BACKGROUND: Predictive biomarkers are needed to identify individuals with metabolic dysfunction-associated steatotic liver disease (MASLD), at high risk for HCC. Our study aimed to determine whether the detection of circulating bacterial DNA could be associated with HCC development in MASLD patients with liver cirrhosis. METHODS: We developed a multicenter prospective cohort of patients with cirrhosis undergoing surveillance for HCC by contrast-enhanced magnetic resonance imaging. In a nested cohort study, we performed 16S rRNA sequencing of cell-free DNA extracted from 343 longitudinal plasma samples collected from 151 MASLD patients with cirrhosis. Among the 151 patients, 25 developed HCC during follow-up. RESULTS: Following in silico decontamination approaches, variations in circulating bacterial DNA profiles were significantly associated with HCC development, cirrhosis severity, and male gender. Many of the identified taxa were well-known human-associated bacteria. Patients who developed HCC during follow-up showed an enrichment in Tsukamurella and Bacteroides, and depletion of Natronomonas. Associations with HCC development remained for the identified bacteria, after adjusting for cirrhosis severity and male gender. Acidobacteriota, Tsukamurella, and Staphylococcaceae showed markedly improved prediction of HCC within 12 months prior to diagnosis [Acidobacteriota: AOR=2.87 (1.36-6.04), p=0.006; Tsukamurella: AOR=2.80 (1.34-5.85), p=0.006; Staphylococcaceae: AOR=2.52 (1.19-5.36), p=0.016]. Circulating bacterial DNA profiles associated with male gender and cirrhosis severity were different from those observed for HCC and showed considerable overlap in significant taxa. These included enrichment of the lineage Gammaproteobacteria, Enterobacterales, and Rheinheimera. CONCLUSIONS: The identified circulating bacterial DNA signatures may have utility in personalized approaches to HCC surveillance in MASLD patients.