Autonomous robotics for transcrestal sinus elevation: submillimetre trueness in a first cohort

Transcrestal sinus floor elevation is a delicate, largely blind procedure: the surgeon lifts the Schneiderian membrane through a narrow osteotomy with limited direct vision, and a perforation can compromise the graft. This study asked whether an autonomous dental implant robotic system can perform it with greater precision and safety. It is a single-center retrospective cohort of 11 patients and 12 implants treated between June 2024 and June 2025, with planned and postoperative cone-beam CT scans superimposed to measure accuracy.

The robot was accurate. Mean global deviation was 0.58 mm at the platform (coronal) level and 0.64 mm at the apex, with a mean angular deviation of just 1.19 degrees, all comfortably submillimetre and sub-two-degree. Critically, there were no Schneiderian membrane perforations and no other intraoperative or postoperative complications, confirmed by intraoperative endoscopic inspection and the Valsalva maneuver. The authors also compared flat versus sloped sinus floor morphologies, the sloped floor being the classically harder anatomy to instrument safely, and found no significant difference in accuracy between them.

The cohort is small and retrospective, so this is a proof of capability rather than definitive evidence. But the combination of submillimetre trueness and a clean safety record, even on sloped floors, is notable. It suggests autonomous robotics could standardize one of the more operator-dependent steps in posterior maxillary implantology, turning transcrestal sinus elevation into a more reproducible, minimally invasive protocol. Larger prospective studies will be needed before that promise is settled.