CUVIS-joint: a technical system profile

In brief

• CUREXO’s CUVIS-joint uses CT, 3D planning, optical tracking and a robotic arm.
• FDA K252037 covers CJ150 bone preparation for primary TKA with stated United U2 compatibility.
• The manufacturer describes active resection, but the platform implements an approved plan under surgeon direction.
• A reported THA authorisation is regional to Korea and does not expand US indications.

Components and intended use

CUVIS-joint combines a preoperative planning workstation and an electromechanical robotic tool. CT is reconstructed into a 3D model on which the team defines resection and component parameters. Optical navigation relates that model to anatomy in theatre.

FDA K252037, dated 25 March 2026, describes CJ150 preparation of the distal femur and proximal tibia in primary total knee arthroplasty and names the United U2 Knee System. It does not establish compatibility with all implants, US hip use or availability in every country.

Workflow and control

Following CT, segmentation and a preliminary plan are completed. In theatre, references are fixed, landmarks registered and correspondence verified. The surgeon assesses joint gaps and may revise the plan. The arm then positions the tool and carries out planned bone preparation.

Manufacturer terms such as “fully automatic” or “active” describe tool motion, not autonomous clinical judgement. The surgeon approves and initiates the step, supervises safety and can stop execution. The clinical team remains responsible for exposure, tissue protection, registration and an alternative strategy.

Typical sequence

1. Clinical indication and protocol CT.
2. Segmentation, 3D analysis and preliminary plan.
3. Tracker placement and anatomical registration.
4. Gap assessment and surgeon-led plan adjustment.
5. Controlled robotic-arm resections.
6. Surface and trial checks, then completion of surgery.

Regional context

CUREXO reported Korean MFDS approval of a CUVIS-joint THA application in 2025. This manufacturer announcement describes a regional extension. It is not evidence of THA approval in the United States, Europe or Russia, and knee evidence cannot be transferred to the hip application.

Evidence

Active resection without the autonomy myth

CUVIS-joint differs from guide-positioning systems because its arm moves the cutting tool along an approved trajectory. Describing this as “the robot operates by itself” omits essential dependencies. The clinical team creates and verifies the plan, registers the patient, establishes exposure, authorises movement, watches the working area and can stop the device.

Automation describes mechanics, not competence. The platform does not diagnose disease, select an indication, choose an implant outside an authorised configuration or assess the entire clinical context. Responsibility does not transfer from clinician to machine.

Data checkpoints

Before resection, the team must confirm patient and side, CT completeness, segmentation of bony surfaces and registration accuracy. Optical references must retain position and visibility. Intraoperative gap assessment may justify changing the preliminary plan; that judgement belongs to the surgeon.

After execution, real surfaces and parameters are checked rather than assumed correct because software completed a step. Current instructions must provide a stop and fallback pathway for technical error or uncertain registration.

Reading the studies

Cadaveric accuracy asks whether geometry can be reproduced in a laboratory model. A case series describes events observed in a clinical cohort. Neither design alone demonstrates faster recovery or fewer revisions. Those conclusions require comparable clinical groups, prespecified outcomes and adequate follow-up.

Excluding cases in which robotic completion was abandoned is especially relevant: it may understate problems across all initiated workflows. A transparent denominator is more informative than an isolated percentage.

A cadaveric study involving three specimens, six knees and three surgeons examined planned versus executed resection thickness and angles. Its very small laboratory sample cannot represent patient outcomes.

A consecutive series of 500 TKAs by one senior surgeon described intraoperative technical events and complications in completed robotic cases. Cases in which robotic execution was abandoned were excluded, and there was no control group or long follow-up. It cannot demonstrate overall safety superiority.

Limits and questions

The CT-based process adds imaging and depends on segmentation, registration and stable tracking. Compatibility is configuration-specific. Active arm movement requires supervision and immediate stop capability. Technical accuracy does not promise less pain, fewer complications or revisions.

Does it act independently? No. It executes a plan under surgeon direction and control.

Can it be used for hips? A Korean regional approval has been reported; that is not a universal indication.

Does it accept any implant? The FDA document names a specific compatibility; others need separate instructions.

This material does not replace individual assessment. Sources reviewed 8 August 2026; scheduled review August 2027. Author: RoboOrthoClub editorial team. Relevant conflicts must be disclosed.