Objectives: Biannual ultrasonography, a globally accepted surveillance method, has low sensitivity in detecting early-stage hepatocellular carcinoma (HCC). We aimed to investigate the effectiveness of a surveillance strategy using annual contrast-enhanced MRI to detect HCCs at early-stage.
Materials and methods: We reviewed the data of 294 patients with consistent annual contrast-enhanced MRI and biannual alpha fetoprotein (AFP) surveillance between 2008 and 2017. Patients were stratified for HCC risk as low-intermediate-high risk group using Toronto risk score. HCCs were classified according to Barcelona Clinic Liver Cancer staging system.
Results: Thirty-five (11.9%) HCCs were detected with annual surveillance MRI. Of those, 30 (85.8%) were early-stage and 15 (42.9%) were very early-stage. The majority of patients (82.9%) with surveillance detected HCC were high risk at the entry. MRI had sensitivity of 83.3 and 80% with a specificity of 95.4 and 91.4%, for detecting early and very early-stage HCC, respectively. Addition of AFP to MRI displayed similar sensitivity and specificity rates to detect early and very early HCCs. The area under the curve of MRI alone and combination with AFP was not statistically different (Any-HCC: 0.905 vs. 0.924; Early-HCC: 0.853 vs. 0.885; Very early-HCC: 0.838 vs. 0.885, respectively, all P values >0.2).
Conclusion: Annual MRI strategy demonstrated a satisfactory performance in the surveillance of HCC, in terms of detecting most of the lesions in earlier curable stages and indicating high sensitivity with no additional benefit of biannual AFP. New risk stratified screening algorithms may further increase the yield of HCC surveillance among cirrhotic patients.