INTRODUCTION

Hepatocellular carcinoma (HCC) is the most common primary malignancy of the liver and fifth most common malignancy worldwide.1 Repeated recurrence after curative treatments is not uncommon, even when HCC is diagnosed at an early stage. Following progression to intermediate stage (Barcelona Clinic Liver Cancer [BCLC]-B) and unresectable status (u-HCC), the previous version of the BCLC staging system recommended transcatheter arterial chemoembolization (TACE).2 Recently, “TACE-refractoriness”3 and “TACE-unsuitable”4, 5 have been proposed as terms to describe patient condition for improvement of prognosis of those with BCLC-B stage u-HCC along with continuous development of tyrosine kinase inhibitors (TKI) and anti-vascular endothelial growth factor receptor (anti-VEGFR) (e.g., sorafenib,6 regorafenib,7 lenvatinib,8 ramucirumab,9 and cabozantinib10). It has also been reported that switching to an TKIs in patients with BCLC-B stage u-HCC classified as beyond up to 7 criteria (UT-7) (seven as the sum of the size of the largest tumor in cm and number of tumors)11 and such conditions contributed to their improved prognosis.

In September 2020, atezolizumab plus bevacizumab treatment (Atez/Bev)12 was approved as a new treatment to be given with an immune checkpoint inhibitor (ICI) and anti-VEGFR for u-HCC. As a result, Atez/Bev has become to be shown as one of therapeutic options for BCLC-B stage HCC as well as TACE in the newest BCLC staging.13 However, few reports have noted details regarding the therapeutic efficacy of Atez/Bev for beyond UT-7 patients with multiple tumors (UT-7 out/multiple) in clinical practice. This study aimed to elucidate therapeutic response of UT-7 out/multiple/Child-Pugh A u-HCC patients who received Atez/Bev.

MATERIALS AND METHODS

From September 2020 to September 2021, Atez/Bev was introduced to 306 Japanese u-HCC patients, including 95 with UT-7 out/multiple patients with Child-Pugh class A, receiving treatment at 21 different institutions (Figure S1). Response Evaluation Criteria in Solid Tumors (RECIST) ver. 1.114 and modified RECIST (mRECIST)15 were used for evaluation of therapeutic response. The initial assessment of the effect of therapy was performed using dynamic-computed tomography (CT) results obtained at approximately six weeks (6W) after introduction of Atez/Bev whenever possible, then additional dynamic-CT examinations were performed as needed depending on patient condition, even before 6W in some cases. After the first 6W, dynamic-CT examinations were performed every 6W and then every 12 weeks after the first six months. Patients with a known history of auto-immune disease were not treated with Atez/Bev. All were examined using upper gastrointestinal endoscopy for surveillance of gastroesophageal varices. When bleeding was detected or if a high risk was present, the patient was treated according to local clinical practice.

Underlying liver disease

Positive anti-HCV findings were considered to indicate that HCC was due to hepatitis C virus (HCV), whereas HCC due to hepatitis B virus (HBV) was determined when the HBV surface antigen was positive. For patients with a history of alcohol abuse (≥60 g/day),16, 17 underlying liver disease was judged as related to alcohol.

Liver function assessment

Child-Pugh classification,18 albumin-bilirubin (ALBI) grade,19, 20 and modified ALBI (mALBI) grade,21 for which ALBI grade 2 was divided into two sub-grades (mALBI 2a and 2b) using an ALBI score of −2.27 as the cut-off value, were used for hepatic reserve function assessment.

HCC diagnosis and treatment

HCC diagnosis was based on an increasing course of alpha-fetoprotein (AFP), as well as dynamic-CT,22 Magnetic Resonance Imaging (MRI),23, 24 and/or pathological findings obtained during the clinical course. BCLC13 and tumor node metastasis (TNM) staging for HCC presented by the Liver Cancer Study Group of Japan (LCSGJ), sixth edition25 (TNM-LCSGJ), were used for evaluations of tumor progression.

Atez/Bev treatment and adverse event assessment

After obtaining written informed consent from the patient, intravenous Atez/Bev treatment, composed of 1200 mg of atezolizumab plus 15 mg/kg of body weight of bevacizumab, was given every three weeks (3W),12 following guidelines for Atez/Bev treatment provided by the manufacturer. Treatment was discontinued following observation of any unacceptable or serious adverse event (AE), or clinical tumor progression. The National Cancer Institute Common Terminology Criteria for Adverse Events, ver5.0,26 was used for assessment of AEs. At the time of Atez/Bev discontinuation, introduction of the next treatment was determined by the attending physician.

After receiving official approval, this study was conducted as a retrospective analysis of database records based on the Guidelines for Clinical Research issued by the Ministry of Health and Welfare of Japan. All procedures were done in accordance with the declaration of Helsinki. Written informed consent was received from each of the enrolled patients.

Statistical analysis

For statistical analysis, Welch t-test, Mann-Whitney's U test, the Kaplan–Meyer method, and a log-rank test were used. p values less than 0.05 were considered to indicate statistical significance. Easy-R (EZR), ver1.53 (Saitama Medical Center, Jichi Medical University, Saitama, Japan),27 a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria), was used to perform all of the statistical analyses.

RESULTS Characteristics of enrolled patients

Three hundred six u-HCC patients were treated with Atez/Bev at our institutions (median PFS was 6.5 months, 95% confidential interval (CI) 5.3–7.8, 6-month survival rate was 88.8%) (Median PFS (mPFS) was 9.7 month (95% confidence interval (CI): 6.5-not applicable (NA)) in mALBI 1, 6.7 months (95%CI: 3.6-NA) in mALBI 2a, 4.5 months (95% CI: 3.8-6.2) in mALBI 2b, and 0.4 months (95%CI: 0.4-NA) in mALBI 3 (P = 0.004)) (Figure S2). Of the 306 u-HCC patients, 95 BCLC-B/multiple/UT-7 out/Child-Pugh A were enrolled. Characteristics of the present cohort (n = 95) are shown in Table 1. Median age was 76 years and males numbered 73 (76.8%). Child-Pugh score 5 was noted in 68 (71.6%), and score 6 in 27 (28.4%). Atez/Bev was given as first-line treatment in 52 (54.7%), while 43 (45.3%) received it as later-line therapy (2nd:3rd:4th:5th = 34:7:1:1). The median observation period was 6.0 months (Table 1). At the time of writing, Atez/Bev had been stopped in 44 (46.3%) (PD:AE:hepatic failure:hemobilia: Atez/Bev curative (ABC) conversion therapy28:others = 30:8:2:1:1:2). Of the 44, reasons of stopping Atez/Bev due to AEs in eight patients were appetite loss in 3, and colitis, liver functional abnormality, hyperthyroidism, urine protein, and fever were 1 patient each.

TABLE 1. Characteristics of BCLC-B patients classified as beyond up to seven criteria with multiple tumors treated by Atez/Bev (n = 95) Age years (Median, IQR) 76 (71–80) Gender (Male:Female) 73:22 BMI kg/m2 (Median, IQR) 23.7 (21.1–25.7) Etiology, HCV:HBV:alcohol:other 25:13:20:37 Positive for diabetes mellitus n (%) 36 (37.9%) Platelets ≥104/μl (Median, IQR) 14.0 (11.0–18.6) AST U/L (Median, IQR) 39 (28–47) ALT U/L (Median, IQR) 26 (18–38) T-bilirubin mg/dL (Median, IQR) 0.7 (0.5–0.9) Albumin g/dL (Median, IQR) 3.8 (3.5–4.1) Prothrombin time % (Median, IQR) 91.1 (83.9–99.0) ALBI score (Median, IQR) −2.56 (−2.25 to −2.76) mALBI grade 1:2a:2b 44:22:29 Child-pugh score (5:6) 68:27 AFP ng/mL (Median, IQR) 21 (6.3–192.8) Elevated AFP (≥400 ng/ml) n (%) 21 (22.1%) TNM LCSGJ sixth (II:III) 17:78 Tumor size, maximum cm (Median, IQR) 3.0 (2.2–5.0) Tumor number (2:3:4:5:≥6) 5:6:1:4:79 Initial dose of bevacizumab mg (Median, IQR) 900 (800–1000) Patients without past history of treatments, n (%) 12 (12.6%) Treatment line (1st:2nd:3rd:4th:5th) 52:34:7:1:1 Past history of TACE alone treatment after development to BCLC-B stage (none:1:2:≥3) 46:17:11:15 (excluded six patients without information) Past history of TACE refractoriness in patients treated with TACE alone treatment history n (%) 39 (44.8%) (excluded eight patients without information) Past history of systemic drug therapy (SOR:LEN:REG:RAM:ICI) 9:42:2:1:1 Number of administrations of atezolizumab in the present observation perioda n (Median, IQR) 5 (3–9) Number of administrations of bevacizumab in the present observation perioda n (Median, IQR) 5 (3–7) Observation period (months) Months (Median, IQR) 6.0 (2.8–8.4) Note: Numbers in parentheses indicate range, unless otherwise noted. Abbreviations: AFP, alpha-fetoprotein; ALBI score, albumin-bilirubin score; ALT, alanine aminotransferase; AST, aspartate transaminase; Atez/Bev, atezolizumab plus bevacizumab; BCLC stage, Barcelona Clinic Liver Cancer stage; BMI, body mass index; HBV, hepatitis B virus; HCV, hepatitis C virus; ICI, immune-check point inhibitor; LEN, lenvatinib; mALBI grade, modified ALBI grade; TNM LCSGJ sixth, tumor node metastasis stage by Liver Cancer Study Group of Japan sixth edition; TACE, transcatheter arterial chemoembolization; RAM, ramucirumab; REG, regorafenib; SOR, sorafenib. a Median (interquartile range). Initial therapeutic response using RECIST at six weeks

In the first evaluation using RECIST performed at approximately six weeks, CR, PR, SD, and PD were 1 (1.2%), 14 (16.5%), 57 (67.0%), and 13 (15.3%), respectively, while CR, PR, SD and PD in 2 (2.5%), 32 (40.0%), 35 (43.7%), and 11 (13.8%) by mRECIST after excluded patients without dynamic-imaging study (n = 5), respectively. Objective response rate (ORR)/disease control rate (DCR) of RECIST and mRECIST at 6W were 17.7%/84.7% and 42.5%/86.2%, respectfully (Table 2). MPFS was 8.0 months (95% CI: 6.5 to NA) (Figure 1a), and overall-survival (OS) rate and PFS rate at six months were 95.8% and 63.2%, respectively. Although there was no significant difference in analysis for mPFS according to mALBI grade, better mPFS was observed in patients with better hepatic function (mALBI 1:2a:2b = 9.7:7.8:6.2  months, p = 0.452) (mALBI 1/2a:2b = 8.0:6.2 months, p = 0.348) (Figure S3).

TABLE 2. Initial therapeutic responses with RECIST and modified RECIST at six weeks CR PR SD PD ORR DCR NE/ND RECIST All (n = 85) 1 (1.2%) 14 (16.5%) 57 (67.0%) 13 (15.3%) 17.7% 84.7% 10 First-line 1 (2.3%) 8 (18.6%) 28 (65.1%) 6 (14.0%) 20.9% 86.0% 9 Later-line 0 (0%) 6 (14.3%) 29 (69.0%) 7 (16.7%) 14.3% 83.3% 1 mRECIST All (n = 80) 2 (2.5%) 32 (40.0%) 35 (43.7%) 11 (13.8%) 42.5% 86.2% 15 First-line 1 (2.3%) 20 (46.5%) 16 (37.2%) 6 (14.0%) 48.8% 86.0% 9 Later-line 1 (2.7%) 12 (32.4%) 19 (51.4%) 5 (13.5%) 35.1% 86.5% 6 Abbreviations: CR, complete response; DCR, disease control rate; mRECIST, modified RECIST; NE, no examination; ND, no data of dynamic study; ORR, objective response rate; PD, progressive disease; PR, partial response; RECIST, Response Evaluation Criteria in Solid Tumors ver. 1.1; SD, stable disease.

(a) Progression-free survival (PFS) of all patients (n = 95). Median PFS was 8.0 months (95% CI: 6.5-not applicable (NA)). (b) Median PFS of patients given atezolizumab plus bevacizumab as first-line use group was 11.1 months (95% CI: 5.25-NA), while that in the later-line use group was 7.8 months (95% CI: 4.6-NA) (p = 0.172)

When therapeutic response was evaluated according to treatment line (first- and later-line), ORR/DCR of RECIST and mRECIST at 6W in the first-line treatment group were 20.9%/86.0% and 48.8%/86.0%, while those in the later-line treatment group were 14.3%/83.3% and 35.1%/86.5%, respectively (p = 0.437 and p = 0.360, respectively). There were significant differences for Child-Pugh score (5:6 = 80.8%:19.2% vs. 60.5%:39.5%, p = 0.040) between the first- and later-line groups, whereas statistical differences were not observed in mALBI grade (grade 1:2a:2b = 50.0%:23.1%:26.9% vs. 41.9%:23.2%:34.9%, p = 0.373) and mPFS (11.1 vs. 7.8 months, p = 0.172) between the groups in the present study, respectively (Figure 1b). There were significant differences in TACE number after development to BCLC-B (none:1:2:≧3 = 34:9:4:3 vs. 12:8:7:12, p = 0.001), while frequency of patients, in whom systemic treatments was introduced due to TACE refractoriness did not show a significant difference (31.3% (15/48) vs. 61.5% (24/39), p = 0.07) between both groups (excluded patients without each information, respectively). There were no significant differences in all AEs between first-line and later-line groups (Table S1).

Adverse events and influence of Atez/Bev on hepatic function

As for AEs (>10%) during the present observation period of Atez/Bev treatment, general fatigue was most frequently observed (any-grade 23.2%; grade 1:2:3 = 17:4:1, 17.9%:4.2%:1.1%), followed by urine protein (any-grade 21.1%; grade 1:2:3 = 9:4:7, 9.5%:4.2%:7.4%), appetite loss (any-grade 20.0%; grade 1:2:3 = 13:3:3, 13.7%:3.2%:3.2%), and hypertension (any-grade 13.7%; grade 1:2:3 = 4:6:3, 4.2:%:6.3%:3.2%). Treatment related AE (any grade) was observed in 57 (60%) and treatment related AE (grade 3 or more) was in 23 (24.2%).

Relative changes of hepatic function are shown in Figure 2. Child-Pugh A was seen in 100% at the baseline, then in 89.8% at 3W, in 94.8% at 6W, and in 91.4% at nine weeks (9W) (Figure 2a). mALBI 1/2a was noted in 69.4% at the baseline, and 57.3%, 65.3%, and 60.0% at 3, 6 and 9W, respectively (Figure 2b). After the introduction of Atez/Bev treatment, its influence on hepatic function was small.

(a) Relative changes in Child-Pugh score after introducing atezolizumab plus bevacizumab. Child-Pugh class A was 100% at the baseline (BL), 89.8% at three weeks (3W), 94.8% at six weeks (6W), and 91.4% at nine weeks (9W). (b). Relative changes in modified albumin-bilirubin grade (mALBI) after introducing atezolizumab plus bevacizumab. mALBI 1/2a was 69.4% at BL, 57.3% at 3W, 65.3% at 6W, and 60.0% at 9W. Observation period of 4 patients were less than 3W and 2 had no data of albumin level at 3W

DISCUSSION

To our best knowledge, the present study is the first to examine the therapeutic efficacy of Atez/Bev for u-HCC patients classified as UT-7 out/multiple/Child-Pugh A. It was reported that mPFS of patients with BCLC-B including UT-7 in (n = 76) is reported as 12.6 months in update data of IMbrave150 trial (median observation period: 15.6 months).29 Although the present observation period was not still long enough, Atez/Bev can be expected a favorable therapeutic response, because the 6 months PFS rate and mPFS of the present UT-7 out/multiple/Child-Pugh A BCLC-B u-HCC cohort were 63.2% and 8.0 months (median observation period: 6.0 months), and the first evaluation of therapeutic efficacy was conducted at approximately 6W following introduction of Atez/Bev and the results were favorable, as ORR/DCR of RECIST and mRECIST were 17.7%/84.7% and 42.5%/86.2%, respectively.

Recently, the “TACE-unsuitable”4, 5 concept was established, which proposes introduction of systemic treatment, such as sorafenib, lenvatinib, or Atez/Bev, but not TACE as first-line therapy for intermediate stage HCC (e.g., UT-7 out/multiple). In addition, the concept “TACE-refractoriness”,3 which recommends switching to systemic treatment for patients without therapeutic response by TACE, has been reported. In a past report,30 ORR and mPFS were 37.8% and 3.0 months in BCLC-B/UT-7 out u-HCC patients treated with TACE alone. Although median observation period of the present cohort was a half year, it is expected that Atez/Bev might behave as better therapeutic option in BCLC-B UT-7 out u-HCC patients than those treated with TACE alone. Moreover, Kudo et al.30 and Tada et al.31 found that the prognosis of BCLC-B stage u-HCC patients classified beyond UT-7 out and initially treated with lenvatinib was better as compared to those treated with TACE. Although Kudo mentioned that hepatic function was maintained in patients treated with lenvatinib in the study,30 there have been few reports which compared relative changes of hepatic reserve function between Atez/Bev and TKI monotherapy. Based on our results (Figure S2c), Atez/Bev might be expected to further improve the prognosis of Child-Pugh A patients, especially those with mALBI grade 1. Although our data did not show significant differences statistically in therapeutic response and mALBI grade between first- and later-line use groups, larger ORR of first-line use group was observed in the present BCLC-B cohort. It was thought to be one of the reasons that percentage of patients with mALBI grade 1 was larger in first-line use group. Further examination to compare the therapeutic efficacy between TKI and Atez/Bev and to elucidate clinical features of patients for obtaining good responses in each therapy for UT-7 out BCLC-B HCC will be needed.

Sequential treatment with TKIs or anti-VEGFR is characterized as “attrition warfare” and good liver function at the time of induction (mALBI 1/2a) has been reported to be a requirement to increase the likelihood of post progression treatment and to obtain a better prognosis with sequential therapy,32-38 even in patients with Child-Pugh A. Using time-dependent receiver operating characteristic analysis, Tada found that mALBI grade is a more useful assessment tool for hepatic function than Child-Pugh.39 To consider increasing chances of post progression treatments with TKI or anti-VEGFR after Atez/Bev, introducing systemic treatments including Atez/Bev at better hepatic reserve function (at least mALBI 2a) as possible, is important.

In a recent report by Terashima, prolonging post-progression survival (PPS) of u-HCC patients was shown to contribute to improve OS, and better hepatic function made making the percentage of PPS greater in OS (Child-Pugh A:B = 54.4 ± 17.6% vs. 32.0 ± 11.6%, p = 0.015).40 To prolong PPS, sequential treatment with multiple drugs has an important role and maintaining hepatic function during systemic treatment is required for increasing the chance of introducing post-progression treatment. It has also been recently reported that hepatic reserve function can be maintained during Atez/Bev treatment.41 The reason for this is thought to be because both of those drugs are antibody agents, which have less impact on hepatic function deterioration as compared to TKI drugs.42, 43 At the time of writing, this phenomenon temporally decline of hepatic function at 3W cannot be explained exactly, because there were no significant differences in AEs (e.g., appetite loss) between patients with and without decline of ALBI score (data not shown). The present findings of favorable therapeutic response and less influence on hepatic function in patients who received Atez/Bev are encouraging and have led us to use that as first-line treatment for u-HCC including UT-7 out/multiple/Child-Pugh A u-HCC. Moreover, they indicated that early therapeutic response was similar between first- and later-line use. Clinical practice data of post progression treatment transition rate after Atez/Bev failure has already showed favorable rate (30/34: 88.2%) also in Japan,44 as well as that from Asian countries other than Japan (49/66: 74.2%).45 It is expected that the use of Atez/Bev, which can easily be transferred to sequential treatment including TKIs and anti-VEGFR, prior to TACE as an initial treatment might prolong the prognosis of BCLC-B u-HCC patients with good hepatic function, but further accumulation of clinical data and trials will be needed to establish an effective treatment strategy.

When PR is obtained with Atez/Bev in UT-7 out/multiple/Child-Pugh A u-HCC patients, a possibility of curative conversion treatment with locoregional therapy (ABC-conversion therapy28) should be considered for obtaining drug-free survival. Of course, it is difficult to challenge ABC-conversion in those patients who show PD, especially with the appearance of extra hepatic metastasis, sequential treatments after Atez/Bev failure is also important clinically.

Although favorable therapeutic responses were observed in the present cohort, this study has some limitations. The observation period was not long enough for evaluation of OS, and the number of patients enrolled was small. Furthermore, there is nearly no information regarding the therapeutic effects of post-progression treatments after Atez/Bev presented in this report. For determining the best order of TKIs and anti-VEGFR after Atez/Bev failure, real-world data accumulation must be continued and a randomized trial (Atez/Bev vs. TACE or sorafenib/lenvatinib) for u-HCC patients classified as UT-7 out/multiple/Child-Pugh A is needed for obtaining solid evidence.

In conclusion, Atez/Bev showed favorable therapeutic response with less influence on hepatic function. In this multiple systemic therapies' era, Atez/Bev might provide a useful therapeutic efficacy in patients that are UT-7 out/multiple u-HCCs.

The present study protocol was granted approval by the Institutional Ethics Committee of Ehime Prefectural Central Hospital (IRB No. 30–66) (UMIN000043219).

CONFLICTS OF INTEREST

Atsushi Hiraoka, MD, PhD: lecture fees; Bayer, Eisai, Eli Lilly, Otsuka.

Takashi Kumada, MD, PhD: lecture fees; Eisai.

Masatoshi Kudo, MD, PhD: advisory role: Eisai, Ono, MSD, Bristol-Myers Squibb, Roche. Lecture fees; Eisai, Bayer, MSD, Bristol-Myers Squibb, Eli Lilly, EA Pharma. Research funding; Gilead Sciences, Taiho, Sumitomo Dainippon Pharma, Takeda, Otsuka, EA Pharma, Abbvie, Eisai.

None of the other authors have potential conflicts of interest to declare.