Surgery from 2,800 km away passed its first randomised test, with 72 patients
A trial across five Chinese hospitals found remote robotic urological surgery non-inferior to local surgery. The sample is small, the surgeons were experts, and the network never failed.
Remote surgery has been demonstrated many times, but almost always as a case series: a surgeon performs a run of operations over a distance, they go well, and the report says so. A trial published in The BMJ on 28 January is the first to randomise patients between a remote surgeon and a local one and compare the results [s1].
The finding is that telesurgery was non-inferior. The more useful information is in how narrow the question was.
The design
Five hospitals in China enrolled patients scheduled for radical prostatectomy or partial nephrectomy between December 2023 and June 2024, randomising them 1:1 to telesurgery or local robotic surgery [s1]. Seventy-two patients were enrolled in the intention-to-treat set — median age 61.0 years (IQR 57.5 to 68.0) in the telesurgery group and 65.0 (56.5 to 70.0) in the local group [s1].
The primary outcome was not survival, or complications, or oncological clearance. It was the probability of surgical success, judged by the medical team against criteria established before the trial began [s1]. The pre-specified non-inferiority margin was an absolute reduction of 0.1 in that probability [s1].
Telesurgery met it. Accounting for clustering by surgeon, the difference in success probability was 0.02 (95% credible interval −0.03 to 0.15), with a Bayesian posterior probability of 0.99 for non-inferiority [s1]. Thirteen clinical secondary outcomes covering the operation and early recovery, plus a measure of medical team workload, did not differ substantially between the groups [s1]. Patients were followed up at four and six weeks [s1].
The engineering result is the real one
The trial also reported technical outcomes, and these are what a reader should take from it. The telesurgery system operated over distances from 1,000 km to 2,800 km with a mean round-trip network latency of 20.1 to 47.5 ms and frame loss of 0 to 1.5 per operation [s1]. The authors describe the system as stable [s1].
That is the claim being tested. A non-inferiority trial with 72 patients and a composite team-judged endpoint cannot tell you whether remote surgery is as safe as local surgery in any general sense. It can tell you that, on the network and hardware used, in these hands, over these distances, the link held and the operations went as planned.
Independent evidence from the same country points the same way. A report in the January issue of the Chinese Journal of Gastrointestinal Surgery describes 37 consecutive 5G remote robotic radical gastrectomies performed between September 2023 and March 2025, across campuses of one hospital and between that hospital and three others [s2]. All 37 were completed, with mean operative time of 204.5 (SD 43.8) minutes, mean intraoperative blood loss of 52.4 (SD 19.1) ml, mean 36.2 (SD 12.2) lymph nodes dissected, and no serious postoperative complications [s2]. Total system latency averaged 219.1 (SD 20.0) ms and round-trip network latency 33.5 (SD 9.4) ms [s2].
What is holding the safety case up
Neither report rests its safety argument on the network. Both rest it on what happens if the network fails.
The gastrectomy series describes the arrangement explicitly: a dedicated backup transmission line, a backup surgical console at the patient's site for immediate hot takeover, a second qualified robotic surgeon physically present at all times, an emergency stop that the local team could activate, and control that could be switched back to the local console within seconds [s2]. The local team held the highest safety responsibility and the highest authority over the robot, and decided whether to accept remote control requests at all [s2].
In other words, every remote operation described here was also a local operation with a second surgeon standing by. That is the correct way to run an early-stage remote surgery programme. It is also the reason these results say little about telesurgery as a way to deliver surgery where no local surgeon exists — which is the use case most often invoked to justify it.
Limits worth stating plainly
The BMJ trial's own scope is narrow: two urological procedures, five hospitals in one country, 72 patients, expert robotic surgeons, and a primary endpoint defined by the treating team rather than an independent adjudicator [s1]. The gastrectomy report is a single-group case series with no comparator [s2]. Neither addresses long-term oncological outcomes, cost, liability, or what happens the first time a link drops mid-dissection with no console at the bedside.
What to watch
Whether subsequent trials use patient-centred primary endpoints rather than team-judged procedural success, and whether any programme reports results from a site that does not have a qualified robotic surgeon in the room.
Sources
- [s1] Reliability of urological telesurgery compared with local surgery: multicentre randomised controlled trial. BMJ, 28 January 2026. https://doi.org/10.1136/bmj-2024-083588
- [s2] Application and experience of 5G-based telesurgery in robotic radical gastrectomy. Chinese Journal of Gastrointestinal Surgery, January 2026 issue. https://doi.org/10.3760/cma.j.cn441530-20250623-00237
Sources
- Reliability of urological telesurgery compared with local surgery: multicentre randomised controlled trial — BMJ , January 28, 2026
- Application and experience of 5G-based telesurgery in robotic radical gastrectomy — Chinese Journal of Gastrointestinal Surgery , January 1, 2026
More on
Two FDA-cleared AI tools read prostate biopsies. A review flags what could go wrong.
A review of Paige Prostate Detect and Ibex Prostate Detect finds the tools mainly help less-specialized pathologists, and warns performance shifts when a tool meets a new population.
Monitoring low-risk prostate cancer matched surgery on survival at 15 years
In a trial of 1,643 men, prostate cancer killed 3.1% of the monitored group and 2.2% of the surgery group — a difference the trial could not distinguish from chance. Metastases were twice as common.
Two June studies test AI in radiology — one for speed, one for accuracy
A draft-reporting model cut documentation time 15.5% across 23,960 radiographs without changing report quality. A separate reader study found AI assistance raised prostate MRI accuracy by 3.3 percentage points.
Screening guidelines disagree because the same trials support more than one answer
US guidance recommends mammography from 40. A Canadian review of the same randomised evidence put the benefit at 0.27 fewer breast cancer deaths per 1,000 women aged 40-49 over ten years.