Robotic anatomical hepatectomy: fewer severe complications, harder cases, longer operations

Design

exploratory post hoc analysis of a randomised controlled trial, no subgroup sample size calculation

Population

57 patients scheduled for minimally invasive anatomical hepatectomy (30 robotic, 27 laparoscopic) from 81 randomised

Primary outcome

severe postoperative complications, Clavien-Dindo grade III or above, within 90 days

Effect

13% vs 37% (OR 0.27, 95% CI 0.07-0.97); operating time 252 vs 190 min

The ROC'N'ROLL trial randomised patients to robotic or laparoscopic liver resection; this exploratory post hoc analysis took the 57 of 81 participants who had been scheduled for anatomical hepatectomy before randomisation — 30 robotic, 27 laparoscopic. The endpoint was severe postoperative complication, Clavien-Dindo grade III or above, within 90 days.

Severe complications occurred in 13% of robotic and 37% of laparoscopic cases (OR 0.27, 95% CI 0.07-0.97). That happened despite the robotic arm being the harder one: median tumour 59 against 40 mm and median IWATE difficulty score 10 against 7, both differences reaching conventional significance. Operating time was longer with the robot, 252 against 190 minutes. Conversion, blood loss, length of stay, functional recovery, margin status, mortality and patient-reported outcomes showed no clear separation.

The design limits what this can carry. It is a post hoc subgroup of a randomised trial with no subgroup-specific sample size calculation, the groups were imbalanced at baseline in the direction that should have disadvantaged the robot, and the complication interval reaches 0.97 — it only just clears no difference.

Read alongside the gastrectomy meta-analysis below, the emerging pattern is that the robot's advantage, where one exists, sits in technically difficult resections rather than across the board. For a unit deciding where to deploy limited robotic time, high-IWATE anatomical liver resection is a defensible answer; routine cases are not.

• Where robotic time is rationed, consider prioritising high-difficulty anatomical liver resections over straightforward ones
• Budget the extra theatre time — the robotic arm ran about an hour longer per case in this analysis
• Score difficulty formally (IWATE or equivalent) when comparing your own outcomes between approaches
• Do not expect gains in blood loss, conversion, margin status or length of stay; none were seen here
• Treat this as a reason to audit your own complication rate by approach, not as evidence to change consent wording

Why it matters

It suggests the robot earns its cost in the hard cases rather than the routine list, which is the opposite of how most units introduce it.

Don't overread it

This is an exploratory post hoc subgroup with baseline imbalance and no subgroup power calculation — it cannot establish a complication benefit.

The statistics, in plain English

An odds ratio of 0.27 with an interval from 0.07 to 0.97 only just excludes 1.0, so the finding is statistically significant but fragile — one or two more complications in the robotic arm would have erased it. With 30 and 27 patients, a difference of 13% against 37% represents roughly four events against ten, which is why the interval is so wide. Because this was a post hoc subgroup with no sample size calculation of its own, the usual protection against a chance finding does not apply; the P value should be read as descriptive.

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.