Brain hemorrhage in patients with advanced hepatocellular carcinoma treated with atezolizumab and bevacizumab: two case reports
Article information
Abstract
A combination of atezolizumab and bevacizumab is currently recommended for treating unresectable advanced-stage hepatocellular carcinoma (HCC), as it has demonstrated superior overall survival and progression-free survival compared to sorafenib. However, concerns have been raised regarding serious adverse events associated with bevacizumab, such as gastrointestinal perforation, fistula, hemorrhage, and arterial thromboembolism. In particular, patients with liver cirrhosis (LC) show an increased risk of variceal bleeding. However, brain hemorrhage associated with the use of bevacizumab in patients with HCC and LC is extremely rare. We encountered two cases of brain hemorrhage in patients with HCC and LC who underwent treatment with atezolizumab and bevacizumab. One patient had no history of hypertension, while the other patient had hypertension that was well-controlled with medication and an unruptured brain aneurysm located on the right side of the anterior communicating artery. Both patients experienced brain hemorrhage after two treatment cycles of atezolizumab with bevacizumab. One patient died due to brain hemorrhage, while the other patient recovered from subarachnoid hemorrhage with successful coil embolization. This case report suggests that if a patient has any high-risk factors associated with brain hemorrhage, physicians should thoroughly consider alternative treatment options for advanced HCC, as brain hemorrhage could be fatal.
Introduction
Hepatocellular carcinoma (HCC) is the sixth most frequent cancer and the third leading cause of cancer-related deaths worldwide [1]. The overall survival of HCC is poor, with a median survival of 6–10 months [2]. Currently, in the unresectable advanced stage, the combination of atezolizumab and bevacizumab is the first-line treatment recommended by the Barcelona Clinic Liver Cancer (BCLC) staging system [3]. It is well-known that the median overall survival and progression-free survival of atezolizumab plus bevacizumab are superior to those of sorafenib in patients with unresectable advanced HCC [4]. Bevacizumab, a recombinant monoclonal antibody targeting vascular endothelial growth factor (VEGF), has been widely used for treating various cancers, including colorectal cancer, non-small cell lung cancer (NSCLC), breast cancer, glioblastoma, and renal cell carcinoma in combination with other agents [5]. In patients with liver cirrhosis (LC), the risk of variceal bleeding increases when bevacizumab is used [6]. However, cases of brain hemorrhage associated with bevacizumab in HCC and LC patients are known to be extremely rare [4]. We recently encountered two cases of brain hemorrhage in patients with HCC and LC who underwent treatment with atezolizumab plus bevacizumab.
Cases
Ethical statements: This study was exempted from review by the Institutional Review Board (IRB) of Kosin University Gospel Hospital (IRB No. 2024-03-004), and the requirement for the patient's informed consent was waived.
1. Case 1
A 70-year-old male patient presenting with stupor mental status was admitted to the emergency room. The patient had previously been diagnosed with alcoholic LC and HCC 5 years ago. There was no history of hypertension or diabetes mellitus. Upon initial diagnosis, HCC was identified as a solitary lesion located in the S2 region, measuring approximately 2.3×2.5 cm. There was no evidence of vascular invasion, bile duct invasion, or distant metastasis. The patient decided to undergo treatment for HCC through transcatheter arterial chemoembolization (TACE). Subsequently, due to the presence of a residual viable portion and recurrent HCC, five sessions of TACE were additionally administered 5 years after the initial treatment. Nevertheless, multiple HCCs were developed in the both hepatic lobes without evidence of distant metastasis (Fig. 1). Laboratory findings at that time were as follows: hemoglobin level (14.0 g/dL), white blood cell (WBC) count (3,900/mm3), platelet count (79,000/mm3), total bilirubin level (1.38 mg/dL), albumin level (3.3 g/dL), alanine transaminase level (47 U/L), aspartate aminotransferase level (17 U/L), prothrombin time (14.9 seconds), and international normalized ratio (INR) of 1.15. Upper endoscopy revealed no esophageal or gastric varices. The patient’s liver function was categorized as Child-Pugh class A with Eastern Cooperative Oncology Group (ECOG) performance status of 0. Upon diagnosis of TACE-refractory multiple HCCs, the patient underwent two cycles of intravenous administration consisting of 1,200 mg of atezolizumab plus 15 mg/kg of bevacizumab every 3 weeks. There were no significant interval changes in systolic or diastolic blood pressure before and after atezolizumab plus bevacizumab treatment. Two days after completing the second cycle of the atezolizumab plus bevacizumab therapy, the patient presented to the emergency room with a stupor mental status. Brain computed tomography (CT) revealed intracranial hemorrhage in the right thalamus and extensive intraventricular hemorrhage (Fig. 2). Despite receiving best supportive care, the patient expired the following day.
Abdominal computed tomography scan images in the arterial phase (A, B), abdominal magnetic resonance images at arterial phase (C, D) of the patient in Case 1 before treatment with atezolizumab and bevacizumab. (A, B) Multiple variable-sized enhancing nodules (arrows) and (C, D) multiple variable-sized high-density nodules (arrows) were observed.
Nonenhanced axial brain computed tomography scan images of the patient in Case 1 on the day of presentation to the emergency room. High-density lesion measuring 3×2 cm (arrows) in the right thalamus area (A) and extensive high density of the right ventricle (arrows) (B) were observed.
2. Case 2
A 67-year-old female patient presented to the emergency room with a severe headache. The patient had a history of LC with chronic viral hepatitis B diagnosed 10 years ago and HCC diagnosed 2 years ago. The patient was receiving entecavir for chronic viral hepatitis B and a calcium channel blocker, an angiotensin receptor blocker, and a beta-blocker for hypertension, with well-controlled blood pressure. The patient had no history of alcohol or smoking. The patient’s HCC was initially identified as a solitary lesion located in the S6 region, measuring approximately 2.6×1.3 cm. There was no evidence of vascular invasion, bile duct invasion, or distant metastasis. Subsequently, the patient underwent an S6 segmentectomy. However, 1 year after the operation, a newly developed enhancing mass with delayed wash-out in the S6 region (1.7 cm) with enlargement and hypermetabolism of the portocaval lymph node (LN) was detected. The patient then received TACE for the S6 HCC and conventional radiotherapy (5,000 cGy) for the portocaval LN. However, at 6 months after combined treatment with TACE and radiotherapy, CT and positron emission tomography-CT revealed enlargement and hypermetabolism of the left supraclavicular LN, right anterior costophrenic LN, and right retrocrural LN with no viable portion in the liver (Fig. 3). Laboratory findings at that time were as follows: hemoglobin level (11.2 g/dL), WBC count (3,220/mm3), platelet count (162,000/mm3), total bilirubin level (0.39 mg/dL), albumin level (4.3 g/dL), aspartate aminotransferase level (29 U/L), alanine aminotransferase level (11 U/L), prothrombin time (14.8 seconds), and INR (1.13). Gastric antral vascular ectasia was observed; however, no esophageal or gastric varices were noted on the upper endoscopy. The patient’s liver function was categorized as Child-Pugh class A with ECOG performance status of 0. Upon diagnosis of stage C according to the BCLC staging system and stage IVA according to the modified Union for International Cancer Control staging system, the patient was scheduled to receive intravenous administration of 1,200 mg of atezolizumab plus 15 mg/kg of bevacizumab. Prior to treatment initiation, a three dimensional angiographic brain CT for brain assessment revealed a 2.6 mm-sized unruptured vascular outpouching aneurysm located on the right side of anterior communicating artery. Throughout two cycles of atezolizumab plus bevacizumab therapy, there had been no side effect. There were no significant interval changes in systolic or diastolic blood pressure before and after atezolizumab plus bevacizumab treatment while the patient was on antihypertensive medication. However, the day before the scheduled third cycle, the patient was admitted to the emergency room due to a severe headache. Initial brain CT showed no evidence of significant bleeding. However, a follow-up brain magnetic resonance imaging 1 day later revealed evidence of suspected subarachnoid hemorrhage (Fig. 4), leading to successful emergent coil embolization (Fig. 5). Two months following coil embolization, the patient was prescribed a 2-month period of sorafenib for treating the HCC. Unfortunately, progression of HCC was observed upon restaging CT. Subsequently, the patient was administered regorafenib for a duration of 3 months. However, the patient was expired due to hepatic failure.
Axial positron emission tomography-computed tomography scan images of the patient in Case 2 before treatment with atezolizumab and bevacizumab. (A) A hypermetabolic lesion (arrow) was observed in the left supraclavicular area. (B) A hypermetabolic lesion (arrow) was observed in the right retrocrural space.
Axial brain computed tomography (CT) and magnetic resonance imaging (MRI) of the patient in Case 2. (A) There was no evidence of brain hemorrhage in the nonenhanced CT scan. (B) Slightly high-density lesions (arrows) were observed in the subarachnoid space, predominantly on the left side, in T2-weighted fluid-attenuated inversion recovery MRI scan image.
Brain angiography images of the patient in Case 2. (A) A saccular aneurysm (arrows) was observed on the right anterior communicating artery. (B) Following coil embolization, occlusion of the aneurysm (arrows) was successful with preserved flow in the parent artery.
Discussion
Bevacizumab primarily suppresses tumors by inhibiting angiogenesis. It binds to circulating VEGF-A isoforms, blocking their interaction with receptors and thereby inhibiting VEGF signaling critical to neovascularization. In vivo studies showed bevacizumab suppresses vascular growth, promotes regression of new vessels, and normalizes vasculature, especially when combined with cytotoxic chemotherapy [5,7-10].
Despite its efficacy, bevacizumab is associated with adverse events such as hypertension, fatigue, diarrhea, abdominal pain, anorexia, epistaxis, pruritus, neutropenia, and thrombocytopenia [5]. Hypertension often occurs dose-dependently but is usually manageable with antihypertensive drugs. Serious adverse events include gastrointestinal perforation and fistula, hemorrhage, and arterial thromboembolism [5]. Hemorrhagic events may manifest as hemoptysis in NSCLC or variceal bleeding in patients with HCC and LC [5,6]. However, brain hemorrhage is extremely rare in patients with HCC and LC receiving atezolizumab plus bevacizumab. In a phase 3 trial of 329 patients with unresectable HCC treated with this combination, only one experienced a subarachnoid hemorrhage [4]. Real-world data also report few cases of bevacizumab-associated brain hemorrhage. In 2011, Nishimura et al. [11] reported two intracranial hemorrhages (ICH) in colon cancer patients without brain metastasis. Tanvetyanon et al. [12] described a case of ICH from a ruptured arteriovenous malformation in an NSCLC patient treated with bevacizumab and chemotherapy. Nguyen et al. [13] reported two ICH cases in patients treated with bevacizumab and low-molecular-weight heparin.
The relationship between bevacizumab and brain hemorrhage remains controversial. A meta-analysis of 17 randomized controlled trials (12,917 patients) showed a higher risk of central venous system (CNS) hemorrhage with bevacizumab (relative risk, 3.09; 95% confidence interval [CI], 1.36–6.99) [14]. However, another meta-analysis of eight studies (8,713 patients) focusing on ICH in patients with brain metastasis found no significant increase in risk (odds ratio, 1.2; 95% CI, 0.69–20.9; p=0.53) [15]. Furthermore, in a retrospective study involving 17,466 patients who received bevacizumab, only 154 CNS hemorrhage occurred in 99 patients [16]. Among the entire cohort, 30% had a history of hypertension, 16% had CNS metastasis, and 13% had primary glioma. Patients received a median of three doses of bevacizumab before experiencing CNS hemorrhage. Of these patients, 48% died due to CNS hemorrhage. The most prevalent predisposing factor for CNS hemorrhage was bleeding associated with the use of medications (e.g., warfarin, nonsteroidal anti-inflammatory drugs, heparin, clopidogrel), followed by thrombocytopenia and a history of hypertension [16].
For safety, Besse et al. [17] and Carden et al. [18] concluded that bevacizumab should not be excluded in advanced cancer patients, even with brain metastasis, due to no observed increased risk of brain hemorrhage. However, many HCC patients have LC and portal hypertension, often leading to thrombocytopenia and prolonged INR, raising bleeding risk. In our experience, two advanced HCC patients without brain metastasis experienced brain hemorrhage. One had well-controlled hypertension and a known brain aneurysm. Although these events may be coincidental, brain hemorrhages have serious consequences that can drastically affect prognosis. Therefore, careful evaluation is essential when considering atezolizumab plus bevacizumab for HCC with LC. Risk factors—such as uncontrolled hypertension, use of anticoagulants, severe thrombocytopenia, or brain abnormalities like arteriovenous malformations or aneurysms—should be thoroughly reviewed. If present, alternative HCC treatments should be considered. Further research is needed to clarify risk factors and mechanisms linking bevacizumab to brain hemorrhage, especially in HCC and LC patients [19].
Notes
Conflicts of interest
Hyun Joon Park is an associate editor of the journal but was not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.
Funding
None.
Author contributions
Conceptualization: BK, HJP. Data curation: BK, JYN. Investigation: BK, SJ. Methodology: HJP, KIS. Supervision: SUL, KIS, JWL. Writing-original draft: BK, HJP. Writing-review & editing: HJP, SUL, KIS, SJ, JYN. All authors read and approved the final manuscript.