Should MRI Replace CT Scan for the Evaluation of Colorectal and Neuroendocrine Liver Metastases?

Introduction: When staging metastatic colorectal or neuroendocrine liver metastases (CRLM, NLM), the use of CT or MRI to evaluate the liver can affect the perceived tumor burden and possibly alter surgical planning. We performed a comparison of the two modalities to determine which study is superior in detecting liver lesions. Methods: A prospectively maintained database was queried for patients with CRLM or NLM who underwent contemporaneous CT and MRI. Two blinded radiologists (RADS1 & RADS2) independently evaluated the images for the number of liver lesions and their sizes. To minimize recall bias, studies were grouped by modality, randomized, and evaluated separately. Results: Our query yielded 91 patients (46 CRLM; 45 NLM). The median age was 59 [interquartile range 51–66]. Low inter-rater variability was observed with intraclass correlation coefficients of 0.902 for CT and 0.953 for MRI. In the CRLM cohort, there was no difference in the number of lesions between CT and MRI (RADS1: 6 vs. 7, P=0.187; RADS2: 7 vs. 7, P=0.360) or the sizes of the smallest lesion (P=0.850). Conversely, in the NLM group, more lesions were seen on MRI than CT (14 vs. 12 lesions, P=0.001). In addition, Eovist-enhanced MRI allowed detection of smaller lesions compared to CT (RADS1: 4mm vs. 6mm, P=0.001; RADS2: 3mm vs. 5mm, P=0.003). Conclusions: For CRLM, CT and MRI appear to be equivalent in quantifying liver lesions. For NLM, our data suggest that more, and smaller, lesions are detected on MRI than CT which may alter treatment planning and surgical approach.