- Design
- Phase 3, multicentre randomised clinical trial
- Population
- 218 patients with resectable oesophageal or junctional adenocarcinoma at 4 high-volume centres
- Primary outcome
- Number of resected lymph nodes
- Effect
- Median 36 vs 32 nodes (P=.005); 90-day mortality 1.0% vs 3.0%
A phase 3 randomised trial at four high-volume centres in Germany, the Netherlands and Switzerland (JAMA Surgery, 30 September 2026) assigned 218 patients with resectable oesophageal or junctional adenocarcinoma to robot-assisted minimally invasive oesophagectomy (RAMIE) or conventional minimally invasive oesophagectomy (MIE); 202 had transthoracic resection, 101 in each arm.
The robotic arm removed more lymph nodes (median 36 vs 32, P=.005), the primary endpoint. Luminal R0 rates were 99.0% and 97.0%. Complications were similar, 90-day mortality was 1.0% versus 3.0% (P=.62), and one-year survival did not differ.
This is the first randomised evidence that the robot changes the oncological operation, not just the surgeon's comfort. Node yield is a surrogate, though: whether four extra nodes translate into better staging or survival will depend on the five-year follow-up. The trial was run in centres already expert in both approaches.
For units with robotic access and experience, this supports RAMIE as a reasonable standard for adenocarcinoma. For units without, it is not a reason to abandon a well-run MIE programme, and the learning curve has its own cost.
- Consider RAMIE for oesophageal adenocarcinoma where robotic expertise and volume already exist
- Do not treat node yield as a survival benefit until long-term results report
- Keep a well-established MIE programme rather than switching without volume and training
- Report node yield and R0 rates in local audit whichever approach you use
Why it matters
It is randomised evidence that the robot can change the oncological quality of the operation, not just its ergonomics.
Don't overread it
A surrogate primary endpoint in expert centres - no survival benefit has been shown yet.
The statistics, in plain English
A median of 36 versus 32 nodes is a clear difference, but node count is a surrogate endpoint. The 90-day mortality difference (1% vs 3%) has P=.62, so it may be chance. One-year survival tells us nothing yet about cure.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for general surgery, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free