First-line durvalumab with arterial chemotherapy in major portal invaded hepatocellular carcinoma: The phase 2 DurHope study with biomolecular analyses.

576 Background: Current clinical trials for advanced hepatocellular carcinoma (HCC) significantly underrepresent patients (pts) with Vp4 portal vein tumor thrombus (PVTT), a poor prognostic factor, including only 0–15% despite their real-world prevalence of 24–49%. Previous FOHAIC-1 study indicated the favorable efficacy of arterial FOLFOX chemotherapy in this population. Preclinical evidence suggested that oxaliplatin in the FOLFOX regimen can induce immunogenic cell death. The DurHope study is a biomolecular exploratory, phase 2 trial of first-line durvalumab combined with arterial FOLFOX in pts with Vp4 PVTT. Methods: This trial enrolled 30 HCC pts with Vp4 (n = 24, 80.0%) or Vp3 (n = 6, 20.0%) PVTT. Arterial FOLFOX was given on day 1–3, plus durvalumab (1120 mg) intravenously on day 7 (± 2) of each 21-day cycle up to 8 cycles. Maintenance durvalumab (1500 mg) was administered every 28 days thereafter until disease progression, unacceptable toxicity, or death. Primary endpoint was 1-year OS rate. Secondary endpoints included progression-free survival (PFS), objective response rate (ORR) per RECIST v1.1, and safety. Single-cell RNA sequencing (scRNA-seq) was performed on pre-treatment biopsies and peripheral blood mononuclear cells (PBMCs) collected at baseline and during treatment from responders (CR/PR) and nonresponders (SD/PD). Results: With a median follow-up of 21.6 months, the 1-year OS rate was 63.3%, median OS was 13.9 months (95% CI, 10.7–NR), and median PFS was 9.0 months (95% CI, 4.5–12.2). ORR was 53.3% (CR: 5/30 [16.7%]; PR: 11/30 [36.7%]), and median duration of response was 9.1 months (> 6 months: 11/16 [68.8%]; > 12 months: 4/16 [25%]). Grade ≥ 3 TRAEs occurred in 43.3% of patients, with no unexpected safety events beyond those of each agent. ScRNA-seq of biopsies revealed enrichment of chemotherapy resistance and immune escape signatures in non-responders, while responders exhibited higher immune cell infiltration and stronger immune-malignant cell interaction. Dynamic scRNA-seq of serial PBMCs demonstrated expanded T cell subsets and increased expression of cytotoxicity-related genes (e.g., GZMB, GNLY, and IFNG) and proliferation marker (MKI67) after treatment, which was more pronounced in responders. This was accompanied by decreased non-classical monocyte frequency and reduced M2-directed monocyte polarization in responders. In vivo mouse models further showed enhanced efficacy of the combination of chemotherapy and PD-L1 blockade, indicating the synergistic effects of this therapeutic approach. Conclusions: Durvalumab plus arterial FOLFOX showed promising efficacy and manageable toxicities in advanced HCC with Vp4 PVTT. Biomolecular exploratory analyses revealed enhanced cellular immune responses triggered by treatment, along with novel molecular biomarkers that may help predict therapeutic response. Clinical trial information: NCT04945720 .