Phase II Trial of Bevacizumab Plus Gemcitabine in Patients With Advanced Pancreatic Cancer
http://www.100md.com
《临床肿瘤学》
the Section of Hematology/Oncology, University of Chicago Medical Center
Cancer Research Center
Department of Health Studies
Department of Radiology, University of Chicago, Chicago
Evanston Hospital, Evanston
Ingalls Hospital, Harvey, IL
Fort Wayne Medical Center, Fort Wayne
Northern Indiana Cancer Research Consortium, South Bend, IN
ABSTRACT
PURPOSE: Vascular endothelial growth factor (VEGF) plays a key role in the biology and prognosis of pancreatic cancer. Inhibitors of VEGF suppress the growth of pancreatic cancer in preclinical models. The objectives of this phase II study were to assess the response rate and overall survival of pancreatic cancer patients who received gemcitabine with the recombinant humanized anti-VEGF monoclonal antibody bevacizumab.
PATIENTS AND METHODS: Patients with previously untreated advanced pancreatic cancer received gemcitabine 1,000 mg/m2 intravenously over 30 minutes on days 1, 8, and 15 every 28 days. Bevacizumab, 10 mg/kg, was administered after gemcitabine on days 1 and 15. Tumor measurements were assessed every two cycles. Plasma VEGF levels were obtained pretreatment.
RESULTS: Fifty-two patients were enrolled at seven centers between November 2001 and March 2004. All patients had metastatic disease, and 83% had liver metastases. Eleven patients (21%) had confirmed partial responses, and 24 (46%) had stable disease. The 6-month survival rate was 77%. Median survival was 8.8 months; median progression-free survival was 5.4 months. Pretreatment plasma VEGF levels did not correlate with outcome. Grade 3 and 4 toxicities included hypertension in 19% of the patients, thrombosis in 13%, visceral perforation in 8%, and bleeding in 2%.
CONCLUSION: The combination of bevacizumab plus gemcitabine is active in advanced pancreatic cancer patients. Additional study is warranted. A randomized phase III trial of gemcitabine plus bevacizumab versus gemcitabine plus placebo is ongoing in the Cancer and Leukemia Group B.
INTRODUCTION
Pancreatic cancer is the fourth leading cause of cancer death in the United States. It is estimated that 31,860 Americans developed pancreatic cancer in 2004, and 31,270 died as a result of this disease.1 The nearly identical incidence and mortality rates reflect the early distant spread of pancreatic cancer and the inadequacy of current therapies.
Single-agent gemcitabine, the standard chemotherapy for this disease, has only modest activity. In the pivotal randomized study that compared gemcitabine to a weekly bolus of 5-fluorouracil, gemcitabine produced an objective response rate of 5%, a median survival of 5.7 months, and a 6-month survival rate of 46%.2 Although combination chemotherapy with gemcitabine plus cytotoxic agents such as 5-fluorouracil, pemetrexed, irinotecan, oxaliplatin, and cisplatin achieves higher response rates than gemcitabine alone, randomized phase III trials of these combinations have not demonstrated a statistically significant improvement in survival over single-agent gemcitabine.3-7 Novel approaches that are based on the biologic characteristics of pancreatic cancer may be more promising than additional trials of chemotherapy doublets.
Vascular endothelial growth factor (VEGF) plays a key role in the growth and metastasis of many tumors including pancreatic cancer.8 VEGF and its receptors, flt-1 (VEGFR-1) and KDR (VEGFR-2) are coexpressed in pancreatic cancer, which suggests that VEGF could have autocrine effects on pancreatic cancer cells that express VEGF receptors and paracrine effects on microvascular endothelial cells.9-11 In pancreatic cancer cell lines, VEGF stimulates cell growth in a dose-dependent manner.9 In animal models, inhibitors of the VEGF tyrosine kinase, as well as anti-VEGF and anti-KDR antibodies, inhibit growth and angiogenesis associated with pancreatic tumors12-16 and potentiate the tumoricidal effect of gemcitabine.12,13,16 High serum VEGF levels correlate with advanced-stage postoperative disease recurrence and decreased survival in patients with pancreatic cancer.17 Several,18-22 although not all,23 investigators have reported that high tumor VEGF expression also correlates with a worse prognosis in this disease.
Bevacizumab (Avastin, Genentech, South San Francisco, CA) is a recombinant humanized anti-VEGF monoclonal antibody. In a phase III randomized trial in patients with advanced colorectal cancer, the addition of bevacizumab to standard chemotherapy resulted in a significant improvement in response, survival, and progression-free survival.24
We report the results of a multicenter phase II trial of bevacizumab given in combination with gemcitabine for patients with previously untreated advanced pancreatic cancer. The objectives of this study were to determine the objective tumor response rate, overall and progression-free survival rates, and toxicity profile.
PATIENTS AND METHODS
Patients
Eligible patients had histologically or cytologically confirmed pancreatic adenocarcinoma that was not amenable to potentially curative surgery; unidimensionally measurable disease as defined by Response Evaluation Criteria in Solid Tumors (RECIST) criteria25; Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2; and adequate bone marrow (granulocytes 1,500/μL, leukocytes 3,000/μL, platelets 100,000/μL), renal (creatinine 1.5 mg/dL or creatinine clearance 60 mL/min), and hepatic (normal total bilirubin and transaminases 2.5x the upper limit of normal) function. No prior chemotherapy for metastatic disease was allowed; prior adjuvant chemotherapy for radiosensitization was permitted if it did not contain gemcitabine or bevacizumab. Prior radiation therapy was allowed provided that the only sites of measurable disease were not located within the radiation port. Adjuvant therapy must have been completed at least 4 weeks before enrollment. Exclusion criteria included pregnancy or lactation; prior history of bleeding diathesis; tumor involvement of major blood vessels; prior cerebrovascular accident, pulmonary embolus, or deep venous thrombosis; myocardial ischemia, infarction, or uncompensated coronary artery disease within the past 6 months; major surgery within the prior 6 weeks; brain metastasis; active second malignancy; uncontrolled intercurrent illness; urine protein 1+ or 500 mg in 24 hours; or ongoing uncontrolled hypertension. Patients who received thrombolytic agents within the previous month or who required full-dose anticoagulation (except for maintenance of indwelling catheters with International Normalized Ratio < 1.5), daily aspirin, or daily nonsteroidal anti-inflammatory agents were not eligible. All patients provided written informed consent according to federal and institutional guidelines.
Treatment
Gemcitabine 1,000 mg/m2 was given intravenously over 30 minutes on days 1, 8, and 15 of a 28-day cycle. Bevacizumab 10 mg/kg was administered intravenously after gemcitabine on days 1 and 15 of each 28-day cycle. The initial bevacizumab dose was given over 90 minutes, and if no infusion reaction developed, the second dose was given over 60 minutes. If no infusion reaction occurred with the second dose, subsequent doses were given over 30 minutes.
At the initiation of this study, treatment was limited to six cycles of bevacizumab and gemcitabine. Patients who did not have disease progression after six cycles could receive additional single-agent bevacizumab until progression. Subsequent data from an ongoing phase III trial in colorectal cancer (Eastern Cooperative Oncology Group 3200) suggested that bevacizumab had marginal activity as a single agent.26 Therefore, the protocol was amended in October 2002 to permit patients with stable or responding disease to continue on both drugs until disease progression. Treatment was discontinued for progressive disease, unacceptable adverse events, or patient withdrawal of consent.
Dose Adjustments
Dose reduction was based on toxicities on the day of treatment. Adverse effects were graded according to National Cancer Institute Common Toxicity Criteria version 2.0. A cycle was not started until the absolute granulocyte count was more than 1.5 x 109/L and the platelet count was more than 100 x 109/L. Patients who developed either neutropenic fever that required antibiotic therapy or bleeding associated with thrombocytopenia received a 25% dose reduction of gemcitabine for subsequent cycles.
Within a cycle, if the granulocyte count was between 0.5 and 0.999 x 109/L or the platelet count was between 50 and 75 x 109/L on the day of treatment, the dose of gemcitabine was reduced by 25%. The dose was held for a granulocyte count less than 0.5 x 109/L or a platelet count less than 50 x 109/L. Patients with nonhematologic toxicities grades 0 to 2 (and grade 3 nausea/vomiting) received full-dose gemcitabine on days 8 and 15. For grade 3 nonhematologic toxicities other than nausea/vomiting, patients received either 75% of the gemcitabine dose or no treatment at the discretion of the treating physician. The gemcitabine dose was held for grade 4 nonhematologic toxicities.
There were no dose modifications of bevacizumab in this study. If gemcitabine was held at the beginning of a new cycle because of toxicity, the bevacizumab dose was also held until the chemotherapy could be given. If gemcitabine was held on day 15 for hematologic toxicity, bevacizumab was given. Bevacizumab was held for bilirubin or hepatic transaminase elevations of grade 3 or higher and was not resumed until they were grade 1 or less. If the hepatic abnormalities recurred with re-treatment, bevacizumab was discontinued permanently. Patients who developed either new proteinuria or an exacerbation of pre-existing proteinuria underwent a 24-hour urine collection for total protein. Patients who developed more than 2 g of proteinuria/24 hours could not receive additional doses of bevacizumab unless the proteinuria improved to less than 2 g/24 hours. Bevacizumab was also discontinued for grade 4 hypertension, uncontrolled grade 3 hypertension despite maximal medical therapy, grade 3 or 4 bleeding requiring transfusion or hospitalization, or any thrombotic events that required systemic anticoagulation.
Study Evaluations
Pretreatment evaluation included a complete medical history and physical exam, complete blood count and differential, chemistry panel (including liver function tests), prothrombin time/partial thromboplastin time, plasma VEGF level, pregnancy test (in women of childbearing potential), and a computed tomography scan of the chest, abdomen, and pelvis. The VEGF concentration in plasma was determined as described previously by using a quantitative sandwich enzyme immunoassay technique (Human VEGF Immunoassay; R & D Systems, Minneapolis, MN).27
A history and physical exam were performed every 14 days. Laboratory tests including complete blood count and differential, serum chemistries, and a prothrombin time/partial thromboplastin time were performed weekly. A urinalysis for assessment of proteinuria was performed every 14 days, on the days when bevacizumab was to be administered. Computed tomography scans were obtained every two cycles.
Response Criteria and Toxicity
Patients received a minimum of two cycles unless unacceptable toxicity or early progression of disease occurred. Patients were evaluated for response according to RECIST criteria after two successive cycles.26 Confirmatory scans were obtained at least 4 weeks after initial documentation of objective complete or partial response.
Statistical Analysis
The primary end point of this study was the objective response rate (complete response + partial response). The trial was conducted by using a Simon optimal two-stage design28 to test the null hypothesis that the response rate was 10% versus the alternative that it was at least 25%. Twenty-one assessable patients were to be enrolled in the first stage. If two or fewer patients had an objective response, the trial would be terminated for lack of efficacy. Otherwise, an additional 29 patients were to be enrolled, and if eight or more objective responses were observed among the 50 patients, the regimen would be considered worthy of additional evaluation in this disease. This design yields at least .90 probability of a positive result if the true response rate is at least 25% and at least .90 probability of a negative result if the true response rate is at most 10%. In addition, 50 patients would be sufficient to distinguish between a 65% 6-month survival rate and a 45% 6-month survival rate. Progression-free and overall survival were calculated by using the method of Kaplan and Meier.29 Median progression-free and overall survival times and their associated 95% CIs were derived as described by Brookmeyer and Crowley.30
A protocol amendment in October 2002 required that both drugs, instead of bevacizumab alone, be continued after six cycles (months) in patients whose disease was nonprogressing. Additional patients were enrolled until a total of 45 were available for evaluation of the modified treatment plan. Primary analysis of response rate and 6-month overall survival used data from all enrolled patients, because the amended treatment plan had no effect on these end points. Analysis of overall survival and progression-free survival was assessed for the pooled group of 52 patients and then was assessed separately for patients who were enrolled before (n = 7) and after the amendment (n = 45). Results from the former analysis are reported here because there were no significant differences between the two groups of patients.
Because bleeding, thrombosis, nephrotic syndrome, and uncontrolled hypertension were potential concerns with the administration of bevacizumab, these events were monitored closely. Specifically, any grade 3 or 4 toxicity for bleeding, thrombosis, or proteinuria or any grade 4 hypertension were reported immediately, and a stopping guideline developed by Goldman31 was used as a guide for early termination as a result of toxicity. The trial would be halted if there was evidence that the true rate of toxicity was increased from an acceptable level of o = 10% to an upper limit of A = 25%. A stopping rule that provides an level of 4% and a power of 77% requires that early termination be considered if three adverse events occur in the first five or fewer patients, four adverse events in the first eight or fewer patients, five adverse events in the first 14 patients, six adverse events in the first 20 patients, seven adverse events in the first 26 patients, or eight adverse events in the first 33 or fewer patients.
Mean baseline levels of plasma VEGF were compared between responders and nonresponders by using a nonparametric Wilcoxon test. Patients were also divided into two cohorts on the basis of whether their baseline VEGF levels were greater than the median value. Log-rank tests were then performed to compare progression-free and overall survival for these cohorts.
Independent Response Audit
A special response review was held on December 30, 2002, at the request of the Investigational Drug Branch, Cancer Therapy Evaluation Program. There were 15 participants including the principal investigator (H.L.K.), a diagnostic radiologist from the Clinical Center of the National Institutes of Health, and medical oncology and regulatory representatives from the Cancer Therapy Evaluation Program and Genentech.
Sixteen patients were assessable for response at the time of the audit; of these, seven cases considered by the principal investigator to meet RECIST criteria for objective response or prolonged stable disease were reviewed. A brief patient summary, which included patient history, treatment, toxicity, and response to treatment, and a copy of the pathology report were provided by the principal investigator for each case selected for review. All baseline and follow-up scans necessary to verify the claimed response were available at the review meeting. The diagnostic radiologist independently measured all target lesions by using calipers, the percent change was calculated, and these values were compared to those obtained previously by the study team at the University of Chicago.
RESULTS
Patient Characteristics
Fifty-two patients at seven centers in the University of Chicago phase II consortium were enrolled in this trial from November 2001 to March 2004. Patient characteristics are listed in Table 1. Forty-eight percent of the patients were women, and most (98%) had an ECOG performance status of 0 or 1. All the patients had metastatic disease, and 83% had liver metastases.
A total of 250 cycles were delivered (median, 4; range, 1 to 18). Seven patients were treated before a protocol amendment in October 2002 mandated that both agents be continued after cycle six in patients with nonprogressing disease. Of the initial seven patients, two had no disease progression by cycle six and were able to receive single-agent bevacizumab for four and five additional cycles, respectively.
Response to Treatment and Survival
All 52 patients were included in an intention-to-treat analysis of response. There were 11 confirmed partial responses, for an overall response rate of 21% (95% CI, 11% to 35%). The median duration of response was 10 months (range, 4.9 to 12.1 months). The first five responses were confirmed by an independent response audit at the National Cancer Institute. Twenty-four patients (46%) had stable disease as their best response, which was maintained for a median of 6.3 months. Thirteen patients (25%) had progressive disease.
At the time of this analysis, no patients remained on the study treatment, and 45 patients (87%) had died. The median follow-up time for all patients was 8.0 months. The Kaplan-Meier curve for overall survival is presented in Figure 1. The estimated median survival was 8.8 months (95% CI, 7.4 to 9.7 months). The 6-month survival rate was 77% (95% CI, 63% to 86%), and the 1-year survival rate was 29% (95% CI, 17% to 42%). Figure 2 shows the Kaplan-Meier curve for progression-free survival. The median progression-free survival was 5.4 months (95% CI, 3.7 to 6.2).
Toxicity
Toxicities are summarized in Table 2. Hematologic toxicities were similar to those observed with single-agent gemcitabine.2 Only 35% of patients developed grade 3/4 neutropenia, 4% experienced grade 3 anemia, and 8% had grade 3 thrombocytopenia. No patient developed neutropenic fever. The most common nonhematologic toxicities included fatigue, hyperglycemia, nausea, and anorexia. Hepatic transaminase elevations were related principally to disease progression.
Several nonhematologic toxicities could be attributed to bevacizumab. Forty-four percent of patients experienced hypertension, including 19% who developed grade 3 hypertension. In most patients, this was managed with standard oral antihypertensive drugs; however, two patients discontinued bevacizumab because of poorly controlled hypertension. No patient experienced a hypertensive crisis.
Although proteinuria occurred in 36% of patients, it was grade 3 in only 2%, and no patient developed nephrotic syndrome. Seven patients (14%) developed grade 3 or 4 deep vein thromboses or pulmonary emboli, which required discontinuation of bevacizumab according to the protocol. Three of these thromboses occurred in the first cycle of treatment. None were associated with indwelling venous catheters. There were no arterial thromboses.
Bleeding episodes, although frequent, were usually mild and included epistaxis in 23% of the patients and gum bleeding in 4% of the patients. A lethal gastrointestinal bleed occurred in a 79-year-old man. Endoscopic retrograde cholangiopancreatography demonstrated an actively bleeding vessel and extensive ulceration resulting from malignant invasion from the duodenal bulb and ampullary region.
The toxicity monitoring boundary was reached when seven of the first 26 patients developed grade 3 or higher bleeding, thrombosis, and/or proteinuria. The trial was continued with close monitoring, however, and only two additional such toxicities occurred in the remaining 26 patients who were enrolled.
Four patients developed visceral perforations, one of which was fatal. Nonfatal perforations included a Mallory-Weiss tear in the setting of malignant duodenal obstruction, perforation with ulceration at a gastrojejunostomy anastomosis, and perforation at both ends of a sigmoid stent 8 days postprocedure. A 72-year-old man who developed abdominal pain, fever, and peritonitis 13 days after his initial dose of bevacizumab died 5 days later without additional intervention as a result of presumed bowel perforation.
Correlative Studies
Pretreatment plasma VEGF levels were obtained in 42 patients and ranged from 0 to 462 pg/mL (median, 84 pg/mL). Although patients who attained a partial response or stable disease had slightly higher baseline VEGF levels than those with progressive disease (mean ± standard deviation, 101 ± 112 and 140 ± 122 v 82 ± 85 pg/mL, respectively), the difference was not statistically significant (Kruskal-Wallis exact test, P = .23). There was no significant difference in overall survival (log-rank test, P = .20) or progression-free survival (log-rank test, P = .37) between patients whose VEGF levels were above or below the median. Figure 3 illustrates the Kaplan-Meier overall survival curve by plasma VEGF level.
DISCUSSION
Novel agents and approaches are needed for the treatment of advanced pancreatic cancer, a disease that has the briefest survival of any solid tumor. Although single-agent gemcitabine provides only a modest survival benefit, countless classical cytotoxic agents, given alone or in combination with gemcitabine, have not proven to be superior.32 We evaluated the combination of gemcitabine with the anti-VEGF monoclonal antibody bevacizumab in 52 patients with advanced pancreatic cancer and observed an objective response rate of 21%, a median progression-free survival of 5.4 months, a 6-month overall survival rate of 77%, and a median survival time of 8.8 months. These data seem superior to comparable phase II and III data for single-agent gemcitabine2-7,32,33 and suggest that the combination of gemcitabine and bevacizumab may have a role in the treatment of this disease.
Although most of the patients in this study had a good ECOG performance status, all had metastatic disease and more than 80% had liver metastases, both of which are poor prognostic factors. Patients with pancreatic cancer are at an increased risk of developing malignancy-related hypercoagulability.34 Because the safety of administering bevacizumab with full-dose anticoagulation was unknown at the initiation of this study, patients who had venous thromboses that required anticoagulation were excluded. It is quite possible that we introduced a selection bias by excluding these patients, because cancer patients who have experienced thromboembolism may have a worse prognosis.35 However, bevacizumab did not seem to increase the expected rate of grade 3/4 thrombosis.5
Gastrointestinal bleeding is an unfortunate late complication of this disease that can occur when the pancreatic tumor invades the duodenum. It is not possible to ascertain whether bevacizumab exacerbated the ultimately fatal bleeding in our patient whose tumor eroded into his duodenum. We would advise caution, however, before considering this drug in any individual who has tumor invasion of an adjacent organ.
Gastrointestinal perforation is an infrequent, life-threatening complication of bevacizumab that may be a result of the effects of bevacizumab on wound healing.24 In the pivotal phase III trial of bevacizumab in colorectal cancer, Hurwitz et al24 observed a 1.5% rate of perforation in the treatment group and none in the placebo control. The 8% rate of visceral perforation in our study is significantly higher than that in the colorectal study and is likely a reflection of our patient population. Because one patient developed a perforation after a colon stent placement and another after severe vomiting from a duodenal obstruction, it would be appropriate to hold additional bevacizumab in these situations.
Because there have been no dose-finding trials of bevacizumab in pancreatic cancer, the optimal dose of this agent for this disease remains unclear. Although it is reasonable to speculate whether fewer toxicities or alternate efficacy might have been observed had we arbitrarily chosen a lower dose than the 10 mg/kg used in this trial, this cannot be definitively ascertained without additional study. A randomized phase II trial in colorectal cancer suggested that a dose of 5 mg/kg every 14 days was more effective than 10 mg/kg,36 and a randomized phase III trial in that disease confirmed the activity of the 5 mg/kg dose.24 Significant activity, with tolerable toxicity, was subsequently reported in a phase III study in colorectal cancer that used a 10 mg/kg dose.26 In a randomized phase II trial in non–small-cell lung cancer, a dose of 15 mg/kg every 21 days was more active than the 7.5-mg dose, and there were fewer episodes of significant bleeding at the higher dose.37 The efficacy and safety of the 15 mg/kg bevacizumab dose in that disease has been confirmed in a randomized phase III trial.38
Preclinical models suggest a key role for VEGF in the progression and prognosis of pancreatic cancer,8-11 and elevated serum17 and tumor18-22 VEGF levels have been shown to correlate with poor prognosis in patients with pancreatic cancer. Serum VEGF levels may be unreliable, because they are derived principally from platelets, whereas tumor VEGF is thought to enrich the plasma fraction.39 Despite its overall negative prognostic value, higher plasma VEGF levels may correspond to tumors that are more VEGF dependent and thus more sensitive to VEGF inhibition.
We were not able to correlate plasma VEGF levels with clinical outcome in this study. VEGF levels during and after bevacizumab administration could not be measured reliably because bevacizumab interferes with the immunoassay. Although we observed a trend toward higher baseline soluble plasma VEGF levels in patients who had a partial response or stable disease, it did not reach statistical significance. This may be because of the small sample size or our choice of surrogate biomarker.
In patients with rectal cancer, Willett et al40 observed no decrease in plasma, serum, or urine VEGF levels when bevacizumab was administered; however, the number of viable, circulating endothelial and progenitor cells was decreased after bevacizumab treatment.
We conclude that the combination of bevacizumab and gemcitabine is active in patients with advanced pancreatic cancer and that additional study is warranted. The Cancer and Leukemia Group B (CALGB) is currently accruing to a 590-patient, double-blind, placebo-controlled, randomized phase IV trial (CALGB 80303) that compares gemcitabine plus bevacizumab to gemcitabine plus placebo using the doses and schedule used in this phase II study. Angiogenic biomarkers, proteomic profiling, and pharmacogenomic predictors are being evaluated to further elucidate the biology of VEGF in this disease, and a clinical economic analysis will delineate the potential economic impact of this combination.
Acknowledgment
We thank Cynthia Bajda for assistance with data management; Amy Brich, RN, BSN, for assistance with patient care and protocol monitoring; and Kristen Kasza, MS, for assistance with statistical analysis.
NOTES
Supported by National Cancer Institute Grant No. N01-CM-17102.
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, May 31-June 3, 2003, Chicago, IL; 40th Annual Meeting of the American Society of Clinical Oncology, June 5-8, 2004, New Orleans, LA; and the American Society of Clinical Oncology Gastrointestinal Cancer Symposium, January 22-24, 2004, San Francisco, CA.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Tiwari RC, Murray T, et al: Cancer Statistics, 2004. CA Cancer J Clin 54:8-29, 2004
Burris HA, Moore MJ, Andersen J III, et al: Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: A randomized trial. J Clin Oncol 15:2403-2413, 1997
Berlin JD, Catalano P, Thomas JP, et al: Phase III study of gemcitabine in combination with fluorouracil versus gemcitabine alone in patients with advanced pancreatic carcinoma: Eastern Cooperative Oncology Group Trial E2297. J Clin Oncol 20:3270-3275, 2002
Richards DA, Kindler HL, Oettle H, et al: A randomized phase III study comparing gemcitabine + pemetrexed versus gemcitabine in patients with locally advanced and metastatic pancreas cancer. J Clin Oncol, 2004 ASCO Annual Meeting Proceedings (Post-Meeting Edition) 22, 2004 (abstr 4007)
Rocha Lima CM, Green MR, Rotche R, et al: Irinotecan plus gemcitabine results in no survival advantage compared with gemcitabine monotherapy in patients with locally advanced or metastatic pancreatic cancer despite increased tumor response rate. J Clin Oncol 22:3776-3783, 2004
Heinemann V, Quietzsch D, Gieseler F, et al: A phase III trial comparing gemcitabine plus cisplatin vs. gemcitabine alone in advanced pancreatic carcinoma. Proc Am Soc Clin Oncol 22:250, 2003 (abstr 1003)
Louvet C, Labianca R, Hammel P, et al: Gemcitabine in combination with oxaliplatin compared with gemcitabine alone in locally advanced or metastatic pancreatic cancer: Results of a GERCOR and GISCAD phase III trial. J Clin Oncol 23(15):3509-3516, 2005
Korc M: Pathways for aberrant angiogenesis in pancreatic cancer. Mol Cancer 2:8, 2003
Itakura J, Ishiwata T, Shen B, et al: Concomitant over-expression of vascular endothelial growth factor and its receptors in pancreatic cancer. Int J Cancer 85:27-34, 2000
Luo J, Guo P, Matsuda K, et al: Pancreatic cancer cell-derived vascular endothelial growth factor is biologically active in vitro and enhances tumorigenicity in vivo. Int J Cancer 92:361-369, 2001
von Marschall Z, Cramer T, Hocker M, et al: De novo expression of vascular endothelial growth factor in human pancreatic cancer: Evidence for an autocrine mitogenic loop. Gastroenterology 119:1358-1372, 2000
Baker CH, Solorzano CC, Fidler IJ: Blockade of vascular endothelial growth factor receptor and epidermal growth factor receptor signaling for therapy of metastatic human pancreatic cancer. Cancer Res 62:1996-2003, 2002
Solorzano CC, Baker CH, Bruns CJ, et al: Inhibition of growth and metastasis of human pancreatic cancer growing in nude mice by PTK 787/ZK222584, an inhibitor of the vascular endothelial growth factor receptor tyrosine kinase. Cancer Biother Radiopharm 16:359-370, 2001
Bockhorn M, Tsuzuki Y, Xu L, et al: Differential vascular and transcriptional responses to anti-vascular endothelial growth factor antibody in orthotopic human pancreatic cancer xenografts. Clin Cancer Res 9:4221-4226, 2003
Tsuzuki Y, Mouta Carreira C, Bockhorn M, et al: Pancreas microenvironment promotes VEGF expression and tumor growth: Novel window models for pancreatic tumor angiogenesis and microcirculation. Lab Invest 81:1439-1451, 2001
Bruns CJ, Shrader M, Harbison MT, et al: Effect of the vascular endothelial growth factor receptor-2 antibody DC-101 plus gemcitabine on growth, metastasis, and angiogenesis of human pancreatic cancer growing orthotopically in nude mice. Int J Cancer 102:101-108, 2002
Karayiannakis AJ, Bolanaki H, Syrigos KN, et al: Serum vascular endothelial growth factor levels in pancreatic cancer patients correlate with advanced and metastatic disease and poor prognosis. Cancer Lett 194:119-124, 2003
Ikeda N, Adachi M, Taki T, et al: Prognostic significance of angiogenesis in human pancreatic cancer. Br J Cancer 79:1553-1563, 1999
Itakura J, Ishiwata T, Friess H, et al: Enhanced expression of vascular endothelial growth factor in human pancreatic cancer correlates with local disease progression. Clin Cancer Res 3:1309-1316, 1997
Knoll MR, Rudnitzki D, Sturm J, et al: Correlation of postoperative survival and angiogenic growth factors in pancreatic carcinoma. Hepatogastroenterology 48:1162-1165, 2001
Niedergethmann M, Hildenbrand R, Wostbrock B, et al: High expression of vascular endothelial growth factor predicts early recurrence and poor prognosis after curative resection for ductal adenocarcinoma of the pancreas. Pancreas 25:122-129, 2002
Seo Y, Baba H, Fukuda T, et al: High expression of vascular endothelial growth factor is associated with liver metastasis and a poor prognosis for patients with ductal pancreatic adenocarcinoma. Cancer 88:2239-2245, 2000
Ellis LM, Takahashi Y, Fenoglio CJ, et al: Vessel counts and vascular endothelial growth factor expression in pancreatic adenocarcinoma. Eur J Cancer 34:337-340, 1998
Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205-216, 2000
Giantonio BJ, Catalano PJ, Meropol NJ, et al: High-dose bevacizumab improves survival when combined with FOLFOX4 in previously treated advanced colorectal cancer: Results from the Eastern Cooperative Oncology Group study E3200. Proc Am Soc Clin Oncol, 2005 (abstr 2)
Salven P, Manpaa H, Orpana A, et al: Serum vascular endothelial growth factor is often elevated in disseminated cancer. Clin Cancer Res 3:647-651, 1997
Simon R: Optimal two-stage designs for phase II clinical trials. Control Clin Trials 10:1-10, 1989
Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Brookmeyer R, Crowley J: A confidence interval for the median survival time. Biometrics 38:29-41, 1982
Goldman AI: Issues in designing sequential stopping rules for monitoring side effects in clinical trials. Control Clin Trials 8:327-337, 1987
Kulke MH: Advanced pancreatic cancer: Is there a role for combination therapy? Expert Rev Anticancer Ther 3:729-739, 2003
Casper ES, Green MR, Kelsen DP, et al: Phase II trial of gemcitabine (2,2'-difluorodeoxycytidine) in patients with adenocarcinoma of the pancreas. Invest New Drugs 12:29-34, 1994
Khorana AA, Fine RL: Pancreatic cancer and thromboembolic disease. Lancet Oncol 5:655-663, 2004
Sorensen HT, Mellemkjaer L, Olsen JH, et al: Prognosis of cancers associated with venous thromboembolism. N Engl J Med 343:1846-1850, 2000
Kabbinavar F, Hurwitz HI, Fehrenbacher L, et al: Phase II randomized trial comparing bevacizumab plus fluorouracil (FU)/leucovorin (LV) with FU/LV alone in patients with metastatic colorectal cancer. J Clin Oncol 21:60-65, 2003
Johnson DH, Fehrenbacher L, Novotny WF, et al: Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non–small-cell lung cancer. J Clin Oncol 22:2184-2191, 2004
Sandler AB, Gray R, Brahmer J, et al: Randomized phase II/III trial of paclitaxel (P) plus carboplatin (CP) with or without bevacizumab (NSC# 704865) in patients with advanced non-squamous non-small cell lung cancer (NSCLC): An Eastern Cooperative Oncology Group (ECOG) Trial-E4599. Proc Am Soc Clin Oncol, 2005 (abstr 4)
Salven P, Orpana A, Joensuu H: Leukocytes and platelets of patients with cancer contain high levels of vascular endothelial growth factor. Clin Cancer Res 5:487-491, 1999
Willett CG, Boucher Y, di Tomaso E, et al: Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 10:145-147, 2004(Hedy L. Kindler, Gregory )
Cancer Research Center
Department of Health Studies
Department of Radiology, University of Chicago, Chicago
Evanston Hospital, Evanston
Ingalls Hospital, Harvey, IL
Fort Wayne Medical Center, Fort Wayne
Northern Indiana Cancer Research Consortium, South Bend, IN
ABSTRACT
PURPOSE: Vascular endothelial growth factor (VEGF) plays a key role in the biology and prognosis of pancreatic cancer. Inhibitors of VEGF suppress the growth of pancreatic cancer in preclinical models. The objectives of this phase II study were to assess the response rate and overall survival of pancreatic cancer patients who received gemcitabine with the recombinant humanized anti-VEGF monoclonal antibody bevacizumab.
PATIENTS AND METHODS: Patients with previously untreated advanced pancreatic cancer received gemcitabine 1,000 mg/m2 intravenously over 30 minutes on days 1, 8, and 15 every 28 days. Bevacizumab, 10 mg/kg, was administered after gemcitabine on days 1 and 15. Tumor measurements were assessed every two cycles. Plasma VEGF levels were obtained pretreatment.
RESULTS: Fifty-two patients were enrolled at seven centers between November 2001 and March 2004. All patients had metastatic disease, and 83% had liver metastases. Eleven patients (21%) had confirmed partial responses, and 24 (46%) had stable disease. The 6-month survival rate was 77%. Median survival was 8.8 months; median progression-free survival was 5.4 months. Pretreatment plasma VEGF levels did not correlate with outcome. Grade 3 and 4 toxicities included hypertension in 19% of the patients, thrombosis in 13%, visceral perforation in 8%, and bleeding in 2%.
CONCLUSION: The combination of bevacizumab plus gemcitabine is active in advanced pancreatic cancer patients. Additional study is warranted. A randomized phase III trial of gemcitabine plus bevacizumab versus gemcitabine plus placebo is ongoing in the Cancer and Leukemia Group B.
INTRODUCTION
Pancreatic cancer is the fourth leading cause of cancer death in the United States. It is estimated that 31,860 Americans developed pancreatic cancer in 2004, and 31,270 died as a result of this disease.1 The nearly identical incidence and mortality rates reflect the early distant spread of pancreatic cancer and the inadequacy of current therapies.
Single-agent gemcitabine, the standard chemotherapy for this disease, has only modest activity. In the pivotal randomized study that compared gemcitabine to a weekly bolus of 5-fluorouracil, gemcitabine produced an objective response rate of 5%, a median survival of 5.7 months, and a 6-month survival rate of 46%.2 Although combination chemotherapy with gemcitabine plus cytotoxic agents such as 5-fluorouracil, pemetrexed, irinotecan, oxaliplatin, and cisplatin achieves higher response rates than gemcitabine alone, randomized phase III trials of these combinations have not demonstrated a statistically significant improvement in survival over single-agent gemcitabine.3-7 Novel approaches that are based on the biologic characteristics of pancreatic cancer may be more promising than additional trials of chemotherapy doublets.
Vascular endothelial growth factor (VEGF) plays a key role in the growth and metastasis of many tumors including pancreatic cancer.8 VEGF and its receptors, flt-1 (VEGFR-1) and KDR (VEGFR-2) are coexpressed in pancreatic cancer, which suggests that VEGF could have autocrine effects on pancreatic cancer cells that express VEGF receptors and paracrine effects on microvascular endothelial cells.9-11 In pancreatic cancer cell lines, VEGF stimulates cell growth in a dose-dependent manner.9 In animal models, inhibitors of the VEGF tyrosine kinase, as well as anti-VEGF and anti-KDR antibodies, inhibit growth and angiogenesis associated with pancreatic tumors12-16 and potentiate the tumoricidal effect of gemcitabine.12,13,16 High serum VEGF levels correlate with advanced-stage postoperative disease recurrence and decreased survival in patients with pancreatic cancer.17 Several,18-22 although not all,23 investigators have reported that high tumor VEGF expression also correlates with a worse prognosis in this disease.
Bevacizumab (Avastin, Genentech, South San Francisco, CA) is a recombinant humanized anti-VEGF monoclonal antibody. In a phase III randomized trial in patients with advanced colorectal cancer, the addition of bevacizumab to standard chemotherapy resulted in a significant improvement in response, survival, and progression-free survival.24
We report the results of a multicenter phase II trial of bevacizumab given in combination with gemcitabine for patients with previously untreated advanced pancreatic cancer. The objectives of this study were to determine the objective tumor response rate, overall and progression-free survival rates, and toxicity profile.
PATIENTS AND METHODS
Patients
Eligible patients had histologically or cytologically confirmed pancreatic adenocarcinoma that was not amenable to potentially curative surgery; unidimensionally measurable disease as defined by Response Evaluation Criteria in Solid Tumors (RECIST) criteria25; Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2; and adequate bone marrow (granulocytes 1,500/μL, leukocytes 3,000/μL, platelets 100,000/μL), renal (creatinine 1.5 mg/dL or creatinine clearance 60 mL/min), and hepatic (normal total bilirubin and transaminases 2.5x the upper limit of normal) function. No prior chemotherapy for metastatic disease was allowed; prior adjuvant chemotherapy for radiosensitization was permitted if it did not contain gemcitabine or bevacizumab. Prior radiation therapy was allowed provided that the only sites of measurable disease were not located within the radiation port. Adjuvant therapy must have been completed at least 4 weeks before enrollment. Exclusion criteria included pregnancy or lactation; prior history of bleeding diathesis; tumor involvement of major blood vessels; prior cerebrovascular accident, pulmonary embolus, or deep venous thrombosis; myocardial ischemia, infarction, or uncompensated coronary artery disease within the past 6 months; major surgery within the prior 6 weeks; brain metastasis; active second malignancy; uncontrolled intercurrent illness; urine protein 1+ or 500 mg in 24 hours; or ongoing uncontrolled hypertension. Patients who received thrombolytic agents within the previous month or who required full-dose anticoagulation (except for maintenance of indwelling catheters with International Normalized Ratio < 1.5), daily aspirin, or daily nonsteroidal anti-inflammatory agents were not eligible. All patients provided written informed consent according to federal and institutional guidelines.
Treatment
Gemcitabine 1,000 mg/m2 was given intravenously over 30 minutes on days 1, 8, and 15 of a 28-day cycle. Bevacizumab 10 mg/kg was administered intravenously after gemcitabine on days 1 and 15 of each 28-day cycle. The initial bevacizumab dose was given over 90 minutes, and if no infusion reaction developed, the second dose was given over 60 minutes. If no infusion reaction occurred with the second dose, subsequent doses were given over 30 minutes.
At the initiation of this study, treatment was limited to six cycles of bevacizumab and gemcitabine. Patients who did not have disease progression after six cycles could receive additional single-agent bevacizumab until progression. Subsequent data from an ongoing phase III trial in colorectal cancer (Eastern Cooperative Oncology Group 3200) suggested that bevacizumab had marginal activity as a single agent.26 Therefore, the protocol was amended in October 2002 to permit patients with stable or responding disease to continue on both drugs until disease progression. Treatment was discontinued for progressive disease, unacceptable adverse events, or patient withdrawal of consent.
Dose Adjustments
Dose reduction was based on toxicities on the day of treatment. Adverse effects were graded according to National Cancer Institute Common Toxicity Criteria version 2.0. A cycle was not started until the absolute granulocyte count was more than 1.5 x 109/L and the platelet count was more than 100 x 109/L. Patients who developed either neutropenic fever that required antibiotic therapy or bleeding associated with thrombocytopenia received a 25% dose reduction of gemcitabine for subsequent cycles.
Within a cycle, if the granulocyte count was between 0.5 and 0.999 x 109/L or the platelet count was between 50 and 75 x 109/L on the day of treatment, the dose of gemcitabine was reduced by 25%. The dose was held for a granulocyte count less than 0.5 x 109/L or a platelet count less than 50 x 109/L. Patients with nonhematologic toxicities grades 0 to 2 (and grade 3 nausea/vomiting) received full-dose gemcitabine on days 8 and 15. For grade 3 nonhematologic toxicities other than nausea/vomiting, patients received either 75% of the gemcitabine dose or no treatment at the discretion of the treating physician. The gemcitabine dose was held for grade 4 nonhematologic toxicities.
There were no dose modifications of bevacizumab in this study. If gemcitabine was held at the beginning of a new cycle because of toxicity, the bevacizumab dose was also held until the chemotherapy could be given. If gemcitabine was held on day 15 for hematologic toxicity, bevacizumab was given. Bevacizumab was held for bilirubin or hepatic transaminase elevations of grade 3 or higher and was not resumed until they were grade 1 or less. If the hepatic abnormalities recurred with re-treatment, bevacizumab was discontinued permanently. Patients who developed either new proteinuria or an exacerbation of pre-existing proteinuria underwent a 24-hour urine collection for total protein. Patients who developed more than 2 g of proteinuria/24 hours could not receive additional doses of bevacizumab unless the proteinuria improved to less than 2 g/24 hours. Bevacizumab was also discontinued for grade 4 hypertension, uncontrolled grade 3 hypertension despite maximal medical therapy, grade 3 or 4 bleeding requiring transfusion or hospitalization, or any thrombotic events that required systemic anticoagulation.
Study Evaluations
Pretreatment evaluation included a complete medical history and physical exam, complete blood count and differential, chemistry panel (including liver function tests), prothrombin time/partial thromboplastin time, plasma VEGF level, pregnancy test (in women of childbearing potential), and a computed tomography scan of the chest, abdomen, and pelvis. The VEGF concentration in plasma was determined as described previously by using a quantitative sandwich enzyme immunoassay technique (Human VEGF Immunoassay; R & D Systems, Minneapolis, MN).27
A history and physical exam were performed every 14 days. Laboratory tests including complete blood count and differential, serum chemistries, and a prothrombin time/partial thromboplastin time were performed weekly. A urinalysis for assessment of proteinuria was performed every 14 days, on the days when bevacizumab was to be administered. Computed tomography scans were obtained every two cycles.
Response Criteria and Toxicity
Patients received a minimum of two cycles unless unacceptable toxicity or early progression of disease occurred. Patients were evaluated for response according to RECIST criteria after two successive cycles.26 Confirmatory scans were obtained at least 4 weeks after initial documentation of objective complete or partial response.
Statistical Analysis
The primary end point of this study was the objective response rate (complete response + partial response). The trial was conducted by using a Simon optimal two-stage design28 to test the null hypothesis that the response rate was 10% versus the alternative that it was at least 25%. Twenty-one assessable patients were to be enrolled in the first stage. If two or fewer patients had an objective response, the trial would be terminated for lack of efficacy. Otherwise, an additional 29 patients were to be enrolled, and if eight or more objective responses were observed among the 50 patients, the regimen would be considered worthy of additional evaluation in this disease. This design yields at least .90 probability of a positive result if the true response rate is at least 25% and at least .90 probability of a negative result if the true response rate is at most 10%. In addition, 50 patients would be sufficient to distinguish between a 65% 6-month survival rate and a 45% 6-month survival rate. Progression-free and overall survival were calculated by using the method of Kaplan and Meier.29 Median progression-free and overall survival times and their associated 95% CIs were derived as described by Brookmeyer and Crowley.30
A protocol amendment in October 2002 required that both drugs, instead of bevacizumab alone, be continued after six cycles (months) in patients whose disease was nonprogressing. Additional patients were enrolled until a total of 45 were available for evaluation of the modified treatment plan. Primary analysis of response rate and 6-month overall survival used data from all enrolled patients, because the amended treatment plan had no effect on these end points. Analysis of overall survival and progression-free survival was assessed for the pooled group of 52 patients and then was assessed separately for patients who were enrolled before (n = 7) and after the amendment (n = 45). Results from the former analysis are reported here because there were no significant differences between the two groups of patients.
Because bleeding, thrombosis, nephrotic syndrome, and uncontrolled hypertension were potential concerns with the administration of bevacizumab, these events were monitored closely. Specifically, any grade 3 or 4 toxicity for bleeding, thrombosis, or proteinuria or any grade 4 hypertension were reported immediately, and a stopping guideline developed by Goldman31 was used as a guide for early termination as a result of toxicity. The trial would be halted if there was evidence that the true rate of toxicity was increased from an acceptable level of o = 10% to an upper limit of A = 25%. A stopping rule that provides an level of 4% and a power of 77% requires that early termination be considered if three adverse events occur in the first five or fewer patients, four adverse events in the first eight or fewer patients, five adverse events in the first 14 patients, six adverse events in the first 20 patients, seven adverse events in the first 26 patients, or eight adverse events in the first 33 or fewer patients.
Mean baseline levels of plasma VEGF were compared between responders and nonresponders by using a nonparametric Wilcoxon test. Patients were also divided into two cohorts on the basis of whether their baseline VEGF levels were greater than the median value. Log-rank tests were then performed to compare progression-free and overall survival for these cohorts.
Independent Response Audit
A special response review was held on December 30, 2002, at the request of the Investigational Drug Branch, Cancer Therapy Evaluation Program. There were 15 participants including the principal investigator (H.L.K.), a diagnostic radiologist from the Clinical Center of the National Institutes of Health, and medical oncology and regulatory representatives from the Cancer Therapy Evaluation Program and Genentech.
Sixteen patients were assessable for response at the time of the audit; of these, seven cases considered by the principal investigator to meet RECIST criteria for objective response or prolonged stable disease were reviewed. A brief patient summary, which included patient history, treatment, toxicity, and response to treatment, and a copy of the pathology report were provided by the principal investigator for each case selected for review. All baseline and follow-up scans necessary to verify the claimed response were available at the review meeting. The diagnostic radiologist independently measured all target lesions by using calipers, the percent change was calculated, and these values were compared to those obtained previously by the study team at the University of Chicago.
RESULTS
Patient Characteristics
Fifty-two patients at seven centers in the University of Chicago phase II consortium were enrolled in this trial from November 2001 to March 2004. Patient characteristics are listed in Table 1. Forty-eight percent of the patients were women, and most (98%) had an ECOG performance status of 0 or 1. All the patients had metastatic disease, and 83% had liver metastases.
A total of 250 cycles were delivered (median, 4; range, 1 to 18). Seven patients were treated before a protocol amendment in October 2002 mandated that both agents be continued after cycle six in patients with nonprogressing disease. Of the initial seven patients, two had no disease progression by cycle six and were able to receive single-agent bevacizumab for four and five additional cycles, respectively.
Response to Treatment and Survival
All 52 patients were included in an intention-to-treat analysis of response. There were 11 confirmed partial responses, for an overall response rate of 21% (95% CI, 11% to 35%). The median duration of response was 10 months (range, 4.9 to 12.1 months). The first five responses were confirmed by an independent response audit at the National Cancer Institute. Twenty-four patients (46%) had stable disease as their best response, which was maintained for a median of 6.3 months. Thirteen patients (25%) had progressive disease.
At the time of this analysis, no patients remained on the study treatment, and 45 patients (87%) had died. The median follow-up time for all patients was 8.0 months. The Kaplan-Meier curve for overall survival is presented in Figure 1. The estimated median survival was 8.8 months (95% CI, 7.4 to 9.7 months). The 6-month survival rate was 77% (95% CI, 63% to 86%), and the 1-year survival rate was 29% (95% CI, 17% to 42%). Figure 2 shows the Kaplan-Meier curve for progression-free survival. The median progression-free survival was 5.4 months (95% CI, 3.7 to 6.2).
Toxicity
Toxicities are summarized in Table 2. Hematologic toxicities were similar to those observed with single-agent gemcitabine.2 Only 35% of patients developed grade 3/4 neutropenia, 4% experienced grade 3 anemia, and 8% had grade 3 thrombocytopenia. No patient developed neutropenic fever. The most common nonhematologic toxicities included fatigue, hyperglycemia, nausea, and anorexia. Hepatic transaminase elevations were related principally to disease progression.
Several nonhematologic toxicities could be attributed to bevacizumab. Forty-four percent of patients experienced hypertension, including 19% who developed grade 3 hypertension. In most patients, this was managed with standard oral antihypertensive drugs; however, two patients discontinued bevacizumab because of poorly controlled hypertension. No patient experienced a hypertensive crisis.
Although proteinuria occurred in 36% of patients, it was grade 3 in only 2%, and no patient developed nephrotic syndrome. Seven patients (14%) developed grade 3 or 4 deep vein thromboses or pulmonary emboli, which required discontinuation of bevacizumab according to the protocol. Three of these thromboses occurred in the first cycle of treatment. None were associated with indwelling venous catheters. There were no arterial thromboses.
Bleeding episodes, although frequent, were usually mild and included epistaxis in 23% of the patients and gum bleeding in 4% of the patients. A lethal gastrointestinal bleed occurred in a 79-year-old man. Endoscopic retrograde cholangiopancreatography demonstrated an actively bleeding vessel and extensive ulceration resulting from malignant invasion from the duodenal bulb and ampullary region.
The toxicity monitoring boundary was reached when seven of the first 26 patients developed grade 3 or higher bleeding, thrombosis, and/or proteinuria. The trial was continued with close monitoring, however, and only two additional such toxicities occurred in the remaining 26 patients who were enrolled.
Four patients developed visceral perforations, one of which was fatal. Nonfatal perforations included a Mallory-Weiss tear in the setting of malignant duodenal obstruction, perforation with ulceration at a gastrojejunostomy anastomosis, and perforation at both ends of a sigmoid stent 8 days postprocedure. A 72-year-old man who developed abdominal pain, fever, and peritonitis 13 days after his initial dose of bevacizumab died 5 days later without additional intervention as a result of presumed bowel perforation.
Correlative Studies
Pretreatment plasma VEGF levels were obtained in 42 patients and ranged from 0 to 462 pg/mL (median, 84 pg/mL). Although patients who attained a partial response or stable disease had slightly higher baseline VEGF levels than those with progressive disease (mean ± standard deviation, 101 ± 112 and 140 ± 122 v 82 ± 85 pg/mL, respectively), the difference was not statistically significant (Kruskal-Wallis exact test, P = .23). There was no significant difference in overall survival (log-rank test, P = .20) or progression-free survival (log-rank test, P = .37) between patients whose VEGF levels were above or below the median. Figure 3 illustrates the Kaplan-Meier overall survival curve by plasma VEGF level.
DISCUSSION
Novel agents and approaches are needed for the treatment of advanced pancreatic cancer, a disease that has the briefest survival of any solid tumor. Although single-agent gemcitabine provides only a modest survival benefit, countless classical cytotoxic agents, given alone or in combination with gemcitabine, have not proven to be superior.32 We evaluated the combination of gemcitabine with the anti-VEGF monoclonal antibody bevacizumab in 52 patients with advanced pancreatic cancer and observed an objective response rate of 21%, a median progression-free survival of 5.4 months, a 6-month overall survival rate of 77%, and a median survival time of 8.8 months. These data seem superior to comparable phase II and III data for single-agent gemcitabine2-7,32,33 and suggest that the combination of gemcitabine and bevacizumab may have a role in the treatment of this disease.
Although most of the patients in this study had a good ECOG performance status, all had metastatic disease and more than 80% had liver metastases, both of which are poor prognostic factors. Patients with pancreatic cancer are at an increased risk of developing malignancy-related hypercoagulability.34 Because the safety of administering bevacizumab with full-dose anticoagulation was unknown at the initiation of this study, patients who had venous thromboses that required anticoagulation were excluded. It is quite possible that we introduced a selection bias by excluding these patients, because cancer patients who have experienced thromboembolism may have a worse prognosis.35 However, bevacizumab did not seem to increase the expected rate of grade 3/4 thrombosis.5
Gastrointestinal bleeding is an unfortunate late complication of this disease that can occur when the pancreatic tumor invades the duodenum. It is not possible to ascertain whether bevacizumab exacerbated the ultimately fatal bleeding in our patient whose tumor eroded into his duodenum. We would advise caution, however, before considering this drug in any individual who has tumor invasion of an adjacent organ.
Gastrointestinal perforation is an infrequent, life-threatening complication of bevacizumab that may be a result of the effects of bevacizumab on wound healing.24 In the pivotal phase III trial of bevacizumab in colorectal cancer, Hurwitz et al24 observed a 1.5% rate of perforation in the treatment group and none in the placebo control. The 8% rate of visceral perforation in our study is significantly higher than that in the colorectal study and is likely a reflection of our patient population. Because one patient developed a perforation after a colon stent placement and another after severe vomiting from a duodenal obstruction, it would be appropriate to hold additional bevacizumab in these situations.
Because there have been no dose-finding trials of bevacizumab in pancreatic cancer, the optimal dose of this agent for this disease remains unclear. Although it is reasonable to speculate whether fewer toxicities or alternate efficacy might have been observed had we arbitrarily chosen a lower dose than the 10 mg/kg used in this trial, this cannot be definitively ascertained without additional study. A randomized phase II trial in colorectal cancer suggested that a dose of 5 mg/kg every 14 days was more effective than 10 mg/kg,36 and a randomized phase III trial in that disease confirmed the activity of the 5 mg/kg dose.24 Significant activity, with tolerable toxicity, was subsequently reported in a phase III study in colorectal cancer that used a 10 mg/kg dose.26 In a randomized phase II trial in non–small-cell lung cancer, a dose of 15 mg/kg every 21 days was more active than the 7.5-mg dose, and there were fewer episodes of significant bleeding at the higher dose.37 The efficacy and safety of the 15 mg/kg bevacizumab dose in that disease has been confirmed in a randomized phase III trial.38
Preclinical models suggest a key role for VEGF in the progression and prognosis of pancreatic cancer,8-11 and elevated serum17 and tumor18-22 VEGF levels have been shown to correlate with poor prognosis in patients with pancreatic cancer. Serum VEGF levels may be unreliable, because they are derived principally from platelets, whereas tumor VEGF is thought to enrich the plasma fraction.39 Despite its overall negative prognostic value, higher plasma VEGF levels may correspond to tumors that are more VEGF dependent and thus more sensitive to VEGF inhibition.
We were not able to correlate plasma VEGF levels with clinical outcome in this study. VEGF levels during and after bevacizumab administration could not be measured reliably because bevacizumab interferes with the immunoassay. Although we observed a trend toward higher baseline soluble plasma VEGF levels in patients who had a partial response or stable disease, it did not reach statistical significance. This may be because of the small sample size or our choice of surrogate biomarker.
In patients with rectal cancer, Willett et al40 observed no decrease in plasma, serum, or urine VEGF levels when bevacizumab was administered; however, the number of viable, circulating endothelial and progenitor cells was decreased after bevacizumab treatment.
We conclude that the combination of bevacizumab and gemcitabine is active in patients with advanced pancreatic cancer and that additional study is warranted. The Cancer and Leukemia Group B (CALGB) is currently accruing to a 590-patient, double-blind, placebo-controlled, randomized phase IV trial (CALGB 80303) that compares gemcitabine plus bevacizumab to gemcitabine plus placebo using the doses and schedule used in this phase II study. Angiogenic biomarkers, proteomic profiling, and pharmacogenomic predictors are being evaluated to further elucidate the biology of VEGF in this disease, and a clinical economic analysis will delineate the potential economic impact of this combination.
Acknowledgment
We thank Cynthia Bajda for assistance with data management; Amy Brich, RN, BSN, for assistance with patient care and protocol monitoring; and Kristen Kasza, MS, for assistance with statistical analysis.
NOTES
Supported by National Cancer Institute Grant No. N01-CM-17102.
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, May 31-June 3, 2003, Chicago, IL; 40th Annual Meeting of the American Society of Clinical Oncology, June 5-8, 2004, New Orleans, LA; and the American Society of Clinical Oncology Gastrointestinal Cancer Symposium, January 22-24, 2004, San Francisco, CA.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Jemal A, Tiwari RC, Murray T, et al: Cancer Statistics, 2004. CA Cancer J Clin 54:8-29, 2004
Burris HA, Moore MJ, Andersen J III, et al: Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: A randomized trial. J Clin Oncol 15:2403-2413, 1997
Berlin JD, Catalano P, Thomas JP, et al: Phase III study of gemcitabine in combination with fluorouracil versus gemcitabine alone in patients with advanced pancreatic carcinoma: Eastern Cooperative Oncology Group Trial E2297. J Clin Oncol 20:3270-3275, 2002
Richards DA, Kindler HL, Oettle H, et al: A randomized phase III study comparing gemcitabine + pemetrexed versus gemcitabine in patients with locally advanced and metastatic pancreas cancer. J Clin Oncol, 2004 ASCO Annual Meeting Proceedings (Post-Meeting Edition) 22, 2004 (abstr 4007)
Rocha Lima CM, Green MR, Rotche R, et al: Irinotecan plus gemcitabine results in no survival advantage compared with gemcitabine monotherapy in patients with locally advanced or metastatic pancreatic cancer despite increased tumor response rate. J Clin Oncol 22:3776-3783, 2004
Heinemann V, Quietzsch D, Gieseler F, et al: A phase III trial comparing gemcitabine plus cisplatin vs. gemcitabine alone in advanced pancreatic carcinoma. Proc Am Soc Clin Oncol 22:250, 2003 (abstr 1003)
Louvet C, Labianca R, Hammel P, et al: Gemcitabine in combination with oxaliplatin compared with gemcitabine alone in locally advanced or metastatic pancreatic cancer: Results of a GERCOR and GISCAD phase III trial. J Clin Oncol 23(15):3509-3516, 2005
Korc M: Pathways for aberrant angiogenesis in pancreatic cancer. Mol Cancer 2:8, 2003
Itakura J, Ishiwata T, Shen B, et al: Concomitant over-expression of vascular endothelial growth factor and its receptors in pancreatic cancer. Int J Cancer 85:27-34, 2000
Luo J, Guo P, Matsuda K, et al: Pancreatic cancer cell-derived vascular endothelial growth factor is biologically active in vitro and enhances tumorigenicity in vivo. Int J Cancer 92:361-369, 2001
von Marschall Z, Cramer T, Hocker M, et al: De novo expression of vascular endothelial growth factor in human pancreatic cancer: Evidence for an autocrine mitogenic loop. Gastroenterology 119:1358-1372, 2000
Baker CH, Solorzano CC, Fidler IJ: Blockade of vascular endothelial growth factor receptor and epidermal growth factor receptor signaling for therapy of metastatic human pancreatic cancer. Cancer Res 62:1996-2003, 2002
Solorzano CC, Baker CH, Bruns CJ, et al: Inhibition of growth and metastasis of human pancreatic cancer growing in nude mice by PTK 787/ZK222584, an inhibitor of the vascular endothelial growth factor receptor tyrosine kinase. Cancer Biother Radiopharm 16:359-370, 2001
Bockhorn M, Tsuzuki Y, Xu L, et al: Differential vascular and transcriptional responses to anti-vascular endothelial growth factor antibody in orthotopic human pancreatic cancer xenografts. Clin Cancer Res 9:4221-4226, 2003
Tsuzuki Y, Mouta Carreira C, Bockhorn M, et al: Pancreas microenvironment promotes VEGF expression and tumor growth: Novel window models for pancreatic tumor angiogenesis and microcirculation. Lab Invest 81:1439-1451, 2001
Bruns CJ, Shrader M, Harbison MT, et al: Effect of the vascular endothelial growth factor receptor-2 antibody DC-101 plus gemcitabine on growth, metastasis, and angiogenesis of human pancreatic cancer growing orthotopically in nude mice. Int J Cancer 102:101-108, 2002
Karayiannakis AJ, Bolanaki H, Syrigos KN, et al: Serum vascular endothelial growth factor levels in pancreatic cancer patients correlate with advanced and metastatic disease and poor prognosis. Cancer Lett 194:119-124, 2003
Ikeda N, Adachi M, Taki T, et al: Prognostic significance of angiogenesis in human pancreatic cancer. Br J Cancer 79:1553-1563, 1999
Itakura J, Ishiwata T, Friess H, et al: Enhanced expression of vascular endothelial growth factor in human pancreatic cancer correlates with local disease progression. Clin Cancer Res 3:1309-1316, 1997
Knoll MR, Rudnitzki D, Sturm J, et al: Correlation of postoperative survival and angiogenic growth factors in pancreatic carcinoma. Hepatogastroenterology 48:1162-1165, 2001
Niedergethmann M, Hildenbrand R, Wostbrock B, et al: High expression of vascular endothelial growth factor predicts early recurrence and poor prognosis after curative resection for ductal adenocarcinoma of the pancreas. Pancreas 25:122-129, 2002
Seo Y, Baba H, Fukuda T, et al: High expression of vascular endothelial growth factor is associated with liver metastasis and a poor prognosis for patients with ductal pancreatic adenocarcinoma. Cancer 88:2239-2245, 2000
Ellis LM, Takahashi Y, Fenoglio CJ, et al: Vessel counts and vascular endothelial growth factor expression in pancreatic adenocarcinoma. Eur J Cancer 34:337-340, 1998
Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205-216, 2000
Giantonio BJ, Catalano PJ, Meropol NJ, et al: High-dose bevacizumab improves survival when combined with FOLFOX4 in previously treated advanced colorectal cancer: Results from the Eastern Cooperative Oncology Group study E3200. Proc Am Soc Clin Oncol, 2005 (abstr 2)
Salven P, Manpaa H, Orpana A, et al: Serum vascular endothelial growth factor is often elevated in disseminated cancer. Clin Cancer Res 3:647-651, 1997
Simon R: Optimal two-stage designs for phase II clinical trials. Control Clin Trials 10:1-10, 1989
Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
Brookmeyer R, Crowley J: A confidence interval for the median survival time. Biometrics 38:29-41, 1982
Goldman AI: Issues in designing sequential stopping rules for monitoring side effects in clinical trials. Control Clin Trials 8:327-337, 1987
Kulke MH: Advanced pancreatic cancer: Is there a role for combination therapy? Expert Rev Anticancer Ther 3:729-739, 2003
Casper ES, Green MR, Kelsen DP, et al: Phase II trial of gemcitabine (2,2'-difluorodeoxycytidine) in patients with adenocarcinoma of the pancreas. Invest New Drugs 12:29-34, 1994
Khorana AA, Fine RL: Pancreatic cancer and thromboembolic disease. Lancet Oncol 5:655-663, 2004
Sorensen HT, Mellemkjaer L, Olsen JH, et al: Prognosis of cancers associated with venous thromboembolism. N Engl J Med 343:1846-1850, 2000
Kabbinavar F, Hurwitz HI, Fehrenbacher L, et al: Phase II randomized trial comparing bevacizumab plus fluorouracil (FU)/leucovorin (LV) with FU/LV alone in patients with metastatic colorectal cancer. J Clin Oncol 21:60-65, 2003
Johnson DH, Fehrenbacher L, Novotny WF, et al: Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non–small-cell lung cancer. J Clin Oncol 22:2184-2191, 2004
Sandler AB, Gray R, Brahmer J, et al: Randomized phase II/III trial of paclitaxel (P) plus carboplatin (CP) with or without bevacizumab (NSC# 704865) in patients with advanced non-squamous non-small cell lung cancer (NSCLC): An Eastern Cooperative Oncology Group (ECOG) Trial-E4599. Proc Am Soc Clin Oncol, 2005 (abstr 4)
Salven P, Orpana A, Joensuu H: Leukocytes and platelets of patients with cancer contain high levels of vascular endothelial growth factor. Clin Cancer Res 5:487-491, 1999
Willett CG, Boucher Y, di Tomaso E, et al: Direct evidence that the VEGF-specific antibody bevacizumab has antivascular effects in human rectal cancer. Nat Med 10:145-147, 2004(Hedy L. Kindler, Gregory )