Utilizing circulating tumor DNA methylation levels to predict pathological and radiological response in hepatocellular carcinoma patients undergoing neoadjuvant and systemic therapy.

e16323 Background: Systemic therapy of hepatocellular carcinoma (HCC) faces several challenges, including tumor heterogeneity and drug resistance. Previous studies have indicated that circulating tumor DNA (ctDNA) may offer valuable insights for predicting the efficacy of systemic therapy in various cancers. Nevertheless, the majority of these studies are predominantly based on mutation detection. Recently, the ctDNA methylation detection has been used for real-time monitoring of tumor dynamic changes and assessment of treatment response. However, its applicability in HCC remains unclear. Methods: For HCC patients undergoing neoadjuvant therapy (immune checkpoint inhibitor [ICI] based), we collected paired plasma samples (n = 54) at baseline (Tb) and prior to surgery (Ts). Serial plasma samples (n = 28) were obtained from HCC patients receiving systemic therapy (ICI based) at baseline (C1) and 3 (C2), 6 (C3), and 9 (C4) weeks following treatment initiation. All samples were tested using the GutSeer panel and the cancer-specific methylation score (CSMS) were calculated. The ctDNA methylation response (CMR) was defined as the difference of CSMS between Tb and Ts greater than 0.01. The concordance between CMR and pathological response (major pathological response [MPR], ≤50% residual viable tumor) and radiological response (assessed by the modified Response Evaluation Criteria in Solid Tumors) was evaluated. Results: We enrolled 27 HCC patients undergoing neoadjuvant therapy. All patients completed at least two courses of ICI and the median treatment-to-surgery interval was 1.7 months. The difference of CSMS (Tb-Ts) was significantly negatively correlated with the proportion of residual viable tumor (R = -0.7, P < 0.001). Among them, 17 patients (63.0%) achieved CMR, of whom 15 (88.2%) were assessed as MPR. In contrast, only 2 (20%) MPR were observed among the remaining 10 patients without CMR. The area under the curve and accuracy of the predictive model based on CSMS were 85.9% and 85.2%, respectively. Additionally, we included 7 HCC patients receiving systemic therapy. Among them, 2 were assessed as partial response with one patient demonstrating a significant reduction in CSMS (~20% decrease from C1 to C4) in advance. As for the 5 patients classified as stable disease or progressive disease, 4 exhibited an early increase in CSMS, including one patient with a rise of ~80% (from C1 to C4). Conclusions: Our study found that ctDNA methylation detection is a promising indicator for monitoring tumor dynamic changes, enabling the prediction of pathological and radiological response in HCC patients during treatment. These discoveries pave the way for the development of effective efficacy surveillance strategies, thereby facilitating individualized management and enhancing the precision of medical interventions.