Meta-analysis of MRI for the diagnosis of liver metastasis in patients with pancreatic adenocarcinoma.

Abstract

Determining if a given patient with primary pancreatic adenocarcinoma (PDA) has hepatic metastasis is important for treatment planning. Several previous studies reported on the diagnostic performance of MRI for liver metastasis in patients with PDA. But the reported data are quite variable. To systematically determine the diagnostic performance of MRI for liver metastasis in patients with PDA, including comparing it with computed tomography (CT). Systemic review and meta-analysis. In all, 457 patients from five eligible articles. Conventional MR sequences with or without contrast enhancement at 1.5T and 3T. Two reviewers independently performed the data extraction. The reviewers identified and reviewed the original articles reporting the diagnostic performance of MRI for liver metastases in patients with PDA, including those articles making comparisons with CT. Meta-analytic summary sensitivity and specificity were calculated on a per-patient basis using a bivariate random effects model. We compared the meta-analytic summary sensitivity and specificity between MRI and CT. The meta-analytic summary sensitivity and specificity were 85% (95% confidence interval [CI], 74-92%; I2 = 0%) and 98% (95% CI, 78-100%; I2 = 85%), respectively. In comparison with CT, MRI showed a higher sensitivity (85% vs. 75%) but similar specificity (98% vs. 94%). MRI had good overall diagnostic performance for liver metastasis in patients with PDA, with a higher sensitivity than CT. 3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:1737-1744.