MRI-based risk factors of hepatocellular carcinoma in patients with chronic liver disease: A prospective observational study

J Magn Reson Imaging. 2020 Feb;51(2):389-396. doi: 10.1002/jmri.26797. Epub 2019 May 23.


Background: MR-based metrics, including hepatobiliary phase (HBP) hypointense nodules without arterial phase hyperenhancement (APHE), and liver stiffness as measured by MR elastography are useful markers to stratify the risk of hepatocellular carcinoma (HCC) development in chronic liver disease patients. However, prospective studies are needed to clarify their utility.

Purpose: To perform a risk analysis of HCC development in chronic liver disease patients, with a focus on MR-based biomarkers.

Study type: Prospective.

Subjects: Consecutive 110 cirrhotic patients (61 males, 49 females) without a history of HCC who matched the inclusion criteria.

Field strength/sequence: 3T/3D gradient-echo T1 -weighted images and MR elastography.

Assessment: Patients underwent MRI for HCC screening and attended follow-up appointments every 3 months. The primary endpoint was the development of hypervascular HCC. Patients were classified according to the presence of an HBP hypointense nodule without APHE (if present in the liver, the patient was classified in nonclean liver group; if absent, clean liver group), and stiffness value on MR elastography (soft liver, <4.0 kPa; stiff liver, ≥4.0 kPa) at the initial examination.

Statistical tests: Risk factors were identified in univariate and multivariate Cox regression models. Incidence rates of HCC development were evaluated using the Kaplan-Meier method.

Results: Patients were classified into clean-liver (n = 76) and nonclean-liver groups (n = 34), and into soft-liver (n = 53) and stiff-liver groups (n = 45). During the follow-up period (median, 21.0 months), 16 patients developed hypervascular HCC. Patients in the nonclean-liver group showed a higher incidence of hypervascular HCC than those in the clean-liver group (3-year HCC incidence rates: 50.4% and 5.7%, respectively; P < 0.05). A nonclean liver was independently associated with hypervascular HCC development (hazard ratio, 18.75; 95% confidence interval, 4.83-128.63; P < 0.0001), but stiff liver was not (1.91; 0.66-6.23; P = 0.23).

Data conclusion: An HBP hypointense nodule without APHE observed on a gadoxetic acid-enhanced MR image is a strong indicator of subsequent development of hypervascular HCC in patients with chronic liver disease.

Level of evidence: 2 Technical Efficacy Stage: 5 J. Magn. Reson. Imaging 2020;51:389-396.

Keywords: gadoxetic acid; hepatocellular carcinoma; liver; magnetic resonance elastography; magnetic resonance imaging; risk analysis.