Diagnostic performance of MDCT, PET/CT and gadoxetic acid (Primovist(®))-enhanced MRI in patients with colorectal liver metastases being considered for hepatic resection: initial experience in a single centre.

PURPOSE: To compare the relative diagnostic performance of MDCT, PET/CT and Primovist-enhanced MRI (P-MRI) in the pre-resection work-up of colorectal cancer (CRC) liver metastases. METHOD AND MATERIALS: This was a retrospective study of consecutive referrals for CRC liver metastases. All patients had MDCT, PET/CT and P-MRI examinations within 3 months of each other. They were divided into 2 groups: resected and unresected. Patients in the resected group underwent liver resection within 3 months of the imaging studies. In the unresected group, patients were unresectable by imaging criteria or are awaiting surgery. Standard of reference (SOR) was intra-operative ultrasound findings and pathology for the resected group. Intermodality comparison was the SOR for the unresected group. Number of lesions identified by each imaging modality for each patient was recorded. Sensitivity (95% CI) and PPV were calculated for each imaging modality in the resected group. RESULTS: There were 19 patients in the resected group and 11 patients in the unresected group. The sensitivity (96%) and PPV (0.91) of P-MRI were both superior to that of MDCT (P = 0.0009) and PET/CT (P = 0.0003). Intermodality comparison showed that P-MRI detected more lesions than MDCT and PET/CT. CONCLUSION: The sensitivity and PPV of P-MRI was superior to that of MDCT and PET/CT. P-MRI probably has the most added value if used after MDCT and PET/CT in patients still considered eligible for liver resection.