Prognostic value of minimal residual disease profiling in resectable hepatocellular carcinoma.

530 Background: Detecting post-surgical residual disease is a critical clinical requirement in resectable hepatocellular carcinoma (HCC). Previous studies focused on specific genomic regions have indicated that ctDNA holds promise as a tool for prognostic assessment. Nevertheless, these methods exhibited limited sensitivity and failed to meet the minimal residual disease (MRD) threshold. Here we report the prognostic value of MRD profiling in HCC patients who have undergone hepatectomy. Methods: This retrospective study involved 88 HCC patients who underwent surgical resections in China, from January to May in 2016. During surgery, tumor and normal tissue specimens were collected. Plasma samples were obtained 7 days post-surgery. PredicineBEACON, a baseline tissue- or blood-informed MRD assay, was used for MRD profiling. This entailed identifying tumor-specific mutations through whole exon sequencing of tissue samples. Subsequently, a personalized MRD panel, selecting up to 50 mutations via bioinformatics pipelines merged with a fixed panel featuring 500 tumor actionable hotspots, was created for each patient. To detect MRD in post-surgery plasma samples, we employed ultra-deep sequencing at a coverage of 100,000X, utilizing the designed panel. Results: In the cohort of 88 patients, the distribution based on Barcelona Clinic Liver Cancer (BCLC) staging was as follows: 79.5% stage A (N=70), 12.5% stage B (N=11), and 8.0% stage C (N=7,). The MRD assay identified 36 patients as MRD+ and 52 patients as MRD-. We observed significant correlations between MRD status and both relapse-free survival (RFS) and overall survival (OS). For MRD+ patients, the median RFS was 17.1 months, while it was not reached for MRD- patients (p=0.0013). Likewise, the median OS for MRD+ patients was 40.1 months, in contrast to it not being reached for MRD- patients (p=0.0012). The MRD positivity rate stood at 100% in stage C patients, a significantly higher percentage compared to the rates observed in stage B (36.4%, p=0.0256) and stage A (36.2%, p=0.0042). Significantly, MRD status emerged as a pivotal prognostic differentiator among clinically low-risk patients in stage A: among MRD+ patients, the median RFS was 23.3 months, whereas it was not reached for MRD- patients (p=0.05). Additionally, the median OS of both MRD+ and MRD- was not reached but their survival curves showed significant differences (p=0.038). Conclusions: This study demonstrated the clinical utility of ctDNA MRD assay in patients with resectable HCC, as notably evident in the robust detection of MRD using plasma samples collected 7 days after surgery. Furthermore, the assessment of post-surgical MRD status provided valuable prognostic insights into both patient survival and the risk of disease relapse. ctDNA MRD status played a pivotal role as a prognostic differentiator, particularly among clinically low-risk patients.