Comparison of four MRI protocols for detection of extrahepatic colorectal cancer metastases

Abstract

Purpose: To compare three magnetic resonance imaging (MRI) protocols containing diffusion-weighted imaging with background suppression (DWIBS) and one traditional protocol for detecting extrahepatic colorectal cancer metastases. Materials and Methods: Thirty patients with extrahepatic colorectal cancer metastases were scanned in three stations from the skull base to the upper thighs using a 1.5T MRI system with six different MRI sequences; transverse and coronal T2-weighted (T2W) turbo spin-echo (TSE), coronal short tau inversion recovery (STIR), 3D T1W TSE, DWIBS, and a contrast-enhanced T1W 3D gradient echo (GRE) sequence. The six sequences were used to build four hypothetical MRI interpretive sets which were read by two readers in consensus, blinded to prior imaging. Lesions were categorized into 13 anatomic regions. Fluorodeoxyglucose / positron emission tomography / computed tomography (FDG-PET/CT) read with full access to prior imaging and clinical records was used as the reference standard. Sensitivity, specificity, and false discovery rate (FDR) were calculated as appropriate and receiver operating characteristic (ROC) curves were constructed. Results: In all, 177 malignant lesions were detected by FDG-PET/CT and distributed in 92 out of 390 scanned anatomic regions. The sensitivity was statistically higher in two out of three sets incorporating DWIBS on a per-lesion basis (66.7%, 63.3%, and 66.7% vs. 57.6%) (P = 0.01, P = 0.11, and P = 0.01, respectively) and in all sets incorporating DWIBS on a per-region basis (75.0%, 75.0%, and 77.2 vs. 66.3%) (P = 0.04, P = 0.04, and P = 0.01, respectively). There was no difference in specificity, FDR, or AUCROC. There was no difference between sets containing DWIBS irrespective of the use of a contrast-enhanced sequence. Conclusion: MRI sets containing DWIBS had superior sensitivity. This sensitivity was retained when omitting a contrast-enhanced sequence. Level of Evidence: 1. Technical Efficacy: Stage 2. J. Magn. Reson. Imaging 2017;46:1619–1630.