Autonomous Robotic and Microrobotic Surgery: Progress and Roadmaps in Five Surgical Areas

Introduction

Robot‑assisted surgery (RAS) has expanded since the first neurosurgical biopsy robot in 1985. Miniaturization enables future systems to navigate natural body openings for partially or fully autonomous treatment, potentially shortening recovery and reducing cost.

Vascular System

Endovascular interventions provide a clear anatomical path, repeatable motions, and rely on imaging rather than tactile feedback, making them the most ready for automation. Two automation directions are identified: (1) AI‑driven planning and guidance that integrates multimodal imaging and procedural knowledge to generate personalized paths; (2) autonomous motion‑planning and navigation controllers that decompose surgery into executable sub‑tasks confirmed by the surgeon before execution.

Current consensus prioritizes automating the most challenging navigation sub‑tasks, while simple actions such as vascular puncture remain manual. Electromagnetically controlled guidewires and ultrasound‑guided helical microrobots have demonstrated multi‑fold improvements in thrombolysis efficiency, but remain in pre‑clinical validation.

Lumen System

Short‑term breakthroughs focus on directed navigation to known targets, leveraging the natural constraints of bodily lumens for automatic path planning. Exploratory navigation that requires patient‑specific reasoning remains more difficult.

Intelligent gastrointestinal capsules and urological micro‑graspers have entered clinical or pre‑clinical stages, enabling autonomous endoscopic inspection, drug delivery, and tissue biopsy, but still face positioning‑accuracy limitations and risk of capsule retention.

Laparoscopic System

Laparoscopic surgery involves deformable soft tissue and significant anatomical variation, raising automation difficulty compared with orthopedics. Commercial systems are expensive and require extensive surgeon training, limiting adoption to high‑volume centers.

Near‑term automation targets low‑level autonomous sub‑tasks such as stable camera positioning, organ retraction, and intra‑operative suction. Real‑time 3D soft‑tissue reconstruction, force‑feedback sensing, multimodal imaging‑driven deformation models, and flexible robotic instruments are required to achieve these tasks.

Digital endoscopic imaging combined with AI analysis can recognize surgical steps and trigger intra‑operative alerts, providing essential data for AI‑assisted systems.

Ophthalmic System

The eye’s optical transparency and anatomical stability make ophthalmology well‑suited for early automation. Partial automation already exists in refractive laser surgery; robotic assistance in vitreoretinal procedures can improve precision from the human tremor limit of 40‑100 µm to 1‑10 µm, reducing micro‑bleeding.

Future automation may include emergency laser iridotomy, intravitreal drug injection, and high‑volume cataract surgery. Required technologies are curved tools matching ocular geometry, real‑time OCT‑microscope image fusion, sub‑micron force sensing, and safety response within 200 ms. Digital twins could support personalized pre‑operative planning and intra‑operative feedback.

Orthopedic System

Rigid bone anatomy and mature image‑guided navigation make orthopedics favorable for automation. Generative AI models can reconstruct 3D anatomy from 2D images, reducing intra‑operative radiation. Smart bone implants equipped with drug‑delivery and sensing capabilities enable long‑term monitoring, targeted anti‑infection treatment, and promotion of bone growth.

Technical Bottlenecks for Microrobotic Surgery

Microrobots span centimeter to nanometer scales, offering access to anatomical regions unreachable by conventional tools. Three primary challenges impede full autonomy:

Energy supply: conventional batteries cannot fit micron‑scale devices; emerging zinc‑air micro‑batteries remain experimental.

On‑board computation and decision‑making: most systems rely on external controllers; true autonomy requires integrating sensing, processing, and communication within a tiny volume.

Safety mechanisms: in‑body tracking and retrieval of micron‑scale robots are difficult; biodegradable or bio‑absorbable materials and robust fail‑safe designs are essential.

A pragmatic near‑term solution is a hybrid control mode where microrobots execute simple local loops while an external system provides global supervision and scheduling, balancing flexibility with clinical safety. Long‑term progress will need standardized multimodal data formats, integration of imaging streams, and alignment of responsibility boundaries with existing surgical training and quality‑control frameworks.

Source: https://www.science.org/doi/10.1126/sciadv.aec4197

Code example

来源：ScienceAI
本文
约2300字
，建议阅读
5
分钟
结合外科实践和工程创新，迈向完全自主的微型辅助机器人。