，做出更准确的判断。Every layer of perception, computation, and collaboration serves one goal — empower the surgeon to make better decisions, naturally.

感知层 · 多模态空间感知 Perception Layer · Multi-modal Spatial Sensing RGB-D · AHAT · 光学追踪 · 超声 · 协作臂主动视觉RGB-D · AHAT · Optical Tracking · Ultrasound · Cobot Vision

RGB-D 深度相机、AHAT 深度感知、光学追踪、术中超声、协作机械臂主动视觉 —— 多模态融合实时构建术野空间模型。RGB-D depth, AHAT, optical tracking, intraoperative ultrasound, and cobot active vision — fusing multi-modal signals into a live spatial model of the surgical field.

计算层 · Spatial AI 核心算法 Cognition Layer · Spatial AI Core LIBR / LIBR+ · 生物力学 · 实时重建 · 语义分割LIBR / LIBR+ · Biomechanics · Real-time Recon · Semantic Segmentation

LIBR / LIBR+ 软组织变形补偿、生物力学建模、实时三维重建、组织语义分割 —— 把空间数据转化为可信赖的术中智能。LIBR / LIBR+ deformation correction, biomechanical modeling, real-time 3D reconstruction, and tissue semantic segmentation — turning spatial data into trustworthy intraoperative intelligence.

协作层 · Surgeon Co-Pilot Co-Pilot Layer · Surgeon Co-Pilot HoloLens / Vision Pro · 语音手势 · 机械臂避障 · Virtual IFSHHoloLens / Vision Pro · Voice & Gesture · Obstacle Avoidance · Virtual IFSH

HoloLens / Vision Pro AR 显示、语音手势免手控制、协作机械臂主动避障、Virtual IFSH 实时虚拟病理 —— 把信息以最自然的方式送达医生眼前。HoloLens / Vision Pro AR display, voice & gesture hands-free control, active obstacle avoidance, and Virtual IFSH real-time pathology — delivering information into the surgeon's eyes at the right moment.

应用场景Applications

从诊室到手术室 · 空间智能的全流程价值From clinic to OR — value across the whole pathway

同一套 Spatial AI 平台，服务患者沟通、医学教学与术中导航三大场景。One Spatial AI platform serving three settings: patient consultation, teaching & training, and the operating room.

患者沟通Patient Consultation

用患者自己的三维解剖与手术方案，帮助患者与家属更直观地理解诊断与手术路径，在混合现实中实现共同决策。Show patients their own anatomy and the surgical plan in 3D — helping patients and families understand the diagnosis and approach, and enabling shared decision-making in mixed reality.

教学与培训Teaching & Training

多用户同步查看同一台全息影像，讲师与学员共享同一空间视角 —— 把复杂解剖与术式讲解从二维幻灯片带入真实三维空间。Multiple users view the same hologram in sync, sharing one spatial perspective — moving complex anatomy and procedural teaching from 2D slides into real 3D space.

手术室Operating Room

术中把术前影像、软组织形变矫正、器械位置与决策辅助实时叠加到医生视野中 —— 亚表面靶点定位、消融导航与实时切缘评估，皆在术野原位完成。Intraoperatively overlay preoperative imaging, soft-tissue deformation correction, instrument position, and decision support in the surgeon's view — subsurface targeting, ablation guidance, and real-time margin assessment, in situ.

临床演示Clinical Demos

看见我们的技术在真实手术场景中工作See the platform working in real surgical settings

MR 超声实时叠加MR Ultrasound Real-Time Overlay · 超声切面直接叠加于患者解剖正确位置 —— ~2mm 精度、< 16ms 延迟。 Ultrasound projected onto correct anatomical position — ~2mm accuracy, <16ms latency.

微波探针导航Microwave Probe Navigation · 实时探针轨迹可视化、距靶点距离测量、消融区预测。 Live probe trajectory, target distance, and ablation zone prediction.

2D 四视图同步导航2D Four-Panel Navigation · 轴位 / 矢状位 / 冠状位 + 3D 模型同步显示。 Axial · Sagittal · Coronal + 3D model in synchronized view.

消融路径规划与可视化Ablation Path Planning & Visualization · 基于患者个体化组织特性，预测消融范围、评估肿瘤覆盖与周围组织风险。 Patient-specific ablation zone prediction with tumor coverage and risk visualization.

最新研究成果Latest Research

两篇新发表 · 两套下一代系统Two new publications · Two next-generation systems

IEEE TBME
2026

完整 MR 肝脏手术导航系统 · 亚表面靶点定位综合验证Comprehensive MR Liver Navigation · Subsurface Target Localization

Bowen Xiang, Jon S. Heiselman, Dingjie Su, Michael I. Miga · IEEE Transactions on Biomedical Engineering

已发表Published DOI ↗

IEEE RA-L
2026

ReSNA · 面向保乳手术的机器人手术导航助手ReSNA · Robotic Navigation Assistant for Breast-Conserving Surgery

Zixuan Liu, Bowen Xiang, Jon Heiselman, Jie Ying Wu, Ingrid Meszoely, Michael I. Miga · IEEE Robotics and Automation Letters

已发表Published DOI ↗

研发中In Dev
2026

自成一体的乳腺 MR 导航系统 · 单头显免外部追踪Self-Contained Breast MR Navigation · Single Headset, No External Tracker

Bowen Xiang, Zixuan Liu, Jon S. Heiselman, Michael I. Miga · 拟投 IEEE TBMETarget: IEEE TBME

研发中In Development

研发中In Dev
2026

无标记 · 免手 · 全自动 · 学习驱动的肝脏 MR 导航Markerless · Hands-Free · Fully Automatic · Learning-Based Liver MR

Bowen Xiang, Jon S. Heiselman, Michael I. Miga · 拟投 IEEE TBMETarget: IEEE TBME