Early experience of atezolizumab plus bevacizumab therapy in Japanese patients with unresectable hepatocellular carcinoma in real-world practice

Background. We aimed to investigate the efficacy and safety of atezolizumab plus bevacizumab therapy in patients with unresectable hepatocellular carcinoma (u-HCC) based on whether they had previously received systemic therapy, as well as the association of atezolizumab plus bevacizumab with early alpha-fetoprotein (AFP) response in real-world practice. Methods. A total of 52 patients with u-HCC were treated with atezolizumab plus bevacizumab between October 2020 and April 2021. The Response Evaluation Criteria in Solid Tumors (RECIST) and modified RECIST were used to evaluate radiological responses. Results. The patients received atezolizumab plus bevacizumab as 1st-line (n = 23), 2nd-line (n = 16), 3rd-line (n = 6), 4th-line (n = 3), 5th-line (n = 3), or 6th-line (n = 1) therapy. According to RECIST, the objective response rate (ORR) and disease control rate (DCR) in all patients were 15.4% and 57.7%. In the 1st-line patients, ORR and DCR based on RECIST 1.1 were 27.3% and 81.8%. The median time to progression (TTP) assessed by RECIST was significantly longer among patients receiving atezolizumab plus bevacizumab as 1st-line therapy than in patients receiving atezolizumab plus bevacizumab as later-line therapy (P < 0.001). Patients with an AFP response (reduction ≥ 20% from baseline) at 6 weeks had a significantly longer TTP assessed by RECIST than those without an AFP response (P = 0.02). Conclusion. Patients who received atezolizumab plus bevacizumab as 1st-line therapy had better clinical outcome than those who received atezolizumab plus bevacizumab in later lines. The AFP response at 6 weeks could be a predictor of disease progression.


Introduction
Hepatocellular carcinoma (HCC) is the most common primary liver cancer and the fourth most common cause of cancer-related deaths worldwide [1]. HCC is hypervascular, and such angiogenesis is driven by growth factors, such as vascular endothelial growth factor (VEGF). Tyrosine kinase inhibitors targeting VEGF signaling, such as sorafenib [2,3], regorafenib [4], and lenvatinib [5], have already been approved for the treatment of unresectable HCC (u-HCC) in many countries. Cabozantinib [6] and ramucirumab [7,8] have been approved as 2nd-or later-line agents for u-HCC in Japan. Ramucirumab was approved for patients with a high alpha-fetoprotein (AFP) level (≥ 400 ng/mL) in July 2019.
VEGF is also associated with cancer immune evasion. Based on this rationale, the combination of programmed cell Yuka Hayakawa and Kaoru Tsuchiya these two authors contributed equally to this work death-ligand 1 (PD-L1) (atezolizumab) and VEGF (bevacizumab) pathway inhibition is believed to be the most effective therapy for advanced HCC. In September 2020, atezolizumab plus bevacizumab [9] was approved for patients with u-HCC in Japan and became the recommended 1st-line therapy. In Japan, agents approved for 1st-and 2nd-line therapies may be used for later-line treatment in real-world practice. In this study, we aimed to investigate the efficacy and safety of atezolizumab plus bevacizumab as 1st-line and later-line therapies in real-world practice. Moreover, we analyzed the factors associated with time to progression (TTP) and the association of atezolizumab plus bevacizumab therapy with the timing of the AFP response and disease progression.

Patients
A total of 52 patients with u-HCC received atezolizumab plus bevacizumab therapy at the Musashino Red Cross Hospital between October 2020 and April 2021. Atezolizumab plus bevacizumab therapy was administered after discussions by experts at a tumor board in the institution. HCC diagnosis and evaluation of disease progression were based on guidelines proposed by the Japan Society of Hepatology [10], the American Association for the Study of Liver Disease [11], and the European Association for the Study of the Liver [12]. Written informed consent was obtained from all patients, and the ethics committee of the Musashino Red Cross Hospital approved the study in accordance with the Declaration of Helsinki.

Treatment protocol
Patients received intravenous atezolizumab (1200 mg) plus bevacizumab (15 mg/kg) every 3 weeks until disease progression or unacceptable toxicity. Treatment interruption was performed based on the manufacturer's guidelines. Dynamic computed tomography was performed at baseline, 6-8 weeks after atezolizumab plus bevacizumab administration, and every 6-8 weeks thereafter. Treatment response was reported based on the Response Evaluation Criteria in Solid Tumors (RECIST) [13] and the modified RECIST [14]. The AFP level was also evaluated at baseline and every 3 weeks thereafter. Adverse events (AEs) were evaluated according to the Common Terminology Criteria for Adverse Events version 5.0. To assess changes in liver function, the albumin-bilirubin (ALBI) score [15,16] was analyzed at baseline and at 3, 6, and 9 weeks after treatment. The data cut-off date of clinical response and ALBI score were defined on July 31, 2021.

Statistical analysis
Overall survival (OS) was measured from the date of atezolizumab plus bevacizumab administration to the date of death from any cause. Patients who were lost to follow-up were censored at their last visit or contact. Patients who were still alive on July 31, 2021 were censored. Progression-free survival (PFS) was measured from the date of atezolizumab plus bevacizumab administration to the date of radiological tumor progression or death from any cause. Data are expressed as median (range) or n (%). Fisher's exact test, Mann-Whitney U test, paired t-test, Kaplan-Meier method, and log-rank test were used for statistical analyses. Statistical significance was set at P < 0.05. Statistical analyses were performed using EZR (Saitama Medical Centre, Jichi Medical University, Shimotsuke, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [17].
Radiological evaluation after atezolizumab plus bevacizumab administration was performed in 46 patients. With regard to the best antitumor response according to the RECIST, 1 patient achieved complete response (CR), 7 patients achieved partial response (PR), 22 patients had stable disease (SD), 16 patients had PD, and 6 patients had no evaluable disease (NE). The objective response

Changes in liver function during atezolizumab plus bevacizumab therapy
The changes in the ALBI score in 52 patients who received atezolizumab plus bevacizumab are shown in Fig. 3 , respectively. There were no significant differences in the median ALBI scores at baseline and at 3 weeks, at baseline and at 6 weeks or at baseline and at 9 weeks in either group (Fig. 4).

Adverse events during atezolizumab plus bevacizumab therapy
AEs observed during atezolizumab plus bevacizumab therapy are shown in Table 3. Any grade AE was observed in 36 of the 52 patients. The rates of total AEs in patients receiving atezolizumab plus bevacizumab as 1st-line therapy (n = 23) and those receiving atezolizumab plus bevacizumab as 2nd-or later-line therapy (n = 29) were 60.9% and 75.9%, respectively (P = 0.37). When comparing AEs with an incidence of 10% or more, the rates of AEs were 56.5% in patients receiving atezolizumab plus bevacizumab as 1stline therapy and 75.9% in patients receiving atezolizumab plus bevacizumab as 2nd-or later-line therapy (P = 0.23).
No infusion reaction was reported in any patient. During the observation period, six patients discontinued atezolizumab plus bevacizumab therapy owing to AEs, as follows: two patients with transaminase increase, one patient with interstitial pneumonia, one patient with meningitis and transaminase increase, one patient with renal dysfunction, and one patient with gastrointestinal bleeding. Four of the six patients were treated with corticosteroids. The rates of No AEs other than those reported in the IMbrave150 trial were observed.

Clinical outcomes in the 1st-line group and the 2ndor later-line group
We compared TTP between the patients treated with atezolizumab plus bevacizumab as 1st-line (n = 22) and 2nd or later-line (n = 24). The baseline characteristics are shown in Table 4. The proportions of patients receiving a full dose of atezolizumab plus bevacizumab in the 1st-line group and the 2nd-or later-line group were 72.7% and 58.3%, respectively (P = 0.23    (Fig. 5).
We next analyzed the factors associated with TTP in patients receiving atezolizumab plus bevacizumab (n = 46). In the multivariate analysis, receiving atezolizumab plus bevacizumab as 2nd-or later-line therapy was the only significant factor (hazard ratio 3.16, 95% confidence interval 1.03-9.74, P = 0.04) ( Table 5).
First-line group: received atezolizumab plus bevacizumab as 1st-line therapy.
The association between the AFP response and therapeutic efficacy in patients receiving atezolizumab plus bevacizumab therapy.
We investigated the AFP response as a predictor of DCR according to RECIST in patients receiving atezolizumab plus bevacizumab therapy. The AFP response was defined as a reduction in AFP from the baseline of ≥ 20%. Thirtytwo patients continued atezolizumab plus bevacizumab therapy for more than 6 weeks and were included in the analysis. At 3 weeks, 12 patients showed an AFP response. The AFP response at 3 weeks had a positive predictive value (PPV) of 91.7% and a negative predictive value (NPV) of 40.0%. At 6 weeks, 12 patients showed an AFP response. The AFP response at 6 weeks had a PPV of 91.7% and an NPV of 40.0%. The TTPs of patients with and without an AFP response at 3 weeks and 6 weeks are presented in Fig. 6. There were no significant differences in the median TTP between patients with and without an AFP response at 3 weeks (3.8 months vs. 4.8 months, P = 0.09). By contrast, there was a significant difference in the median TTP between patients with and without an AFP response at 6 weeks (2.6 months vs. 4.7 months, P = 0.02). The results were similar when patients with a baseline AFP < 20 ng/mL were excluded. Although there were no significant differences in the median TTP between patients with and without an AFP response at 3 weeks (3.8 months vs. 4.1 months, P = 0.08), there was a significant difference in the median TTP between patients with and without an AFP response at 6 weeks (2.5 months vs. 4.7 months, P = 0.001) (Fig. 7).
We also analyzed the association between the AFP response at 6 weeks and therapeutic efficacy stratified by patients with baseline AFP < 20 ng/mL, 20-400 ng/mL, and > 400 ng/mL (Table 6). Among 10 patients with a baseline AFP < 20 ng/mL, 3 had an AFP response at 6 weeks.  According to the RECIST, 2 patients achieved PR and 1 had SD. Among 10 patients with a baseline AFP 20-400 ng/mL, 4 had an AFP response at 6 weeks. One patient achieved PR, 2 had SD, and 1 had PD. Among 12 patients with a baseline AFP > 400 ng/mL, 5 had an AFP response at 6 weeks. One patient achieved CR, 1 achieved PR, and 3 patients had SD. There were no significant differences in the median TTP among patients with baseline AFPs < 20 ng/mL, 20-400 ng/mL, and > 400 ng/mL (5.2 months, 4.0 months, and 4.8 months, respectively; P = 0.19) in patients with an AFP response at 6 weeks (Fig. 8).

Discussion
Atezolizumab plus bevacizumab is recommended as a 1stline therapy for advanced HCC in many guidelines, including the ASCO [18] and ESMO guidelines [19]. Some recent studies have reported the efficacy and safety of atezolizumab plus bevacizumab in Japanese patients with u-HCC in realworld practice [20][21][22]. To the best of our knowledge, this is the first report to reveal the difference in clinical outcome between patients receiving atezolizumab plus bevacizumab as 1st-line and 2nd-or later-line therapies and the association between the timing of the AFP response and therapeutic efficacy.
We included patients who received atezolizumab plus bevacizumab as 2nd-or later-line therapy in this study. The IMbrave150 trial [23] included only patients who had not previously received systemic therapy, and the ORR and DCR were 27.3% and 73.6%, respectively, according to the RECIST. Recent studies [20][21][22] showed there were no significant differences in radiological responses to atezolizumab plus bevacizumab between 1st-line and 2nd-or laterline patients. In our study, the median TTP was significantly shorter in the 2nd-or later-line group than in the first-line group. In patients previously treated with other agents, dose modification and interruption of bevacizumab were more frequent than in patients receiving atezolizumab plus bevacizumab as 1st-line therapy. However, there were no significant differences in the rate of AEs between the two groups.
Previous reports reported deterioration in liver function during the early period among patients treated with sorafenib [24] and lenvatinib [25]. Sangro et al. [26] reported that patients receiving nivolumab were more likely to have preserved liver function than patients receiving sorafenib. Kudo et al. [27] reported that ramucirumab did not worsen the ALBI score during the early treatment period. Changes in liver function during atezolizumab plus bevacizumab therapy were not examined in the IMbrave150 trial. Recent studies [20,21] showed the liver function was slightly worsened after 2-3 weeks and recovered at 6 weeks during atezolizumab plus bevacizumab therapy. In our study, the ALBI score at 3 weeks transiently worsened compared to baseline but improved at 6 weeks and 9 weeks. The temporary worsening of liver function may be related to the therapeutic response, which leads to appetite loss and fatigue during the early treatment period. Liver function and these AEs improved spontaneously at 6 and 9 weeks. Atezolizumab plus bevacizumab therapy can be maintained while preserving liver function in many patients.
In this study, we focused on the AFP response at 3 and 6 weeks after the administration of the combination therapy. There were no significant differences in the median TTP between patients with and without an AFP response at 3 weeks. By contrast, there was a significant difference in  Fig. 8 The comparison of time to progression among baseline AFP in patients with an AFP response at 6 weeks Abbreviation: AFP, alpha-fetoprotein the median TTP between patients with and without an AFP response at 6 weeks. Previous studies reported early AFP response is useful to predict response to sorafenib [28,29], lenvatinib [30] and checkpoint inhibitor therapy [31]. However, in patients receiving atezolizumab plus bevacizumab therapy, the antitumor response is delayed in some patients, and early treatment decisions based on the AFP response should be avoided. In Japan, five regimens (sorafenib, regorafenib, lenvatinib, ramucirumab if AFP ≥ 400 ng/mL, and cabozantinib) can be administered after atezolizumab plus bevacizumab therapy based on the national health insurance guidelines. Because of the small cohort and short observation period, we could not examine the established clinical outcomes of post-combination therapies in real-world practice. Among the patients who received lenvatinib after combination therapy, the ORR and DCR were 50% and 75%, respectively. Aoki et al. [32] revealed that lenvatinib demonstrated considerable antitumor effects with acceptable safety in patients with progressive and u-HCC when administered immediately after PD-1/PD-L1 blockade failure (ORR 55.6%, DCR 86.1%). Further multicenter studies on treatment after atezolizumab plus bevacizumab therapy are necessary to develop an effective treatment strategy for patients with u-HCC.
There are certain limitations to our study. The study was a retrospective single-center study with a small number of patients and a short observation period, and tumor response was not evaluated in all patients.
To our knowledge, our report is the first to reveal the difference in clinical outcome between 1st-line and 2ndor later-line atezolizumab plus bevacizumab therapy and the association between the timing of the AFP response and therapeutic efficacy. Our data can provide useful information to develop useful therapeutic strategies for patients with u-HCC and effectively administer combination therapy.

Conclusion
Atezolizumab plus bevacizumab therapy was found to have similar radiological responses to the phase III IMbrave150 trial regimen in u-HCC patients when administered as 1stline therapy. Patients receiving atezolizumab plus bevacizumab therapy as 1st-line therapy tended to have longer TTP than those receiving atezolizumab plus bevacizumab as 2ndor later-line therapy. The AFP response at 6 weeks could be a predictive indicator of disease progression.