Factors associated with the response to atezolizumab/bevacizumab combination therapy for hepatocellular carcinoma

Abstract Background and Aim The purpose of this study was to analyze factors associated with the overall survival (OS) of atezolizumab/bevacizumab combination therapy for advanced hepatocellular carcinoma (aHCC). We also assessed the OS of patients with ineffective therapy and those who discontinued treatment owing to adverse events (AEs). Methods This retrospective multicenter study involved 139 patients with aHCC who received atezolizumab/bevacizumab combination therapy between November 2020 and September 2022. Results The median duration of treatment was 136.5 days, and the median observation period was 316 days. The overall response rate was 40%, and the disease control rate was 78% according to mRECIST criteria. Grade ≥2 AEs occurred in 63 patients (43%) and led to treatment discontinuation in 16 patients. Multivariate analysis revealed that treatment response and occurrence of grade ≥2 AEs after therapy, as well as low level of albumin‐bilirubin (ALBI) grade and low level of des‐gamma carboxy prothrombin (DCP) before therapy, were extracted as factors that contributed to OS. Log‐rank tests with the Kaplan–Meier method showed significant differences in OS among these factors. The OS of patients who discontinued owing to AEs was significantly shorter than that of other patients. Conclusion Not only factors before therapy but also treatment response and the appearance of AEs are involved in OS for atezolizumab/bevacizumab combination therapy. Although the development of AEs also contributed to OS, appropriate management of AEs is important to avoid discontinuing treatment with this combination.


Introduction
Atezolizumab/bevacizumab combination therapy is becoming established as a first-line treatment for hepatocellular carcinoma (HCC). 1,2 Evidence suggests that the combination is safe to use and effective, even in elderly patients, because it has less severe side effects than conventional cytotoxic anticancer agents. 3,4 Recent realworld studies have also reported its efficacy in patients with Child-Pugh B liver disease. 5 In this multicenter study, the efficacy of atezolizumab/bevacizumab combination therapy was examined, and factors contributing to overall survival (OS) were analyzed.
On the other hand, immune checkpoint inhibitors (ICI) are associated with characteristic adverse events (AEs), which are reported to occur in approximately 20% of patients, 6 and some patients do not achieve a therapeutic response because they stop treatment owing to AEs. However, some studies have shown that the appearance of AEs during treatment with an ICI is associated with their efficacy, 7 and the implications of AE for treatment efficacy are not yet fully understood. In this study, the safety of atezolizumab/bevacizumab combination therapy and the impact of AE on treatment efficacy were also examined. therapy for HCC at various institutions from November 2020 to September 2022 (Kobe University Hospital, Hyogo Prefectural Hyogo Cancer Center, Hyogo Prefectural Kakogawa Medical Center, Yodogawa Christian Hospital, Kobe Asahi Hospital, Osaka Saiseikai Nakatsu Hospital, Kitaharima Medical Center, Akashi Medical Center, Hyogo Prefectural Awaji Medical Center, and Sanda City Hospital). HCC was diagnosed based on an increasing trend of serum level of α-fetoprotein (AFP) and/or des-gamma carboxy prothrombin (DCP), as well as typical imaging findings and/or pathological findings.
Patients positive for hepatitis B virus surface antigen (HBsAg) or anti-hepatitis C virus (HCV) were etiologically included in HCC due to hepatitis viral infection. patients with a history of alcohol abuse of 60 g/day or more were included as alcoholics. 8 The treatment schedule was followed as approved regimens of atezolizumab and bevacizumab for advanced HCC. Both drugs were given at 21-day intervals, with their continuation at the physician's discretion depending on AEs and patient's performance status.
As a rule, imaging evaluations were performed with contrast-enhanced CT or MRI scans every 6-8 weeks, but intervals were allowed to vary according to the patient's condition and the attending physician's judgment. The treatment efficacy was evaluated according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST) based on contrast-enhanced computed tomography or contrast-enhanced magnetic resonance imaging. 9 AEs were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0 (CTCAE version 5.0). According to the instructions, when grade 3 or more severe AEs occurred, dose reduction or temporary interruption was provided until symptoms subsided to pharmaceutically manageable grades 1 or 2.
This was a retrospective analysis of records stored in a database, and official approval was received based on the Guidelines for Clinical Research issued by the Ministry of Health and Welfare of Japan. This study was approved by the Ethics Committee of Kobe University Graduate School of Medicine and each participating institution (B200286).
Statistical analysis. Factors associated with the efficacy of atezolizumab/bevacizumab combination therapy were analyzed using t tests and the log-rank test with the Kaplan-Meier method to evaluate OS. To identify which factors significantly contributed to OS, Cox proportional hazards model was used for univariate and multivariate survival analysis. Statistical analyses were performed using SPSS version 28.

Results
Patient responses and characteristics of nonresponders. Overall, 133 patients were classified as Child-Pugh A, 102 as ALBI grade 1 or 2a, and 56 were within the up-to-seven criteria ( Table 1). Esophageal varices were found in 43 patients, including after prophylactic hemorrhage treatment. The etiology comprised hepatitis viral infection in 75 patients (HBV in 27 and HCV in 48 patients, respectively), alcoholic cirrhosis in 43 patients, and nonalcoholic steatohepatitis in 19 patients. Comorbidities included hypertension in 81 patients and diabetes in 35 patients. One hundred and six patients had received prior therapy before atezolizumab/bevacizumab combination therapy, including chemotherapy in 51 patients.
The median duration of treatment was 136.5 days (IQR: 63-227.5 days, range: 20-601 days), median number of treatment sessions was 6.0 cycles (IQR: 3.0-10.0 cycles, range: 1-28), and the observation period ranged from 23 to 733 days (median 316 days). The best response, based on mRECIST, was complete response (CR) in 6 patients, partial response (PR) in 48 patients, stable disease (SD) in 55 patients, and progressive disease (PD) in 30 patients. Therefore, the overall response rate (ORR) was 40%, and the disease control rate (DCR) was 78%. The log-rank test with the Kaplan-Meier method revealed that patients with ORR and DCR experienced a survival benefit compared with the other patients ( Fig. 1). The patients were therefore classified as either responders (patients who achieved CR or PR) or non-responders (patients with SD or PD), and the characteristics of these two groups were compared, as summarized in Table 1. The responders included greater proportions of patients who had received prior therapy, including chemotherapy, and who met the upto-seven criteria. The responders also received atezolizumab/ bevacizumab combination therapy for longer and had a survival benefit compared with the non-responders.
We also investigated possible factors associated with treatment response. Grade ≥2 AEs occurred in 63 patients (43%), of whom 16 stopped treatment owing to the AEs.
Examination of grade ≥2 AEs. Figure 3 shows the frequency of grade ≥2 AEs and the proportion of patients who discontinued treatment owing to AEs. The most frequent AE was proteinuria (27 patients, 42.2%), and 6 patients (9.4%) discontinued owing to proteinuria. Hypertension occurred in 15 patients (23.4%), but none led to treatment discontinuation. Liver dysfunction occurred in 12 patients (18.8%), and 4 patients (6.3%) discontinued. Other AEs included bleeding in 5 patients (7.8%), skin symptoms in 7 patients (10.9%), thrombocytopenia in 3 patients (4.7%), fatigue in 2 patients (3.1%), and lung injury in 2 patients (3.1%). By grade, 36 patients suffered grade 2 AEs,  Log-rank tests with the Kaplan-Meier method were performed for patients with grade ≥2 AEs. Figure 2d shows that the OS of patients who discontinued owing to AEs was significantly shorter than that of the other patients (P < 0.025).

Figure 2
Kaplan-Meier plots of overall survival. Overall survival was determined according to the serum DCP level before therapy (a), albuminbilirubin (ALBI) score before therapy (b), and occurrence of grade ≥ 2 adverse events after starting therapy (c). Overall survival in patients with grade ≥ 2 adverse events was also compared between patients who continued treatment and patients who discontinued treatment owing to adverse events (d). DCP, des-gamma carboxy prothrombin; ABLI, albumin-bilirubin.
Y Yano et al.

Discussion
In this study, low DCP level and ALBI 1/2a before therapy were expected to be predictive factors of efficacy. It was supportive that many factors have been reported to predict the efficacy of atezolizumab/bevacizumab combination therapy, including the ALBI score before treatment and changes in the AFP and DCP levels after the start of treatment. [10][11][12] It is also reasonable that achievement of CR/PR is related to OS. On the other hand, our study also revealed that patients with AEs are involved in OS. With regard to ICIs, which were introduced in recent years, the appearance of immune-related AEs in patients with lung cancer or malignant melanoma indicates activation of the patient's immune response, which contributes to the antitumor effect and prolonged OS. [13][14][15][16] Regarding HCC, although the significance of immune-related AEs has not been fully demonstrated, these events were shown to be associated with PFS. 7 The results of this study support this concept. However, a survival benefit was not observed in patients who discontinued treatment owing to AEs. Therefore, it is important to continue treatment by reducing the dose or adjusting the duration of administration as appropriate.
In this study, hypertension, proteinuria, and bleeding, which are considered to be associated with bevacizumab, were the most frequent events, whereas immune-related AEs of grade ≥2 were less frequent, occurring in 18 patients. Interstitial pneumonia occurred in two patients and led to treatment discontinuation in both patients. The complications of interstitial pneumonia can be severe, demonstrating the need for caution. 17 Proteinuria with grade ≥2 was observed in 26 patients and resulted in treatment discontinuation in six of these patients. Because it was recently shown that changes in proteinuria early in treatment are associated with grade ≥2 proteinuria, physicians should be vigilant for early changes in proteinuria after starting treatment. 18 Nevertheless, many patients with bevacizumab-related AEs could continue atezolizumab as monotherapy. Thus, it is important to maintain the antitumor effects by continuing treatment.
Bevacizumab-related bleeding was reported as a grade ≥2 AE in seven patients. The atezolizumab/bevacizumab combination therapy was discontinued in four patients with ruptured esophageal varices reported as grade ≥3 AEs and in one patient with cerebral hemorrhage. All of the patients included in this study underwent endoscopy and received prophylactic treatment prior to the start of therapy. However, in prior reports, atezolizumab/bevacizumab combination therapy was associated with worsening esophageal varices and a high risk of rebleeding in patients with a history of variceal rupture. [19][20][21] Therefore, endoscopic examination is considered necessary during treatment, especially in high-risk patients.
Currently, atezolizumab/bevacizumab combination therapy is often chosen as first-line systemic chemotherapy for HCC, whereas lenvatinib and sorafenib are often used as second-line therapy. Durvalumab combined with tremelimumab has also shown promising efficacy, and its use may be considered in the future in consideration of the patient's general condition and the risk of AEs. 22 Limitations of this study include the fact that it was a multicenter study, there were no criteria for starting or changing chemotherapy, and switching to atezolizumab monotherapy was at the discretion of the primary care physician.
In conclusion, not only factors before therapy but also treatment response and the appearance of AEs are involved in OS for atezolizumab/bevacizumab combination therapy. Although the development of AEs also contributed to OS, appropriate management of AEs is important to avoid discontinuing treatment with this combination.

Ethics statement
The study was carried out in accordance with the Declaration of Helsinki and was approved by the institutional review board of the Kobe University Graduate School of Medicine and by the institutional review boards of the participating hospitals (no. B200286). Efficacy factors of atezo/bev for HCC Y Yano et al.