Effectiveness and Safety of Atezolizumab Plus Bevacizumab in Unresectable Hepatocellular Carcinoma: A Multicenter, Retrospective Real-World Study in China

Abstract

Background:

Atezolizumab combined with bevacizumab (Atezo + Bev) has received regulatory approval as a first-line systemic therapy for unresectable hepatocellular carcinoma. This study aimed to evaluate the effectiveness and safety of atezolizumab combined with bevacizumab (Atezo + Bev) in treating unresectable hepatocellular carcinoma (uHCC) in a real-world Chinese population, focusing on both first-line and second-line treatment settings.

Methods:

In this multicenter, retrospective study, patients with uHCC treated with Atezo + Bev were included at 5 centers in China from Jan. 2021 to Jan. 2023. Treatment efficacy was assessed using RECIST 1.1 and mRECIST criteria. Overall response (ORR), disease control rates (DCRs), time to disease progression (TPP), progression-free survival (PFS), and overall survival (OS) were calculated.

Results:

The study included 48 patients, with a median age of 58 years, among which 35 and 13 patients received Atezo + Bev as first- and second-line treatment, respectively. The ORR and DCR were 39.6% and 70.8% according to RECIST 1.1, and 60.4% and 75.0% according to mRECIST, respectively. With a median follow-up of 14.5 months, the median PFS was 8.5 months (95% CI [4.4, 11.2]) in the first-line treatment group, while in the second-line group it was 5.1 months (95% CI [2.1, 7.9]). The median OS was not reached. Adverse events of any grade were observed in 75% of patients (n = 36), most commonly being thrombocytopenia (27.1%), lymphopenia (25%), and abnormal liver function (14.6%).

Conclusion:

This study confirmed the practical efficacy and safety of Atezo + Bev in real-world Chinese patients with uHCC, both as a first-line and second-line treatment. Further studies are warranted to validate these findings and optimize treatment strategies.

Introduction

China stands as a nation with one of the highest incidences of hepatocellular carcinoma (HCC) worldwide, leading the globe in the number of HCC cases. According to an epidemiological survey in 2022, the number of new HCC diagnoses in China in 2022 was 367 700, accounting for 42.5% of the global HCC patient population. ¹ Hepatocellular carcinoma, the predominant form of primary liver cancer, is one of the leading causes of cancer mortality across the world. ² While patients diagnosed in the early stages of HCC can attain favorable prognoses through curative interventions such as liver resection, radiofrequency ablation, and liver transplantation, because of the subtle early symptoms of HCC, many patients have progressed to advanced stages by the time of initial diagnosis. ³ Unfortunately, the prognosis for advanced HCC remains dismal due to the limited treatment options, which contributes to the high mortality rate associated with HCC. ⁴ Therefore, there is an urgent need to explore potential and effective treatment strategies for advanced liver cancer to improve patient outcomes.

In recent years, the latest advances in systemic therapies, including molecular targeted drugs and immune checkpoint inhibitors, have disrupted the traditional treatment pattern for advanced HCC. Historically, sorafenib or lenvatinib have served as the first-line molecular targeted agents, offering modest survival benefits for patients with advanced HCC. ⁵ Previous clinical trials have reported that immune checkpoint inhibitors, including nivolumab ⁶ and pembrolizumab, ⁷ showed the promising anti-tumor efficacy and safety for unresectable HCC. However, the limited tumor response rates and suboptimal survival outcomes associated with monotherapy using either molecular targeted drugs or immune checkpoint inhibitors have underscored the need for more effective treatment strategies. Recently, immune-based combination therapies, such as dual immune checkpoint inhibitors, molecular targeted agents combined with immune checkpoint inhibitors, and immune checkpoint inhibitors combined with anti-angiogenic agents, have exhibited enhanced anti-tumor activity and superior survival outcomes compared to systemic monotherapy. A landmark study, the IMbrave150 trial, evaluated the efficacy of atezolizumab in combination with bevacizumab (Atezo + Bev) as a first-line treatment for patients with unresectable hepatocellular carcinoma (uHCC), has shown that Atezo + Bev demonstrated statistically and clinically significant survival benefits compared to sorafenib for patients with previously untreated HCC. The regimen achieved an objective response rate (ORR) of 30%, a median progression-free survival (PFS) of 6.9 months, and a median overall survival (OS) of 19.2 months. ⁸ Remarkably, in the Chinese cohort of the IMbrave150 trial, the OS extended to an impressive 24 months, ⁹ further underscoring the potential of this combination therapy in improving outcomes for advanced HCC patients, establishing Atezo + Bev as the first first-line combination therapy for HCC.

In real-world clinical practice, the complexity of patient scenarios often deviates from the standardized treatment protocols recommended by guidelines due to factors such as varying tolerance levels, accessibility to treatment regimens, and economic constraints. ¹⁰ There may be cases where first-line treatment regimens are used for second-line or subsequent treatments or in conjunction with localized treatments such as trans-arterial chemoembolization (TACE) or radiotherapy,^11,12 as well as in the treatment of populations excluded from clinical research. Consequently, it is particularly important to understand the efficacy, safety, and treatment patterns of Atezo + Bev treatment regimen in real-world clinical settings. Clinicians across different nations have undertaken explorations and compilations of such data. Another multicenter real-world study involving 51 patients found that only 19 patients (37.3%) received the Atezo + Bev treatment regimen as a first-line treatment, whereas 22, 6, and 4 patients utilized it as second, third, and fourth-line treatments, respectively. ¹³ However, the reported real-world studies of Atezolizumab + Bevacizumab treatment for HCC mainly focus on patients in Japan or Western countries.^14,15 Despite China’s high incidence and mortality rates for liver cancer, there is currently limited real-world evidence for the application of Atezolizumab + Bevacizumab in China, and more local evidence is needed to further provide real-world evidence for the combination treatment of Atezolizumab and Bevacizumab in Chinese liver cancer patients.

Methods

This multicenter, retrospective study has involved several leading medical institutions in China, The study has received approval from the Institutional Review Board on 17 October 2023. The reporting of this study adheres to STROBE guidelines. ¹⁶

Study population

Including the First Affiliated Hospital of Army Medical University, the Second Affiliated Hospital of Army Medical University, the First Affiliated Hospital of Chongqing Medical University, the Second Affiliated Hospital of Chongqing Medical University, and the Chongqing University Affiliated Tumor Hospital. Patients diagnosed with unresectable primary HCC and treated with Atezolizumab combined with Bevacizumab from January 2021 to January 2023 were included in the study. Proper clinical information was available, and the treatment response was evaluated by dynamic CT or MRI at baseline and every 6 weeks. Patients Inclusion criteria: Age 15-75 years, Child-Pugh score 5-8 (Child-Pugh A or B), ECOG performance status 0-1, and At least one measurable lesion per RECIST 1.1. ¹⁷ Patients were excluded if they had concurrent severe comorbidities, including serious cardiovascular and cerebrovascular diseases, severe liver dysfunction (bilirubin > 2× upper limit of normal or transaminases > 5× upper limit of normal), severe renal impairment (eGFR < 45 ml/min), and hematological diseases, had active infectious diseases, such as acute viral hepatitis, Prior liver transplantation, Pregnancy or lactation, Concurrent autoimmune disease requiring systemic immunosuppression, had insufficient clinical data, including incomplete baseline data, follow-up data, or outcome data that would compromise the study’s analytical requirements.

Treatment procedure

Patients with unresectable primary HCC who were enrolled underwent treatment with atezolizumab combined with bevacizumab (Atezo + Bev). All participants signed informed consent for treatment, patients received intravenous injections of atezolizumab (1200 mg) plus bevacizumab (15 mg/kg of body weight) every 3 weeks. The safety profile of the combination therapy was evaluated at 6 weeks following treatment initiation. Treatment was discontinued in cases of unacceptable adverse events (AEs) or upon imaging-confirmed disease progression (PD).

Data collection and follow-up

Clinical baseline characteristics and follow-up data of eligible patients were systematically collected and analyzed through comprehensive medical record review, including gender, age, ECOG PS score, ¹⁸ cause, tumor characteristics (maximum tumor size and number of tumors), serum biomarkers (alpha-fetoprotein [AFP] levels, alanine aminotransferase [ALT], albumin [ALB], aspartate aminotransferase [AST], and total bilirubin [TBIL]), liver function assessment (albumin-bilirubin [ALBI] grade), and disease progression indicators (presence or absence of major vascular invasion, extrahepatic metastasis, and portal vein tumor thrombosis [PVTT]) classification. ¹⁹

Outcomes and assessments

Imaging examinations were evaluated and reviewed independently by two experienced researchers, in accordance with the RECIST 1.1 ¹⁷ and mRECIST criteria, ⁸ and was centrally reviewed by Professor Li Xiaoming, Department of Radiology. The ORR is defined as, number of clinical complete response (CR) patients + number of clinical partial response (PR) patients, / total number of patients × 100%. The disease control rate (DCR) is defined as, number of clinical CR patients + number of clinical PR patients + number of stable disease (SD) patients, / total number of patients × 100%.^20,21 Adverse events and reactions were systematically assessed using the Common Terminology Criteria for Adverse Events version 5.0 (CTCAE 5.0). ²²

Statistical analysis

The measurement data were assessed for normality using the Shapiro-Wilk test. Data following a normal or approximately normal distribution were expressed as mean ± standard deviation (Mean ± SD), while skewed data were presented as median (Median) with interquartile range (IQR) or minimum and maximum values (Min, Max). Categorical data were summarized using frequency (N) and percentage (%). The patient prognosis, including OS, PFS, and time-to-progression (TTP), was evaluated using Kaplan-Meier curves, with median survival times and 95% confidence intervals (CI) calculated. Intergroup comparison was conducted using the Log-rank test. Factors influencing the outcomes of OS, PFS, and TTP were analyzed through univariate COX regression, with variables showing P < 0.1 included in the multivariate COX regression model. All tests were two-sided, with a significance level of α = .05, where P < .05 was considered statistically significant. Statistical analyses were performed using SAS 9.4 (SAS Inc., Cary, N.C., USA).

Results

Patient characteristics

A total of 48 patients diagnosed with uHCC and treated with atezolizumab in combination with bevacizumab (Atezo + Bev) between January 2021 and January 2023 were included in this study (Figure 1). Table 1 summarized the baseline characteristics of the patients before treatment. The median age of the patients was 58 years, with 41 out of 48 patients being male (85.4%). According to the Barcelona Clinic Liver Cancer (BCLC) staging system, 2 (4.2%), 10 (20.8%), and 36 (75.0%) patients were classified as stages A, B, and C, respectively. In terms of liver function, 16 and 32 patients were categorized as ALBI grade 1 and 2, respectively. This study included high-risk patients, defined by the following criteria: tumor occupancy exceeding 50% of the liver volume (18.8%), tumor invasion of the main portal vein trunk and/or the contralateral portal vein branch (Vp4, 2.1%), and bile duct invasion (10.4%). In addition, 26 patients exhibited vascular invasion (54.2%), 15 patients had portal vein thrombus (31.2%), and 16 patients presented with extrahepatic metastases (33.3%). Furthermore, 62.5% of patients exceeded the up-to-7 criteria.

Figure 1.

Flowchart of the study.

Table 1.

Baseline characteristics.

Characteristics	All patients (n = 48)
Age (years), Median (IQR)	58.37 (52.82, 65.99)
< 60	27 (56.2)
⩾ 60	21 (43.8)
Gender, n (%)
Male	41 (85.4)
Female	7 (14.6)
Smoking status, n (%)
Never smoked	24 (50.0)
Long-term smoker	13 (27.1)
Former smoker who has quit	9 (18.8)
Unknown	2 (4.2)
Cause, n(%)
Hepatitis B virus (HBV)	12 (25.0)
Hepatitis C virus (HCV)	2 (4.2)
Hepatitis B virus (HBV) + Alcohol-related liver disease	30 (62.5)
Other/unknown	4 (8.3)
Whether consuming alcohol, n (%)
Alcohol dependence (average daily alcohol consumption > 30 grams)	13 (27.1)
Never consume alcohol	20 (41.7)
Consume alcohol but no alcohol dependence	13 (27.1)
Unknown	2 (4.2)
Initial treatment or recurrence, n (%)
Initial treatment	22 (45.8)
Recurrence	26 (54.2)
Histological classification, n (%)
Histological classification not performed	28 (58.3)
Hepatocellular type	19 (39.6)
Mixed type	1 (2.1)
Number of lesions, n (%)
1	25 (52.1)
2-3	9 (18.8)
>3	14 (29.2)
Portal vein tumor thrombus (PVTT), n (%)
Yes	15 (31.2)
No	33 (68.8)
Vascular invasion, n (%)
Yes	26 (54.2)
No	22 (45.8)
Extrahepatic metastasis, n (%)
Yes	16 (33.3)
No	32 (66.7)
Cirrhosis, n (%)
Yes	28 (58.3)
No	20 (41.7)
Alpha-fetoprotein (AFP), n (%)
> 400μg/L	23 (47.9)
⩽ 400μg/L	24 (50.0)
Missing	1 (2.1)
Eastern Cooperative Oncology Group Performance Status (ECOG PS), n (%)
0	32 (66.7)
1	16 (33.3)
Child-Pugh score, n (%)
A	46 (95.8)
B	2 (4.2)
Albumin-Bilirubin (ALBI) Grade, n (%)
1	16 (33.3)
2	32 (66.7)
BCLC stage, n (%)
A	2 (4.2)
B	10 (20.8)
C	36 (75.0)
CNLC staging, n (%)
I	3 (6.3)
II	9 (18.8)
III	36 (75.0)
Prior treatment, n (%)
Surgery	14 (29.2)
Ablation	8 (16.7)
TACE	27 (56.2)
Radiotherapy	3 (6.3)
Systemic therapy	13 (27.1)
Combination treatment, n (%)
TACE	21 (43.8)
Ablation	5 (10.4)
Nonconcurrent combination therapy	22 (45.8)

IQR, interquartile range; CNLC, staging China Liver Cancer Staging; BCLC stage, Barcelona Clinic Liver Cancer; TACE, transcatheter arterial chemoembolization.

In addition, 13 patients (27.1%) had previously undergone systemic therapy, with Atezo + Bev administered as a subsequent treatment. The median number of Atezo + Bev treatment cycles was 8 cycles, and 26 patients (54.16%) received local therapy (TACE and/or Ablation) concurrently with atezolizumab-bevacizumab. Two BCLC stage A patients received first-line Atezo + Bev: one had a prior hepatectomy with recurrent tumor; the other had a large tumor in a small liver, both leaving < 30 % future remnant. Both were deemed unresectable after thorough multidisciplinary evaluation, prompting systemic therapy.

Treatment efficacy

We evaluated the treatment response to Atezo + Bev using dynamic CT or MRI, based on the RECIST 1.1 and mRECIST criteria. As shown in Table 2, based on RECIST 1.1, complete response (CR) was observed in 7 patients (14.6%) and partial response (PR) was observed in 12 patients (25.0%). Stable disease (SD) was observed in 15 patients (31.2%) and progressive disease (PD) in 11 patients (22.9%), while 3 patients were missing. When assessed by mRECIST, 9 patients achieved CR (18.8%), 20 patients achieved PR (41.7%), 7 patients had SD (14.6%), and 12 patients had PD (25.0%). The ORR and DCR were 39.6% and 70.8%, respectively, assessed by RECIST 1.1, meanwhile 60.4% and 75.0% assessed by mRECIST. The intra- and extrahepatic organ-specific response rate (OSRR) were 59.4% and 62.5%, respectively.

Table 2.

Summary of treatment response.

All patients (n = 48)	RECIST 1.1	mRECIST
CR, n (%)	7 (14.6)	9 (18.8)
PR, n (%)	12 (25.0)	20 (41.7)
SD, n (%)	15 (31.2)	7 (14.6)
PD, n (%)	11 (22.9)	12 (25.0)
NE, n (%)	3 (6.2)	0
ORR (CR + PR), n (%)	19 (39.6)	29 (60.4)
DCR (CR + PR + SD), n (%)	34 (70.8)	36 (75.0)

CR, complete response; PR, partial response; SD, stable disease; PD, progressive disease; NE, not estimated; ORR, objective response rate; DCR, disease control rate; RECIST, Response Evaluation Criteria in Solid Tumors; mRECIST, modified Response Evaluation Criteria in Solid Tumors.

After a median follow-up of 14.5 months, the median PFS was 7.6 months (95% CI [4.4, 9.6]) (Figure 2A). The median OS was not reached (Figure 2B).

Figure 2.

Progression-free survival (A) and overall survival (B) in all patients.

Clinical outcomes in the first-line group and the second-line group

We compared tumor responses between the patients treated with Atezo + Bev as first-line therapy (n = 35) and those as second-line therapy (n = 13), as detailed in Table 3. Among 35 patients who received Atezo + Bev as first-line treatment, 5 patients achieved CR (14.3%), 10 patients achieved PR (28.6%), 11 patients had SD (31.4%), and 6 patients had PD (17.1%) according to RECIST 1.1. The ORR was 42.9% and the DCR was 74.3%. In the second-line treatment group (n = 13), according to RECIST 1.1, 2 patients achieved CR (15.4%), 2 patients achieved PR (15.4%), 4 patients had SD (30.8%), and 5 patients had PD (38.5%). The ORR and DCR in this group were 30.8% and 61.5%, respectively.

Table 3.

Tumor response of patients treated with Atezo + Bev in first-line and second-line therapy.

All patients (n = 48)	RECIST 1.1		mRECIST
	First-line (n = 35)	Second-line (n = 13)	First-line (n = 35)	Second-line (n = 13)
CR, n (%)	5 (14.3)	2 (15.4)	6 (17.1)	3 (23.1)
PR, n (%)	10 (28.6)	2 (15.4)	18 (51.4)	2 (15.4)
SD, n (%)	11 (31.4)	4 (30.8)	4 (11.4)	3 (23.1)
PD, n (%)	6 (17.1)	5 (38.5)	7 (20.0)	5 (38.5)
NE, n (%)	3 (8.6)	0	0	0
ORR (CR + PR), n (%)	15 (42.9)	4 (30.8)	24 (68.6)	5 (38.5)
DCR (CR + PR + SD), n (%)	26 (74.3)	8 (61.5)	28 (80.0)	8 (61.5)

The median PFS assessed by RECIST 1.1 was 8.5 months (95% CI [4.4, 11.2]) in the first-line group and 5.1 months (95% CI [2.1, 7.9]) in the second-line group. Although the median PFS of patients receiving Atezo + Bev as first-line treatment was extended compared to those receiving Atezo + Bev as second-line treatment, the difference did not reach statistical significance (P = .059) (Figure 3).

Figure 3.

PFS after stratification by the treatment line of Atezo + Bev.

Safety and tolerability

Table 4 summarized the safety of 48 patients receiving Atezo + Bev, highlighting the incidence of five common adverse events (AEs). A total of 36 patients (75%) experienced AEs of any grade. The most frequently reported AEs were thrombocytopenia (n = 13, 27.1%), lymphopenia (n = 12, 25%), and abnormal liver function (n = 7, 14.6%). Two cases (4.2%) of grade ⩾ 3 AEs (gastrointestinal bleeding) were observed.

Table 4.

Summary of adverse events.

Events, n (%)	All patients (n = 48)
Any AE, any grade	36 (75.0)
Any AE, grade ⩾ 3	2 (4.2)
Specific AE, any grade (incidence > 5%):
Thrombocytopenia	13 (27.1)
Lymphocytopenia	12 (25.0)
Abnormal liver function	7 (14.6)
Diarrhea	3 (6.3)
Gastrointestinal bleeding	3 (6.3)
Specific AE, grade ⩾ 3:
Gastrointestinal bleeding	2 (4.2)

AE, Adverse Event.

Prognostic factor analysis

The median PFS of patients with baseline AFP levels ⩽400 μg/mL was 8.02 months, and the median OS was not reached. In contrast, for patients with baseline AFP levels > 400 μg/mL, the median PFS was 6.900 months, and the median OS was 21.32 months. However, survival curve comparisons using the Log-rank test showed no statistically significant difference between the two groups for either PFS (P = .6341) or OS (P = .4200) (Figure 4). Using either RECIST1.1 OR mRECIST, the ≤ 400 μg/L subgruop consistently showed numerically higher objective response rates (ORR) and disease-control rates (DCR) than the > 400 µg/L subgroup, Progressive disease was reported more frequently in patients with AFP > 400 µg/L under both criteria. These descriptive data suggest a trend toward better tumor control in the lower-AFP cohort, without implying statistical significance (Table 5).

Figure 4.

PFS and OS after stratification by the baseline AFP levels.

Table 5.

Tumor response of patients treated with Atezo + Bev with different baseline AFP levels (⩽ 400 μg/L vs. > 400 μg/L).

All patients (n = 47)	RECIST 1.1		mRECIST
	AFP ⩽ 400 μg/L (n = 24)	AFP > 400 μg/L (n = 23)	AFP ⩽ 400 μg/L (n = 24)	AFP > 400 μg/L (n = 23)
CR, n (%)	4 (16.7%)	3 (13.0%)	4 (16.7%)	4 (16.7%)
PR, n (%)	6 (25.0%)	6 (26.1%)	10 (41.7%)	9 (39.1%)
SD, n (%)	8 (33.3%)	7 (30.4%)	5 (20.8%)	2 (8.7%)
PD, n (%)	4 (16.7%)	7 (30.4%)	4 (16.7%)	8 (34.8%)
NE, n (%)	2 (8.3%)	0	1 (4.2%)	0
ORR, n (%)	10 (41.7%)	9 (39.1%)	14 (58.3%)	13 (56.5%)
DCR, n (%)	18 (75.0%)	16 (69.6%)	19 (79.1%)	15 (65.2%)

Next, we investigated the association between survival outcomes following Atezo + Bev treatment and baseline ALBI scores, as well as the presence of extrahepatic metastasis. As illustrated in Figure 3, patients with ALBI grade 1 and grade 2 exhibited comparable median PFS of 8.0 months (95% CI [2.4, NA]) and 7.6 months (95% CI [4.4, 10.2]), respectively. The median metastasis-specific PFS was 7.9 months (95% CI [4.1, 10.3]) for intrahepatic metastasis and 7.6 months (95% CI [3.6, 11.2]) for extrahepatic metastasis (Figure 5).

Figure 5.

PFS after stratification by the ALBI grade.

Discussion

In this real-world study we report our experience with atezolizumab plus bevacizumab in Chinese patients with HCC. Only those with complete baseline and follow-up data required for response and safety assessment were included in the analysis. When assessing the efficacy and safety of the treatment regimen, real-world data (RWD) and clinical trial results provide complementary information. Clinical trials are typically conducted under strictly controlled conditions with stringent inclusion and exclusion criteria, allowing for highly reliable data. However, these conditions may not fully capture the complexity and diversity of patient populations encountered in real-world clinical practice. In contrast, RWD is derived from real routine medical records, encompassing a broader patient population and a wider range of variables, thereby providing a more accurate reflection of real-world treatment outcomes. In our study, the median PFS reported was 7.6 months, and the ORR evaluated by RECIST 1.1 was 39.6%, both higher than those reported in the pivotal phase III IMbrave150 trial (median PFS: 6.9 months; ORR: 30%), ⁸ and the Chinese subgroup analysis of the IMbrave150 study (median PFS: 5.7 months; ORR: 24.6%), ⁹ The Chinese subpopulation had a markedly higher prevalence of HBV-related HCC (≈ 85 %) than the global cohort (≈ 48 %), Chronic HBV infection is associated with a more immunogenic tumor microenvironment, potentially enhancing sensitivity to PD-(L)1 blockade. ²³

Real-world evidence often highlights the influence of comorbidities, adherence, and socioeconomic factors that are under-represented in registration trials. In our cohort, 27 % of patients received atezolizumab plus bevacizumab (Atezo + Bev) as second-line therapy; their ORR was 30.8 % and DCR 61.5 % by RECIST 1.1, with a median PFS of 5.1 months (95% CI [2.1, 7.9]). These outcomes are in line with retrospective reports suggesting comparable radiographic activity between treatment-naïve and previously treated patients.^13,24,25 Pivotal trials limited A + B to Child-Pugh A, but real-world data show Child-Pugh B7 patients may still benefit without extra toxicity, ²⁶ Child-Pugh -B8/9 do not, baseline liver function must guide decisions, and Child-Pugh B use warrants prospective validation. Preliminary data from the 2025 ESMO GI-presented TALENTACE study ²⁷ likewise indicate that TACE combined with atezolizumab plus bevacizumab yields a high ORR, aligning with our observation received synchronous or sequential TACE. This pattern needs prospective validation. A Japanese multicenter study including 64 patients—of whom 46 did not meet IMbrave150 criteria—reported an overall ORR of 5.2 % at 6 weeks and 10.0 % at 12 weeks, with DCRs of 82.8 % and 84.0 %, respectively, without significant differences between eligible and noneligible subgroups (P > .05). ²⁸ Grade ⩾ 3 adverse-event rates and treatment-discontinuation frequencies were also similar between these subsets.^8,28 Taken together, the available data suggest that Atezo + Bev retains clinical activity when used beyond the first-line setting, although prospective validation is warranted.

In terms of safety, there are sometimes differences in AE profiles between clinical trial results and RWD. The spectrum of adverse reactions in Chinese patients also presents distinct features, ²⁹ such as a higher incidence of specific types of hepatotoxicity or immune-related side effects. In the IMbrave150 clinical trial, hypertension was a common AE in the Atezo + Bev treatment group (29.8%), ² whereas in this study, the incidence of hypertension was lower, at less than 5%. Regarding grade ⩾ 3 AEs, our study reported an incidence of 4.2% (n = 2), manifested as gastrointestinal bleeding, which is consistent with the reported incidence of grade ⩾ 3 gastrointestinal bleeding in clinical trials (6.4%). The lower AE incidence observed in this study may be attributed to a shorter observation period.

Therefore, greater emphasis should be placed on the application of RWD. It is crucial for developing individualized treatment plans to recognize the differences in patient baselines, efficacy, and safety between clinical trials and the real world. On this basis, further exploration of the specific mechanisms underlying these differences will aid in the development of more targeted therapeutic strategies. Literature reporting long-term follow-up results, as well as this study, have demonstrated that some patients can still achieve significant survival benefits through subsequent treatments after initial treatment failure, indicating the importance of multi-line treatment strategies.^30,31 In our study, we also tried to use Atezo + Bev therapy in patients who progressed on standard frontline or multiple lines of therapy. Current ASCO guidelines state that Atezo + Bev may be considered as second-line therapy in select patients who have progressed on or are intolerant of first-line sorafenib or lenvatinib. ³² At the same time, greater attention should be given to patients’ quality of life. Incorporating patient-reported outcomes (PROs) into the efficacy assessment framework will enable a more comprehensive evaluation of treatment effects.^33,34

Although this study provides valuable information, there are still some limitations and potential biases. First, the diversity of data sources may lead to inconsistencies in data quality, which could affect the reliability of the results. Second, insufficient sample size and representativeness may result in bias in statistical inferences, especially in subgroup analyses. For example, liver function has significant prognostic implications for survival in HCC patients. ³⁵ As one of the key indicators for assessing liver function, a post hoc analysis of the IMbrave150 study suggested that the ALBI grade appears to predict the efficacy of Atezo + Bev and sorafenib. ³⁶ However, in this study, no significant difference in PFS was observed between patients with baseline ALBI grades 1 and 2 receiving Atezo + Bev. This finding may be partly attributed to the limited RWD sample size and also potentially due to confounding from multiple factors. The retrospective nature of RWD limits the ability to determine causal relationships and makes it difficult to eliminate the impact of confounding variables.^37,38 In addition, the follow-up period in this study was relatively short, necessitating the accumulation of further treatment experience and longer-term observation analyses to confirm the real-world evidence for this treatment strategy. ³⁹

Based on the analyses and discussion above, this study provided certain enlightened values for clinical practice and future research. Clinicians should adopt a holistic approach when selecting treatment plans, taking into account factors such as patient age, comorbidities, and lifestyle. ⁴⁰ Furthermore, future clinical research should place greater emphasis on interdisciplinary collaboration, leveraging the advantages of basic research, translational medicine, and clinical practice to drive continuous innovation in treatment strategies. ⁴¹ To enhance the reproducibility and practicality of research, it is recommended to establish standardized data-sharing platforms that facilitate multi-center cooperation and communication. In this way, not only can research efficiency be improved, but the widespread application and promotion of research outcomes can also be ensured. Moreover, there is a need to strengthen the application of emerging technologies and methods, such as artificial intelligence and big data analytics, to improve the accuracy and depth of data analysis.^42,43 Finally, the importance of patient education and engagement should also be emphasized, encouraging patients to actively participate in the treatment decision-making process to collectively improve treatment outcomes and quality of life. ⁴⁴

In short, by thoroughly discussing the above points, we can not only better understand the significance of current research findings but also provide valuable references for optimizing treatment strategies for unresectable HCC. Future research should continue to build on the existing foundation, constantly exploring novel treatment methods and techniques to ultimately achieve the goal of improving patient prognosis.

Limitations and Future Research

The cohort comprised 48 consecutive, retrospectively identified patients treated between January 2021 and January 2023. No formal sample-size calculation was performed, and the dataset was closed in January 2023; regulatory approvals precluded further accrual. Median follow-up was 14 months, shorter than the 27 month update of IMbrave150, rendering OS not reached. Consequently, efficacy estimates may be affected by both patient selection and the limited observation period. Recently presented data (ESMO GI 2025, TALENTACE) ²⁷ suggest that TACE combined with atezolizumab plus bevacizumab produces favorable effective; these findings mirror the real-world outcomes observed here. Prospective trials with predefined inclusion criteria and extended follow-up are required to define the population most likely to achieve sustained benefit.

Conclusions

This study confirmed the practical efficacy and safety of the combination of atezolizumab and bevacizumab (Atezo + Bev) as a first-line treatment in real-world Chinese patients with uHCC and highlighted its potential in a second-line setting within a real-world context. Large-scale studies are planned to further validate these findings and optimize treatment strategies.