Hepatocellular carcinoma (HCC) is the fifth most common cancer and the third leading cause of cancer-related mortality worldwide, accounting for approximately 700,000 deaths annually (1). It is a troublesome tumor with poor prognosis because of frequent late diagnoses with advanced stage, limiting the potential for effective locoregional therapies such as hepatic resection, radiofrequency ablation (RFA) or transarterial chemoembolization (TACE), etc. Hence, systemic therapy is the main therapeutic modality for advanced HCC (2). Sorafenib, the first approved agent for systemic therapy of advanced HCC, is a multi-targeted tyrosine kinase inhibitor (mTKI) that can target several protein receptors such as vascular endothelial growth factor receptor (VEGFR), or platelet-derived growth factor receptor (PDGFR) to impair vascular angiogenesis as well as block several cell signaling pathways such as Raf-1, B-Raf, and kinase activity in the Ras/Raf/MEK/ERK signaling pathways to inhibit tumor proliferation (3). Approval of sorafenib is according to two randomized, double-blind, phase III clinical trials, where sorafenib significantly improved overall survival (OS) in patients with advanced HCC compared with placebo (4, 5); however, the progression of other first-line or following second-line systemic therapies for advanced HCC was disappointing, until 2017, when two second-line agents, regorafenib and nivolumab, have since demonstrated their therapeutic effectiveness for the treatment of HCC (6, 7). The phase III RESORCE study demonstrated that sequential administration of sorafenib followed by that of regorafenib extended patient survival (median survival time: 26.0 months for sorafenib–regorafenib vs. 19.6 months for sorafenib–placebo) (6). In addition, regorafenib also prolonged progression-free survival (PFS) compared with placebo (3.1 months vs 1.5 months, p<0.001). The recent development of cancer immunotherapies using immune checkpoint inhibitors (ICIs) targeting cytotoxic T-lymphocyte-associated protein-4 (CTLA-4) and anti-programmed cell death protein-1 (PD-1) has dramatically changed the landscape of cancer therapy and prolonged the survival of patients with different malignancies (8). Studies evaluating anti-PD-1/PD-L1 monoclonal antibodies as single agents in pre-treated patients with advanced HCC showed encouraging results (9, 10). Indeed, the blockage of PD-1/PD-L1 expression and tumor-infiltrating lymphocytes (TIL) in patients with HCC leads to the discontinuation of immunosuppressive effect by the cancer cells and therefore reactivates cytotoxic T-cells to identify and eradicate the cancer cells (11). Nivolumab was the first second-line treatment for patients with advanced HCC to be approved, and based on the phase I/II Checkmate 040 study (7), the objective response rate (ORR) was 20%, the disease control rate (DCR) was 64%, and PFS was 4.1 months for patients after sorafenib failure. Although regorafenib and nivolumab both showed significant therapeutic efficacy compared with placebo, which systemic therapy should be applied following sorafenib for patients with advanced HCC was still a critical issue in real clinical practice. Consequently, this study aimed to appraise therapeutic efficacy and safety of two second-line therapies, regorafenib and nivolumab, for patients with advanced HCC after sorafenib failure.

Nivolumab, an immune checkpoint inhibitor that blocks PD-1, has recently become one of the effective treatment options for many malignancies, including non-small-cell lung cancer, melanoma, renal cell carcinoma, HCC, and the like (7, 13–15). Although there was no significant difference of OS in the CheckMate-459 trial, being the phase III study of nivolumab versus sorafenib as first-line systemic therapy in advanced HCC, nivolumab still demonstrated survival benefit for patients with radiologic response (16). In addition, in the previous CheckMate-040 trial, a phase I/II uncontrolled trial, nivolumab showed durable treatment response and prolonged long-term survival for patients in advanced stage after sorafenib failure (7). Among the total population of 182 patients, there was a promising ORR of 14%, a CR of 3%, and a DCR of 55%, with a 12-month survival rate of 55%. The current study also had a treatment response of nivolumab with an ORR of 16% and a DCR of 44%. In addition, median OS of Nivolumab in the current study was 14 months, equivalent to that of the CheckMate-040 trial (7). It seems that nivolumab monotherapy shows a promising treatment response and an improved survival outcome as second-line therapy for patients with advanced HCC, regardless of being in a clinical trial or in clinical real-world settings.

Regorafenib is an orally administered TKI that is structurally similar to sorafenib but with additional blockage of fibroblast growth factor receptor pathway (17, 18). Therapeutic efficacy of regorafenib has been approved by the phase III RESCORCE trial (6) that included 573 patients with advanced HCC and Child-Pugh class A who tolerated sorafenib but with tumor progression. Under the 2:1 randomized assignment to regroafenib or placebo, regorafenib significantly improved median OS compared with placebo (10.6 months vs 7.8 months; HR: 0.62, P<0.0001). Moreover, regorafenib also had prolonged. TTP (3.1 months vs 1.5 months, p < 0.0001), improved ORR (10.6% vs 4.1%, p=0.0047) and DCR (65.2% vs 36.1%, p<0.0001) in comparison with placebo. Therefore, current international HCC treatment guidelines identify regorafenib as the standard of care for HCC patients with advanced stage who have tolerated sorafenib but progressed. In the current study, treatment response of regorafenib with an ORR of 6.4% and a DCR of 31.9% was inferior to that of the RESORCE trial, but our median OS was equivalent (11 months vs 10.6 months).

In recent years, various systemic therapeutic options have been approved, so it is a critical issue for clinicians to decide on what is an appropriate second-line systemic treatment option after sorafenib failure? The current study compared the efficacy of nivolumab and regorafenib, the most frequently used ICI and TKI, for HCC patients with advanced stages where sorafenib treatment failed. In a mathematical Markov model reported by Cabibbo et al. that simulated treatment effect of sequential systemic therapies among patients with advanced HCC based on data of clinical trials, the simulated estimates of median OS were significantly higher for sofafenib followed by nivolumab compared to sorafenib followed by regorafenib (27 months vs 18 months) (19).

In the current study, we found that using nivolumab had a trend of better ORR and DCR than using regorafenib, but there was no statistical difference. Furthermore, median TTP and OS were not significantly different between the two groups in Kaplan-Meier survival analysis (TTP: 3 months vs 2.6 months, p = 0.786; OS: 14 months vs 11 months, p = 0.763 for the Nivolumab group vs Regorafenib group). Our finding is compatible to two previous Korean studies. Lee et al. reported that for 102 and 48 patients who were treated with nivolumab and regorafenib respectively, mOS was 5.9 and 6.9 months respectively (P = 0.77) (20). There was no obvious difference in DCR between nivolumab and regorafenib groups (50.0% vs. 47.1%; P =0.58). Another larger-sized study including 223 advanced HCC patients treated with regorafenib and 150 patients treated with nivolumab indicated that PFS (7.1 weeks for Nivolumab group vs 12 weeks for Regoranib group; P = 0.150), TTP (7.9 weeks vs 12.1 weeks; P = 0.680), and OS (32.6 weeks vs 30.9 weeks; P = 0.154) did not differ significantly between patients with nivolumab or regorafenib (21); however, the ORR was significantly higher in the Nivolumab vis-à-vis the Regorafenib group (13.3% vs. 4.0%; P = 0.002). It seems that the Nivolumab group might have superior treatment response, but clinical treatment outcomes such as TTP or OS might not be different.

The current study found that liver function reserve, achieved disease control and afforded post-treatment were independent factors associated with mortality for patients with advanced HCC receiving second-line treatment after sorafenib failure in multivariate analysis. However, using nivolumab or regorafenib was not related to overall survival, no matter for either univariate or multivariate analysis. Lee et al. reported that nivolumab was associated with prolonged OS (vs. regorafenib: HR, 0.54; 95% CI, 0.30–0.96; P =0.04) (20). However, in that study, OS did not differ in either group according to Kaplan Meier survival analysis (5.9 months in Nivolumab group vs 6.9 months in Regorafeib group, p=0.77) and univariate Cox regression analysis (Nivolumab (vs. regorafenib); HR: 1.081 (95%CI: 0.644–1.813) P=0.77). Choi et al. also reported that OS was consistent between these two groups by multivariable-adjusted, propensity score-matched and inverse probability treatment-weighted (IPTW) analyses (21). After cessation of regorafenib or nivolumab, the current study found 37.8% of patients (34.4% of Nivolumab group/39.7% of Regorafenib group) could afford following therapies. Patients with post-treatment had significantly superior median OS than those without (17.1 months vs 5.4 months, p<0.001), meaning that more than one-third of patients could maintain good liver function reserve and adequate performance status after experiencing nivolumab or regorafenib following sorafenib therapy. Regarding the impact of HCC etiology on treatment outcome, the current study found that hepatitis status was not associated with overall survival. Previous sub-analysis of CheckMate 040 study has indicated that the median OS of nivolumab in HCC patients with HBV or HCV was similar (22). In our Nivolumab group, patients with HBV or HCV also had insignificant median OS (6.3 months vs 10.3 months, p=0.885). Regarding the Regorafenib group, median OS from beginning sorafenib was almost statistically significant in HBV-related HCC or HCV-related HCC (18.9 months vs 14.2 months, p=0.051). This differs from the meta-analysis indicating that there is improved OS for patients negative for HBV and positive for HCV when treated with sorafenib (23).

Despite the fact that there was no OS difference between the Regorafenib group and the Nivolumab group, the two regimens really increased survival benefits for those patients who failed sorafenib treatment. Our previous study reported that the median OS of sorafenib use was only 8 months in the era of no effective sequential systemic therapies offered (24). The current study indicated overall OS from the beginning of sorafenib use was extended to 17.3 months in the Regorafenib group and 21.9 months in the Nivolumab group respectively. Consequently, application of regorafenib or nivolumab is approved as a potential second-line therapy followed sorafenib in clinical practice. Further well-designed prospectively randomized clinical trials are required to determine when and how to use regorafenib or nivolumab following sorafenib treatment for patients with unresectable HCC. Moreover, since resistance to targeted or ICI-based therapeutic drugs remains one of the main challenges for HCC treatment, other mechanisms blocking HCC cells such as poly (ADP-ribose) polymerase (PARP) inhibitor might be the future study frontier (25).

Concerning treatment safety, the current study found that the Regorafenib group had significantly higher proportions of TRAE than the Nivolumab group (68% vs 37.5%, p=0.006). The safety of regorafenib in the current study was demonstrated to be consistent with its safety profile in previous studies (6, 21, 23), with the leading four adverse events being HFSR (23.8%), diarrhea (11.9%), fatigue (10.2%) and elevated ALT (10.2%). Moreover, six patients (10.2%) had severe TRAE requiring permanent cessation of regorafenib, including one with HFSR, two with fatigue and three with hyperbilirubinemia. Compared with Regorafenib, the Nivolumab group had lower incidence of TRAE during treatment, with 37.5% including fatigue (12.1%), dermatitis (9.3%) and hyperbilirubinemia (6.2%). Only two patients (6.2%) had severe TRAE over grade 3, and both presented with hyperbilirubinemia. Although using regorafenib has more TRAE and poorer life quality than using nivolumab, the non-invasive oral-administered route and obviously cheaper price appear as advantages of regorafenib over nivolumab in real world consideration.

There are some limitations in the current study. Firstly, this was a retrospective study so that some biochemical and clinical data were not available at medical chart review. Approximately 20% of the patients that lacked image examinations following treatment might generate deviated assessments of treatment response. Secondly, in clinical real practice, baseline characteristics of Nivolumab and Regorafenib groups including liver function reserve and tumor pattern were not consistent, which might lead to confounding bias in the analysis. Thirdly, due to the small sample size of the enrolled patients, the analysis of TTP or OS might be affected by extreme values. Further large sample-sized studies are required to reduce these possible statistical biases.

In clinical practice, for patients with advanced HCC who failed sorafenib treatment, there was optimal survival benefit no matter whether using nivolumab or regorafenib as the second-line therapy. The Nivolumab group seemed to have lower TRAE incidence and a trend of better tumor response compared with the Regorafenib group; however, their TTP and OS did not differ significantly.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Kaohsiung Chang Gung Memorial Hospital (IRB No: 202100227B0). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

Y-HK and J-HW made substantial contributions to the study conception, design, analysis, and interpretation of the data. Y-HY, Y-YC, K-MK, C-HH, S-NL, T-HH, and C-HC contributed to the acquisition of the data. The first draft of the manuscript was written by Y-HK and J-HW. J-HW commented on subsequent versions of the manuscript. All authors approved the final manuscript submitted to the journal.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.