Resecting liver metastases from uncommon GI primaries: survival favours surgery early

Design

Systematic review and meta-analysis of 10 comparative observational studies; random-effects model

Population

Patients with metachronous liver metastases from non-colorectal, non-neuroendocrine GI cancers

Primary outcome

Overall survival at 1, 3 and 5 years

Effect

1-year survival relative risk 1.93 (1.26–2.98); 3- and 5-year trends non-significant

Hepatic resection is standard for colorectal and neuroendocrine liver metastases, but its role in metachronous liver metastases from other gastrointestinal primaries — non-colorectal, non-neuroendocrine cancers such as gastric, biliary or pancreatic — is unsettled, with no guidelines. This meta-analysis pooled ten comparative studies of surgery versus non-surgical management.

Surgically treated patients had longer median overall survival, with a significant advantage at one year and a favourable but non-significant trend at three and five years. Heterogeneity was high, and the surgical cohorts were almost certainly a selected, fitter group with resectable disease.

So this supports considering resection in a carefully selected patient at a centre with hepatobiliary expertise, framed as an individualised multidisciplinary decision — not as evidence that surgery itself confers the survival seen here. The selection effect is the central caveat.

• Overall survival favoured surgery at 1 year (relative risk 1.93, 95% CI 1.26–2.98).
• At 3 and 5 years the trend favoured surgery but was not significant (relative risk 2.45, 0.99–6.11; and 2.26, 0.97–5.29).
• Heterogeneity was substantial (I² 67–86%), and surgical cohorts were likely selected for fitness and resectability.
• Discuss resection for metachronous liver metastases from uncommon GI primaries in a multidisciplinary setting with hepatobiliary input.

Why it matters

It gives the first pooled signal for a decision usually made case by case without any guideline to lean on.

Don't overread it

This was observational with probable selection bias — fitter patients with resectable disease were operated on, so the survival gap partly reflects who was chosen, not surgery alone.

The statistics, in plain English

A relative risk of 1.93 for one-year survival means survivors were nearly twice as common after surgery, but the wide intervals at 3 and 5 years crossing 1.0 mean the longer-term benefit is unproven. High I² values (67–86%) show the studies disagreed a lot, so the pooled number is an average over quite different patients.

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.