**Background**
Microsurgery involves operating on small, delicate structures under magnification, and is used in ophthalmology, otolaryngology, neurosurgery, reconstructive surgery, and urology. Despite its benefits, manual microsurgery faces four major challenges: (1) manipulation of micron-scale targets requiring high precision, where even slight tremor (RMS amplitude ~182 µm) can cause injury; (2) limited perceptual feedback due to restricted microscope field of view and imperceptible tool–tissue forces; (3) prolonged uncomfortable surgeon postures leading to fatigue; and (4) extensive training requirements. Microsurgery robotic (MSR) systems aim to overcome these limitations by providing tremor filtering, motion scaling, force feedback, and automation. This review provides a comprehensive technical overview of MSR systems, covering mechanism design, sensing, human–machine interaction (HMI), and automation, and discusses classic systems and future directions.
**Methods**
The authors conducted a narrative review of MSR-related literature published from 2000 to 2022, identified via Google Scholar keyword searches. They categorized MSR technologies into four areas: (1) operation modes and mechanism designs (handheld, teleoperated, co-manipulated, partially automated; RCM vs. non-RCM structures); (2) sensing and perception (imaging modalities: MRI, CT, OCT, NIRF, exoscope, UBM; 3D localization; force sensing using electrical strain gauges and fiber Bragg grating (FBG) sensors); (3) HMI (force feedback—direct and sensory substitution; tremor filtering; motion scaling; virtual fixtures; extended reality (XR) for training and intraoperative guidance); and (4) automation (current semi-automated methods and potential machine learning approaches like Learning from Demonstration and Reinforcement Learning). Classic MSR systems (NeuroArm, REMS, MUSA, IRISS, Preceyes Surgical System, Co-Manipulator System) were described in detail with parameters such as DOF, workspace, precision, and research progress.
**Key Results**
The review identifies that manual microsurgery precision is limited by physiological tremor (RMS ~182 µm), while many microsurgical tasks require accuracy of 10–25 µm. MSR systems achieve enhanced precision: for example, the Preceyes Surgical System (PSS) has intrinsic precision of 10 µm; the IRISS achieves positional precision of 27 ± 3 µm; the MUSA system offers slave end-effector precision of about 70 µm (bidirectional 30–40 µm). Force sensing using FBG sensors can achieve lateral resolution down to 0.15 mN. In clinical studies, the PSS enabled cannulation of ~60 µm diameter pig retinal venules and was used in first-in-human robot-assisted subretinal drug delivery (12 patients). The Co-Manipulator system achieved successful retinal vein cannulation in 15 out of 18 pig eyes and in a phase I clinical trial with 4 patients, maintaining injection for 10 minutes. The NeuroArm system has been used in dozens of clinical neurosurgery cases. The REMS system demonstrated sub-millimeter accuracy (average registration error 0.46 ± 0.22 mm) in cadaver models. The MUSA system completed first-in-human robot-assisted supermicrosurgical lymphatico-venous anastomosis. Automation examples include semi-automated cataract removal and automated retinal vein cannulation with targeting within 20 μm.
**Clinical Implications**
MSR systems have the potential to significantly improve surgical outcomes by enhancing precision, stability, and safety in microsurgery. They can reduce complications from tremor and limited perception, enable procedures that are difficult or impossible manually (e.g., retinal vein cannulation, supermicrosurgical anastomosis), and improve surgeon ergonomics. However, challenges remain including clinical acceptance, ethical/legal concerns, need for interdisciplinary development, human factors integration, visualization limitations, control robustness, network security, and training requirements. Future directions include further human factors consideration, multiple sensor fusion, and higher levels of autonomy using machine learning. The review emphasizes that MSR systems are not yet widely adopted but show great promise for expanding the capabilities of microsurgery across multiple specialties.