- © 2009 by American Society of Clinical Oncology
Phase II Trial of the Combination of Bevacizumab and Erlotinib in Patients Who Have Advanced Hepatocellular Carcinoma
- Melanie B. Thomas,
- Jeffrey S. Morris,
- Romil Chadha,
- Michiko Iwasaki,
- Harmeet Kaur,
- Elinor Lin,
- Ahmed Kaseb,
- Katrina Glover,
- Marta Davila and
- James Abbruzzese
- From the Department of Gastrointestinal Medical Oncology; Department of Biomathematics, Division of Diagnostic Imaging; and Department of Gastroenterology, Hepatology, and Nutrition, The University of Texas M. D. Anderson Cancer Center, Houston, TX.
- Corresponding author: Melanie B. Thomas, MD, Division of Hematology and Oncology, Medical University of South Carolina, Hollings Cancer Center, 86 Jonathan Lucas St, MSC 955, Charleston, SC 29425; e-mail: thomasmb{at}musc.edu.
Abstract
Purpose The study objective was to determine the proportion of patients with hepatocellular carcinoma (HCC) treated with the combination of bevacizumab (B) and erlotinib (E) who were alive and progression free at 16 weeks (16-week progression-free survival [PFS16]) of continuous therapy. Secondary objectives included response rate, median PFS, survival, and toxicity.
Patients and Methods Patients who had advanced HCC that was not amenable to surgical or regional therapies, up to one prior systemic treatment; Childs-Pugh score A or B liver function; Eastern Cooperative Oncology Group performance status 0, 1, or 2 received B 10 mg/kg every 14 days and E 150 mg orally daily, continuously, for 28-day cycles. Tumor response was evaluated every 2 cycles by using Response Evaluation Criteria in Solid Tumors Group criteria. A total of 40 patients were treated.
Results The primary end point of PFS16 was 62.5%. Ten patients achieved a partial response for a confirmed overall response rate (intent-to-treat) of 25%. The median PFSevent was 39 weeks (95% CI, 26 to 45 weeks; 9.0 months), and the median overall survival was 68 weeks (95% CI, 48 to 78 weeks; 15.65 months). Grades 3 to 4 drug-related toxicity included fatigue (n = 8; 20%), hypertension (n = 6; 15%), diarrhea (n = 4; 10%) elevated transaminases (n = 4; 10%), gastrointestinal hemorrhage (n = 5; 12.5%), wound infection (n = 2; 5%) thrombocytopenia (n = 1; 2.5%), and proteinuria, hyperbilirubinemia, back pain, hyperkalemia, and anorexia (n = 1 each).
Conclusion The combination of B + E in patients who had advanced HCC showed significant, clinically meaningful antitumor activity. B + E warrant additional evaluation in randomized controlled trials.
INTRODUCTION
Hepatocellular carcinoma (HCC; hepatoma) is the fifth leading cause of cancer death worldwide, and the majority of patients reside in Asian countries.1,2 Most HCC develops in the setting of liver cirrhosis from varied causes, including viral hepatitis, excessive alcohol consumption, hemachromatosis, and obesity and metabolic syndrome that leads to steatosis.3,4 As a consequence of such varied etiologies, HCC is a heterogeneous malignancy with complex carcinogenesis.5,6 Despite advances in early detection, liver transplantation, and liver-directed therapies for patients who have early disease, a majority of patients who have HCC present with advanced disease that is not amenable to curative or regional treatment. HCC is a chemotherapy-resistant tumor, and the median survival for patients who have advanced disease is 6 to 8 months. Despite numerous clinical trials of a wide variety of cytotoxic agents to treat HCC, the dismal survival has remained unchanged.7 Recently, the focus has shifted to the evaluation of biologic or targeted agents in HCC on the basis of the principal molecular mechanisms underlying this disease. The approval of sorafenib for patients who have advanced HCC, on the basis of the significant survival benefit demonstrated in a large phase III trial,8 represents the first meaningful progress in outcome improvement in HCC. However, continued improvement in the treatment of patients who have advanced HCC remains important.
The human liver is a rich source of growth factor production during fetal development. This is also true for the adult liver after trauma,9,10 hepatectomy,11,12 and the liver damage produced by inflammation, fibrosis, and cirrhosis.13–19 Increased growth factor expression, including hepatocyte (HGF), epidermal, vascular endothelial (VEGF), insulin-like, and transforming growth factors, have been implicated in the progression from normal liver to cirrhosis, dysplastic nodules (DN), and overt HCC.12,20–27 HCC is a highly vascular tumor, and increased expression of VEGF ligand and VEGF receptor (VEGFr) have been found in both plasma and HCC tumor tissue, respectively. VEGFr elevation in HCC tissue is associated with poor prognosis.28,29 Membranous epidermal growth factor receptor (EGFR) overexpression, measured by immunohistochemistry, has been identified in a significant majority of HCC tumor specimens in several studies.30–36 Previous clinical trials of erlotonib (E) in the treatment of HCC suggest that the agent may provide disease stabilization.30,36 The importance of growth factors in the initiation, development, and poor prognosis of HCC provides rationale for evaluating combinations of targeted agents in HCC. Bevacizumab (B; Avastin; Genentech BioOncology, South San Francisco, CA) is a monoclonal antibody that binds circulating VEGF-A ligand and is approved by the US Food and Drug Administration for the treatment of metastatic colorectal, non–small-cell lung, and breast cancers. Erlotinib (E; Tarceva; OSI Pharmaceuticals, Boulder, CO) is an oral tyrosine kinase inhibitor that blocks phosphorylation at the intracellular portion of the EGFR transmembrane receptor. Erlotinib is currently approved for the treatment of advanced adenocarcinoma of the pancreas and non–small-cell lung cancer.
PATIENTS AND METHODS
This study was a phase II, open-label, single-arm, single-institution, investigator-initiated trial. The trial was approved by the institutional review board at The University of Texas M. D. Anderson Cancer Center and was conducted in accordance with the principles of the Declaration of Helsinki. All patients signed informed consent. Recruitment notification was posted on www.clinicaltrials.gov.
Eligibility Criteria
Patients who had advanced HCC and whose disease was not amenable to surgical or regional therapies were evaluated for this study. Eligibility criteria included age ≥ 18 years; histologically confirmed HCC; Childs-Pugh A or B cirrhosis; no prior malignancy; and Eastern Cooperative Oncology Group performance status of 0, 1, or 2. Organ function requirements included hemoglobin ≥ 10 gm/dL; absolute peripheral granulocyte count ≥ 1,500 mm3; serum albumin ≥ 2.5 gm/dL; platelet count ≥ 40,000 mm3; ALT and AST levels up to five times the institutional normal limits; bilirubin ≤ 2.0 gm/dL; and prothrombin time prolonged not more than 3 seconds greater than institutional normal limits. Patients with New York Heart Association Class II or greater heart failure or uncontrolled dysrhythmia were ineligible. Prior allowed treatments included surgery, transarterial chemoembolization (TACE), one systemic therapy regimen (excluding EGFR- or VEGF-targeted agents), and external-beam radiotherapy. Patients who had a history of significant gastrointestinal bleeding that required intervention within the prior 3 months were ineligible. After two episodes of gastrointestinal bleeding, the protocol was modified to require screening of all potential patients who had any evidence of portal hypertension by using PillCam ESO (Given Imaging ltd, Yokneam, Israel).37 Patients found to have esophageal varices on screening examination underwent esophagogastroduodenoscopy and were treated with banding or sclerotherapy. If follow-up esophagogastroduodenoscopy showed successful treatment of varices, the patient then was eligible for trial participation. Seven of 40 patients were screened with PillCam ESO (manufacturer name); one patient was ineligible because of the presence of gastric varices, because these lesions are difficult to treat.
Treatment and Dose Modification
All patients were treated with B 10 mg/kg on day 1 and day 15 of each 28-day cycle, and all received E 150 mg orally once daily. If a grade 3 or 4 National Cancer Institute Common Toxicity Criteria adverse event occurred, E was held until resolution of the event to baseline. Dose reductions of E were allowed to 100 mg or 50 mg per day. No dose re-escalation was allowed. No dose reductions of B were allowed; however, B could be held or discontinued for grade 3 or higher drug-related adverse events, such as bleeding, hypertension, proteinuria, thrombotic events, bowel obstruction, or wound healing complications. In the event of grade 3 or higher drug-related toxicity, treatment with either agent was held for up to 21 days until resolution to grade 1 or less. Treatment of common adverse effects seen with this regimen, such as hypertension, acne, dry skin, diarrhea, and anorexia, was allowed with standard interventions.
Disease Assessment
Objective response to therapy was assessed by using the Response Evaluation Criteria in Solid Tumors criteria.38 Three-phase computed tomography (CT), or MRI if the patient could not tolerate iodinated contrast material, was used for tumor assessment. Measurable disease was defined by the presence of at least one lesion in which the longest diameter was ≥ 10 mm. All other lesions less than 10 mm were considered nonmeasurable disease. A sum of the longest dimension (LD) of each lesion was calculated and reported as the baseline sum LD. All measurable lesions identified at baseline were observed on reassessment, and response was documented according to Response Evaluation Criteria in Solid Tumors criteria. Partial response (PR) was defined as a decrease of at least 30% in the sum of the LDs of the target lesions compared with baseline. The appearance of new lesions or a 20% increase in the sum of the LDs of the target lesions compared with baseline was considered progressive disease (PD). Tumor response (stable disease [SD], PR, or complete response) was assessed radiographically every 8 weeks. Patients were re-evaluated for disease status 4 weeks after achieving a PR or complete response to confirm the response. Patients were observed until death or until a maximum of 3 years post-enrollment. All tumor measurements were performed by the investigator and were reviewed by a collaborating diagnostic radiologist.
Statistical Methods
The primary end point was the dichotomous end point of progression-free survival after 16 weeks of therapy (PFS16). Our primary objective was to estimate the proportion of patients receiving B + E who were alive and progression free after 16 weeks of therapy (π BE). Previous prospective trials of existing cytotoxic chemotherapy in HCC had shown the median progression-free survival (PFSevent) to be approximately 16 weeks.39 Because objective tumor responses were not expected with the use of biologic agents in HCC, the combination of B + E was considered potentially active if the PFS16 was comparable to that of historical controls. All study end points, including survival, PFS16, PFSevent, and overall survival (OS), were defined as the time from initiation of therapy until documented disease progression or death. The design of Thall and Simon was used. We estimated that the distribution of the probability (πs) of PFS at 16 weeks for patients on standard treatments has a mean of 0.45 and a 95% prior probability interval of 0.30 to 0.60 (ie, Prob[0.30 less than πs less than 0.60] = 0.95). To protect against potential futility, the trial was planned to stop for lack of efficacy if the number of patients alive and progression free at 16 weeks was less than or equal to the following bounds: zero of five, one of 10, two of 14, three of 18, four of 22, five of 26, six of 31, seven of 34, or eight of 38. The data were analyzed on a near-continuous basis to determine whether the stopping criteria were met. A sample size of 40 patients assured that our 95% credible interval on πbev + erl would have a width of no more than 0.30.
Secondary objectives included response rate, median PFS and OS, toxicity, and tolerability. Data also were analyzed to ascertain whether there was any correlation of response with any patient or treatment characteristics. Wilcoxon rank sum tests were used to assess the difference in continuous variables between groups; χ2 tests or Fisher's exact tests were used to assess the association between categoric variables. Univariate logistic regression models were used to estimate the relative rate of responses and the 95% confidence intervals. Kaplan-Meier methods were applied to generate OS and PFS curves. In addition, univariate Cox proportional hazards models were fit to evaluate the association between patient characteristics and OS or PFS. All computations were carried out by using SAS (SAS Institute, Cary, NC).
RESULTS
Patient Characteristics
A total of 40 patients who had advanced HCC were enrolled at The University of Texas M. D. Anderson Cancer Center. Thirty-five patients completed at least 8 weeks of therapy, but the analysis was performed on an intent-to-treat basis; thus, all 40 patients were considered in final analyses. Table 1 lists the characteristics of the enrolled patients. The median age was 64 years; 31 patients (77%) were men (77%); 27 patients (67.5%) had pathologic or radiographic evidence of cirrhosis; 25 patients (62%) were white; 35 patients (87.5%) had Childs-Pugh A liver function; eight patients (20%) had prior systemic therapy; 19 patients (47.5%) had a Cancer of the Liver Italian Program (CLIP) score of 0 to 2; and 21 patients (52.5%) had a CLIP score of 3 to 4. The CLIP score is a prognostic system that has been prospectively validated in several studies and that assigns points for the following: Childs-Pugh score, tumor morphology (solitary, ≤ 50% of the liver, massive), serum α-fetoprotein, and presence or absence of portal vein thrombus (PVT).41 A majority (65%) of the patients had tumors that were advanced stage C on the basis of the Barcelona Clinic Liver Cancer staging classification.40
Patient Characteristics
Response and Survival
At the 16th week, 25 patients (62.5%) were alive and progression free, which was significantly greater than the null-hypothesized value of 45% (P = .022). The best tumor responses for all 40 patients were as follows: 10 patients had confirmed PRs for an overall response rate of 25%; 17 had SD or minor response; and three patients progressed by 8 or 16 weeks (Table 2). Reasons for withdrawal from the study prior to week 8 included rapid PD (n = 1), withdrawal of consent (n = 2), gastrointestinal hemorrhage (n = 1), and clinical decline caused by underlying liver disease (n = 1). All patients who achieved PR underwent confirmatory imaging 4 weeks after the PR. The median PFS was 39 weeks (95% CI, 26 to 45 weeks) or 9.0 months; the median OS was 68 weeks (95% CI, 48 to 78 weeks) or 15.7 months (Fig 1). In univariate Cox proportional hazards models, acneiform rash grade 2 or greater (P = .024), Childs-Pugh score (B v A; P = .004), and response at 16 weeks of therapy (PD v PR + SD; P = .037) were significantly associated with survival. Radiographic images depicting examples of confirmed PRs for three patients are shown in Figures 2 to 4.
Summary of Efficacy End Points
Kaplan-Meier estimates of overall survival and progression-free survival. OS = overall survival; PFS = progression free survival.
Representative images of partial tumor responses. (A) Baseline computed tomography (CT) of the abdomen of 54-year-old male with hepatitis B virus; (HBV), hepatitis c virus; (HCV), cirrhosis, large hepatocellular carcinoma (HCC) in right lobe liver. Serum alpha fetoprotein (AFP) 323,205 ng/mL. (B) Restaging CT abdomen after 8 weeks bevacizumab and erlotinib (B + E), partial response. Serum AFP 1,167 ng/mL. CT abdomen obtained at 12 confirmed PR; serum AFP 188 ng/mL.
Computed tomography of the abdomen of a 63-year-old woman with alcohol-related cirrhosis, hypervascular hepatocellular carcinoma in right lobe liver. (A) Baseline; (B) decreased tumor vascularity, increased necrosis after 8 weeks and (C) 16 weeks of bevacizumab and erlotinib.
Computed tomography of the abdomen of a 76-year-old man with multifocal hypervascular (A) hepatocellular carcinoma in right lobe liver, baseline. (B) Decreased tumor vascularity, partial tumor response after 6 months bevacizumab and erlotinib.
Toxicity and Dose Modification
All patients were assessable for toxicity. Adverse events are listed in Table 3. Nine patients each required one dose reduction of E (150 mg to 100 mg daily), and two patients each required two dose reductions (to 50 mg daily) of E. Reasons for E dose reductions included fatigue, anorexia, and acne. Seven patients were removed from the study for grade 3 or greater drug-related toxicity, including fatigue (n = 1), proteinuria (n = 2), delayed wound healing (n = 1), and gastrointestinal hemorrhage (n = 5). Other reasons for discontinuation of therapy included progressive disease (n = 3), withdrawal of consent (n = 4), or disease-related decline in PS (n = 6).
Toxicity Associated With Bevacizumab Plus Erlotinib
Two patients who had known portal hypertension developed life-threatening gastrointestinal hemorrhage while on study. One patient recovered, and the other patient succumbed within 30 days as a result of complications of the bleeding event. One patient had less than grade 1 esophageal varices but was receiving concurrent warfarin at the time of the bleeding event, and the prothrombin time (PT) was significantly prolonged. Two patients had known portal hypertension, but they had poor tolerance of beta blockade therapy and, thus, remained at risk for variceal bleeding. After implementation of PillCam (Given Imaging ltd) screening, which was utilized because of its low cost and low morbidity, there were no additional episodes of variceal hemorrhage.
Overall, the frequency of grade 3 to 4 events in this study was low relative to studies of cytotoxic agents in HCC.42 Several grades 1 to 2 drug-related adverse events were commonly experienced by a majority of patients and represented chronic adverse effects that affected patient quality of life.43 These included dry skin, acne, anorexia, diarrhea, dry mouth/mucositis, and fatigue. These occurred most frequently during the first 1 to 4 months of therapy, and they generally resolved and/or responded to supportive care measures, including oral and topical antibiotics for acne, low-dose steroids for appetite stimulation and fatigue, and antidiarrheal medications.
DISCUSSION
Patients who have advanced HCC that is not amenable to treatment with surgery, local ablation, or regional therapy represent the majority of patients who present with this deadly malignancy.44,45 For several decades, the survival of such patients has remained unchanged, at a median of 6 to 8 months. HCC is unique among other solid-organ malignancies in that it develops in a damaged vital organ that is essential for metabolism of most chemotherapeutic agents. Because of underlying hepatic dysfunction and portal hypertension, it is challenging to safely treat patients who have HCC. Some cytotoxic regimens have shown encouraging response rates, but they have resulted in significant adverse effect profiles that are unacceptable in the palliative treatment setting.42 As several targeted biologic agents have been approved in other common tumors, these drugs are now being studied in HCC. Data is emerging that treatment with biologic agents offers improved survival and reasonable toxicity for patients who have advanced HCC.46,47
Sorafenib, an oral multikinase inhibitor, blocks tumor cell proliferation by targeting the Raf/MEK/ERK intracellular signaling pathway, and it exerts an antiangiogenic effect by targeting VEGFr and platelet-derived growth factor receptor β (PDGF-β)48 An initial large, phase II trial of sorafenib in patients who had advanced HCC showed a median OS of 9.2 months, and 41.6% of patients achieved SD, PR, or minor response for at least 16 weeks.47 Subsequently, a randomized, prospective, double-blind, placebo-controlled trial of sorafenib (400 mg orally twice daily) was conducted in 602 patients who had advanced HCC. The dual primary end points were median OS of 46 weeks versus 34 weeks (P = .00058) and median time to progression (TTP) of 5.5 months versus 2.8 months (P = .000007). Principal treatment-related adverse events included diarrhea (39%), hand-foot skin reaction (21%), alopecia (14%), anorexia (14%), and nausea (11%).8 This significant improvement in survival led to the US Food and Drug Administration approval of sorafenib for the treatment of patients who have advanced HCC, and sorafenib has become the standard of care for systemic treatment of these patients.
This study of B + E is the first report of clinically meaningful results in patients who have advanced HCC that uses a combination of novel agents that modify growth factor signaling. The principal results of this trial included a PFS16 of 62.5%, a confirmed RR of 25%, median PFS of 9.0 months, and median OS of 15.6 months, all of which appear favorable relative to the results from both phase II and phase III sorafenib trials (Table 4). In addition, the patient population in the B + E trial has somewhat poorer prognosis compared with that in the sorafenib trials, as 12.5% of patients in this study had Childs-Pugh B disease, and 27.5% had received prior therapy. As with other single-arm, phase II trials, the limitations of this study include single institution enrollment, small sample size, lack of a control arm, and institutional and investigator selection bias. Patients referred to a tertiary care cancer center may represent a more favorable patient population, particularly in HCC. However, it should be noted that 52% of the patients enrolled on this trial had a CLIP score of 3 to 4. Historically, the median survival for patients who have CLIP scores ≥ 3 is in the range of 1 to 5 months.41,49 Thus, a majority of the patients enrolled were of statistically poor prognosis, yet the median survival for the 40 treated patients was 15.6 months.
Summary of End Points of Sorafenib and B + E in HCC
Are B + E safe in patients who have advanced HCC? As is well documented in published reports in other tumor types, B has known adverse effects, including both minor and major bleeding50,51 and an increased risk of venous and arterial thrombotic events.51,52 bleeding is clearly a concern in cirrhotic patients who may have portal hypertension, gastroesophageal varices, thrombocytopenia, and coagulopathy. Patients who have cirrhosis have an estimated annual risk of variceal bleeding of 5% to 15%, but this varies with the degree of hepatic dysfunction.53,54 In this trial, there were four grade 3 and one grade 4 gastrointestinal bleeding events (12.5%), one of which was fatal. After implementation of PillCam screening for gastroesophageal varices, there were no additional episodes of gastrointestinal bleeding. In addition to active screening for varices, all patients who had evidence of portal hypertension were managed according to current published guidelines.55 There were no exacerbations of HCC-related PVT; in fact, tumor response was seen in several patients with PVT. There were four patients who experienced grades 3 or 4 liver dysfunction (elevated transaminases or bilirubin), but these resolved in all instances with temporary discontinuation of E and a restart at a lower dose.
Although the overall incidence of grades 3 to 4 toxicity is low in this study, there was one on-study death as a result of a gastrointestinal bleeding event. Several chronic, grades 1 to 2 adverse events were common in a majority of patients and required close patient monitoring and support, particularly during the first several months of therapy. Overall, the combination of B + E appears safe and tolerable in patients who have advanced HCC.
In conclusion, the results of this trial suggest encouraging progress in the development of combination growth factor receptor blockade as treatment for HCC. We believe the combination of B + E provides meaningful patient benefit and should be compared with the current standard of care in randomized trials of patients who have advanced HCC.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Melanie B. Thomas, Genentech (C) Stock Ownership: None Honoraria: None Research Funding: Melanie B. Thomas, Genentech Expert Testimony: None Other Remuneration: None
AUTHOR CONTRIBUTIONS
Conception and design: Melanie B. Thomas, James L. Abbruzzese
Administrative support: Michiko Iwasaki
Provision of study materials or patients: Melanie B. Thomas, Michiko Iwasaki, Ahmed Kaseb, Katrina Y. Glover, Marta Davila
Collection and assembly of data: Melanie B. Thomas, Jeffrey S. Morris, Romil Chadha, Michiko Iwasaki
Data analysis and interpretation: Melanie B. Thomas, Jeffrey S. Morris, Romil Chadha, Harmeet Kaur, E. Lin
Manuscript writing: Melanie B. Thomas, Jeffrey S. Morris, James L. Abbruzzese
Final approval of manuscript: Melanie B. Thomas, Jeffrey S. Morris, Harmeet Kaur, Ahmed Kaseb, Katrina Y. Glover, Marta Davila, James L. Abbruzzese
Footnotes
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
-
Clinical Trials repository link available on JCO.org.
- Received May 27, 2008.
- Accepted August 19, 2008.
