Direct antiviral agents (DAAs) have excellent efficacy against chronic hepatitis C virus (HCV) infection. Despite this strength, recent studies raised concerns about an unexpected hepatocellular carcinoma (HCC) occurrence rate after DAA therapy. In this exploratory case-control study, we evaluated the potential use of miRNAs as serum biomarkers for the detection of early HCC in DAA-treated patients. In the discovery phase, the circulating miRNome was assessed in 10 matched patients with (HCC+) or without HCC (HCC−) occurrence. Microarray analysis was performed before (T0) and after one month of the DAA therapy (T1). MiRNAs discriminating HCC+ and HCC− patients were validated in 60 samples by means of RT-qPCR. We estimated the time-averaged difference of a given miRNA between HCC+ and HCC− patients using a bootstrapped random-effect generalized least square regression model (RE-GLS). At T0, miR-1207-5p, miR-1275, miR-3197, miR-4443, miR-3178, miR-483-5p, miR-4706, miR-4793-3p and miR-1246 discriminated HCC+ from HCC− patients (p < 0.05). At T1, only miR-1180-3p, miR-1228-3p, miR-4329 and miR-4484 (p < 0.05) discriminated HCC+ from HCC− patients. The subsequent validation phase identified miR-3197 as changing with both disease and time. Our results suggest that patients might be already committed to HCC occurrence before DAA therapy. MiR-3197 shows some potential for the identification of patients at risk of HCC during DAA treatments.