Pancreatic cancer staging: liver MRI for small metastases, CT for vessels and resectability

Design

Systematic review and bivariate meta-analysis of 40 direct comparative diagnostic studies

Population

Patients with suspected pancreatic ductal adenocarcinoma who had at least two imaging tests

Primary outcome

Pooled sensitivity and specificity for detection, liver metastases, vascular invasion and resectability

Effect

Liver metastases: MRI sensitivity 0.89 (0.67 to 0.97) versus CT 0.60 (0.44 to 0.74)

A meta-analysis of 40 direct comparative studies, in which patients suspected of pancreatic ductal adenocarcinoma had at least two imaging tests, compared CT, MRI, endoscopic ultrasound and FDG-PET/CT against pathology where available.

For detecting the primary tumour, CT, MRI and endoscopic ultrasound did not differ significantly. For liver metastases, MRI was more accurate than CT, with sensitivity 0.89 (95% CI 0.67 to 0.97) versus 0.60 (0.44 to 0.74) and specificity 0.91 versus 0.84. CT outperformed MRI for vascular invasion. For resectability, CT had higher specificity (0.88, 0.76 to 0.94) than MRI (0.63, 0.48 to 0.76). Endoscopic ultrasound was more sensitive than CT or FDG-PET/CT for lymph node detection.

The authors suggest liver MRI before potentially curative surgery to exclude small metastases, with CT to look for other sites and exclude unresectable disease. The confidence intervals around the MRI liver figures are wide, so local practice and availability still matter.

• No significant difference between CT, MRI and endoscopic ultrasound for detecting the tumour.
• MRI was more sensitive than CT for liver metastases (0.89 versus 0.60).
• CT was better for vascular invasion and for specificity on resectability.
• Endoscopic ultrasound was more sensitive for lymph nodes.
• Consider liver MRI when CT shows a resectable tumour and curative surgery is planned.

Why it matters

It shows the two modalities answer different staging questions, so the best pathway is a combination.

Don't overread it

Pooled sensitivity from direct comparisons with wide intervals does not show that adding MRI changes survival, and it does not replace multidisciplinary staging.

The statistics, in plain English

A sensitivity of 0.60 means CT missed about four in ten liver metastases in these studies, versus about one in ten for MRI. The interval for MRI sensitivity (0.67 to 0.97) is wide, so the true advantage could be smaller. Pooled results from heterogeneous studies depend on how each defined its reference test.

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.