Neoadjuvant transhepatic arterial infusion chemotherapy (HAIC) with FOLFOX regime plus cadonilimab (PD-1/CTLA-4 bispecific antibody) for resectable multinodular CNLC Ib/IIa hepatocellular carcinoma (CAR_Hero study).

4139 Background: The recurrence rate of hepatocellular carcinoma (HCC) remains high, with multinodular HCC being a well-defined high-risk factor for recurrence. However, standardized neoadjuvant or adjuvant therapies for HCC have yet to be definitively established to effectively improve survival outcomes. Methods: In this ongoing single-center, phase 2, open-label, prospective cohort clinical trial, eligible pts were randomly assigned (1:1:1) to three arms (15 pts per arm). Neoadjuvant therapies included: (A) 2 cycles cadonilimab (6mg/kg Q2W); (B) once FOLFOX- HAIC and 2 cycles cadonilimab; (C) once FOLFOX-HAIC. Pts receive scheduled surgery on day 21-28 and postoperative adjuvant HAIC one month after surgery. The primary endpoints were major pathologic response (MPR, defined as ≤50% residual living tumor) and the 1-year recurrence-free survival (RFS) rate. Secondary endpoints included overall response rate (ORR, assessed per RECIST 1.1) and treatment-related adverse events (TRAEs). Additionally, a direct hepatectomy cohort was retrospectively collected as reference data. Results: A total of 42 pts were enrolled. Among them, 2 pts withdrew due to their desire to pursue conversion therapy. 38 pts underwent hepatectomy and were included in the efficacy analyses (A: 14pts, B: 14pts, C: 12pts). The median age was 55 years (range: 32-72), with 90.5% being male and 90.5% infected with hepatitis B virus. Arm B had the highest MPR rate of 78.6%, significantly higher than Arms A (35.7%) and C (20.0%) ( P = 0.011). Additionally, Arm B had the highest ORR (A: 14.3%; B: 40.0%; C: 8.3%), and lowest MVI detection rate (A: 50.0%; B: 21.4%; C: 40.0%). Focal heterogeneity was partially observed. The DCR was 100%. The most common TRAEs were elevated aspartate transaminase (64.3%) and alanine aminotransferase (59.5%). Grade 3-4 TRAEs occurred in 3 pts(hepatic dysfunction and erythema annulare). In Arms A and B, 4 pts experienced a delay in scheduled surgery by 2-4 weeks. The combination of HAIC and cadonilimab did not lead to a significant increase in TRAEs. After propensity score matching, the direct hepatectomy cohort was screened. Kaplan-Meier analysis revealed that the neoadjuvant cohort had a longer recurrence-free survival (RFS) time compared to the direct hepatectomy cohort (median RFS not reached vs. 24.7 months; P = 0.0048) and a lower MVI detection rate (36.8% vs. 52.6%). Conclusions: Neoadjuvant FOLFOX-HAIC combined with cadonilimab had a considerable antitumor activity, and a manageable safety for the resectable multinodular HCC. It brought the fewer MVIs of tumor and a better RFS. Clinical trial information: ChiCTR3000033692 .