Comparison of extracellular and hepatobiliary MR contrast agents for the diagnosis of small HCCs.

BACKGROUND & AIMS: The aim of this study was to compare the performance of MRIs with extracellular contrast agents (ECA-MRI) to HB contrast agents (HBA-MRI) for the non-invasive diagnosis of small HCCs in a head-to-head comparison.

METHODS: Between August 2014 and October 2017; 171 cirrhotic patients, each with 1 to 3 nodules measuring 1 to 3 cm, were included across eight centers. All patients had both an ECA-MRI and an HBA-MRI within a month. The non-invasive diagnosis of HCC was made when the nodule was hyper-enhanced at the arterial phase (HA) with wash-out at the portal phase (PP) and/or delayed phase (DP) for the ECA-MRI, or PP and/or HB phase (HBP) for the HBA-MRI. The gold standard was defined by a composite algorithm previously published.

RESULTS: 225 nodules, of which 153 were HCCs and 72 were not HCCs, were included. Both MRI sensitivities were similar (71.2% [63.4-78.3]). Specificity was 83.3% [72.7-91.1] for the ECA-MRI and 68.1% [56.0-78.6] for the HBA-MRI. With regard to HCCs, on ECA-MRI, 138 were HA, 84 had wash-out at PP and 104 at DP; on HBA-MRI, 128 were HA, 71 had wash-out at PP and 99 at HBP. For nodules from 2 to 3 cm, sensitivity and specificity were similar with 70.9% [57.1-82.4] and 75.0% [47.6-92.7] respectively. For nodules from 1 to 2 cm, specificity dropped to 66.1% [52.2-78.2] for the HBA-MRI vs. 85.7% [73.8-93.6] for the ECA-MRI.

CONCLUSIONS: HBA-MRI specificity is lower than ECA-MRI for diagnosing small HCCs on cirrhotic patients. These results raise the question of the proper use of HBA-MRI in algorithms for the non-invasive diagnosis of small HCCs.