Robotic oesophagectomy took more lymph nodes than thoracolaparoscopic, with no safety cost

Design

Phase 3, multicentre, randomised clinical trial

Population

202 patients with resectable oesophageal or junctional adenocarcinoma (cT1-4a, N0-3, M0) at 4 European centres

Primary outcome

Number of resected lymph nodes

Effect

Median 36 robotic vs 32 thoracolaparoscopic (P = .005); 90-day mortality 1.0% vs 3.0% (P = .62)

A phase 3 randomised trial in JAMA Surgery (30 September 2026) compared robot-assisted minimally invasive oesophagectomy with conventional thoracolaparoscopic oesophagectomy for resectable adenocarcinoma of the oesophagus or oesophagogastric junction. Four high-volume centres in Germany, the Netherlands and Switzerland randomised 218 patients from 2021 to 2025; 202 had transthoracic resection.

The primary end point was the number of nodes removed. The robotic arm retrieved a median of 36 against 32 (P = .005). Luminal R0 resection was 99% and 97%. Complications were similar, 90-day mortality was 1% and 3%, and there was no difference in survival at one year.

Node count is a surrogate. It is linked to staging accuracy and, in observational data, to survival, but this trial has not yet shown that four more nodes mean longer life; that needs the planned five-year follow-up. What it does show is that a robotic approach in experienced hands is at least as safe and gives a more thorough lymphadenectomy.

For units choosing how to invest, it supports robotic oesophagectomy as a reasonable standard where the platform and volume exist. It is not a reason to move patients from an experienced thoracolaparoscopic team to an inexperienced robotic one.

• Report and audit lymph node yield for every oesophagectomy, whatever the approach.
• Where a robotic platform and volume exist, robotic oesophagectomy is a reasonable standard for adenocarcinoma.
• Do not move patients from an experienced minimally invasive team to a robotic programme still on its learning curve.
• Tell patients the survival benefit of the extra nodes is not yet known.

Why it matters

It is randomised evidence for an approach that has spread mostly on enthusiasm and cohort data.

Don't overread it

Lymph node count is a surrogate, and one-year survival did not differ.

The statistics, in plain English

A median difference of four nodes was statistically significant, meaning unlikely to be chance. Whether it is clinically significant depends on whether those nodes change staging or survival, which this report cannot yet answer. Mortality of 1% against 3% sounds different but rests on one versus three deaths, and the P value of .62 shows it could easily be chance.

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.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.