A Robot That Sutures Blood Vessels on Its Own
Vascular anastomosis is the surgical connection of two blood vessels.
It is a core step in organ transplants, reconstructive surgery, and tissue transfers.
The work demands years of training, which limits how many surgeons can do it well.
A Johns Hopkins team has now built a robot that places most of these stitches by itself.
The study is accepted in IEEE Transactions on Medical Robotics and Bionics, and its publication record is dated September 14, 2026.
What μSTAR Actually Does
The system is called the Micro Smart Tissue Autonomous Robot, or μSTAR.
It uses a robotic arm that carries a suturing tool fitted with two sensors.
The first sensor is optical coherence tomography, a fiber-optic probe that reads tissue depth in real time.
The second is a tiny camera that photographs each stitch site before and after the needle passes.
The probe finds the edge of the vessel so the needle enters at the right distance from the cut.
The camera feeds a neural network that checks whether a stitch was missed.
A motorized clamp rotates the vessel between stitches, so the robot can reach every side.
How It Compared With Surgeons
The team tested the robot on five-millimeter pig arteries and compared it with three surgeons, each repairing five vessels.
The robot placed 90% of its sutures without human intervention.
Its stitch depth averaged 1.54 millimeters against a 1.5 millimeter target.
Its stitch spacing varied by 30%, compared with a range of 27% to 62% across the three surgeons.
Leak pressure, the main test of whether the repair holds, showed no statistically significant difference between the robot and the surgeons.
Lumen narrowing, which measures how much the repair squeezes the vessel, was 26% for the robot and ranged from 21% to 71% for the surgeons.
What the Test Did Not Prove
The robot was much slower, at about 353 seconds per stitch compared with about 141 seconds for the surgeons.
A graduate student tied the knots, and a person also pulled each suture through and cut it.
The robot used a thicker 3-0 suture, while the surgeons used a finer 6-0 suture that is closer to clinical practice.
The vessels were about five millimeters wide, much larger than the sub-millimeter vessels where microsurgery is hardest.
The robot arm lost its connection five times during the experiments, and one vessel slipped enough to cause a crossed stitch.
The stitch-checking network reached 87% accuracy on its test images, and the edge-finding probe located the vessel edge in about 90% of scans.
All of this was done on tissue outside the body, so no living patient was involved.
Why It Matters
Revision rates for manual vessel repair reach as high as 7.9% in the studies the authors cite.
A robot that places stitches with consistent depth and spacing targets that variability directly.
The result is a research milestone, not a clinical tool, and the gap between the two is the part worth tracking.
The authors name the next steps themselves: finer sutures, automated knot tying, faster motion, and smaller vessels.
Each of those steps will show how close vascular robots are to doing more than assisting.
The Autonomy Registry checks claims like this one against primary sources and places them on the six-level autonomy ladder.
