Bortezomib, Paclitaxel, and Carboplatin As a First-Line

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector ORIGINAL ARTICLE

Bortezomib, Paclitaxel, and Carboplatin as a First-Line Regimen for Patients with Metastatic Esophageal, Gastric, and Gastroesophageal Cancer Phase II Results from the North Central Cancer Treatment Group (N044B)

Aminah Jatoi, MD,* Shaker R. Dakhil, MD,† Nathan R. Foster, MD,* Cynthia Ma, MD,‡ Kendrith M. Rowland, Jr., MD,§ Dennis F. Moore, Jr., MD,† Anthony J. Jaslowski, MD,ʈ Sachdev P. Thomas, MD,¶ Mark D. Hauge, MD,# Patrick J. Flynn, MD,** Philip J. Stella, MD,†† and Steven R. Alberts, MD*

they manifested tumor progression or severe adverse events. All were Purpose: This study was undertaken to explore the response rate of monitored for tumor response as well as other clinical outcomes. a first-line, three-drug regimen that consisted of bortezomib, pacli- Results: The cohort included 35 eligible patients with a median age taxel, and carboplatin in patients with metastatic adenocarcinoma of of 59 years (range, 36–78) and an Eastern Cooperative Oncology the esophagus, gastroesophageal junction, or gastric cardia. Group performance score of 0, 1, and 2 in 60%, 34%, and 6% of Patients and Methods: Patients with the above diagnosis and patients, respectively. Although this regimen was well tolerated, the acceptable organ function were treated intravenously on a 21-day tumor response rate was lower than that anticipated at 23% (95% 2 cycle with the following: bortezomib 1.2 mg/m on days 1, 4, and 8; confidence interval: 10%, 40%), thereby prompting premature study 2 paclitaxel 175 mg/m on day 2; and carboplatin with an area under closure. There were no complete responses. The median survival for the curve of 6 on day 2. Patients received indefinite treatment unless the cohort was 8.9 months (95% confidence interval: 5.9, 12.8). Conclusion: As prescribed in this trial and for this indication, this regimen does not merit further testing. *Mayo Clinic and Mayo Foundation, Rochester, MN; †Wichita Community (J Thorac Oncol. 2008;3: 516–520) Clinical Oncology Program, Wichita, KS; ‡Washington University, St. Louis, Missouri; §Carle Cancer Center CCOP, Urban, IL; ʈSt. Vincent Regional Cancer Center CCOP, Green Bay, WI; ¶Illinois Oncology Research Assn. CCOP, Peoria, IL; #CentraCare Clinic, St. Cloud, MN; ortezomib is a modified dipeptidyl boronic acid that **Metro-Minnesota Community Clinical Oncology Program, St. Louis inhibits the 26S proteasome and, in turn, leads to chaos Park, MN; and ††Michigan Cancer Research Consortium, Ann Arbor, MI. B 1 Disclosure: The authors declare no conflicts of interest. within the ubiquitin proteasome pathway. The latter is crit- This study was conducted as a collaborative trial of the North Central Cancer ical to cell protein regulation, and plays a key role in the Treatment Group and Mayo Clinic and was supported in part by Public degradation of damaged or senescent proteins. In the setting Health Service grants: CA-25224, CA-37404, CA-15083, CA-63848, of cancer, bortzomib’s effects on the 26S proteasome result in CA-35195, CA-35090, CA-35269, CA-37417, CA-52654, CA-35267, CA-63849, CA-35113, CA-60276, CA-35119, CA-35103, and CA-35431. altered protein homeostasis and disruption of cell signaling Additional participating institutions include: Siouxland Hematology-Oncol- pathways, thereby detracting from the cancer cell’s ability to ogy Associates, Sioux City, IA 51105 (Donald B. Wender, MD); Mis- remain viable.1 Indeed, proof of concept has already been souri Valley Cancer Consortium, Omaha, NE 68131 (Gamini S. Soori, established: bortezomib provides a survival advantage to MD); Mayo Clinic Jacksonville, Jacksonville, FL 32224 (Edith Perez, MD); Duluth CCOP, Duluth, MN 55805 (Daniel A. Nikcevich, MD); chemotherapy-refractory multiple myeloma patients, and this Medcenter One Health Systems, Bismarck, ND 58506 (Edward J. Wos, drug is the first such class of agents approved by the Food and D.O.); Meritcare Hospital CCOP, Fargo, ND 58122 (Preston D. Steen, Drug Administration for a cancer indication.2,3 MD); Upstate Carolina CCOP, Spartanburg, SC 29303 (James D. Bear- This study therefore sought to determine the response den, III, MD); Abbott Northwestern Hospital, Minneapolis, MN 55407 (Daniel J. Schneider, MD); Mayo Clinic Scottsdale, Scottsdale, AZ rate of a three-drug regimen that included bortezomib in 85259-5404 (Tom R. Fitch, MD); Mobile Infirmary Medical Center, patients with metastatic cancer of the esophageal cancer, Mobile, AL 36607 (Paul Schwarzenberger, MD). gastroesophageal junction, and gastric cardia, herein referred Address for correspondence: Aminah Jatoi, MD, Mayo Clinic, 200 First to as “esophageal cancer.” This issue is relevant for several Street, SW, Rochester, MN 55905. E-mail: [email protected] Copyright © 2008 by the International Association for the Study of Lung reasons. First, esophageal cancer cells do also rely on an Cancer intricate cell signaling system, and it is plausible that protea- ISSN: 1556-0864/08/0305-0516 some inhibition might disrupt one or more of these pathways

516 Journal of Thoracic Oncology • Volume 3, Number 5, May 2008 Journal of Thoracic Oncology • Volume 3, Number 5, May 2008 Metastatic Esophageal, Gastric, and Gastroesophageal Cancer

and thereby result in the death of esophageal cancer cells.4 In addition, all patients had to have had the following Pathways involving such mediators as NF kappa B, BAX, laboratory parameters within 14 days before registration: (1) and P27 have been implicated in therapy resistance in esoph- absolute neutrophil counts Ն1.5 ϫ 103 cells/mL; (2) platelet ageal cancer, and they are targets of bortezomib.5 Second, count Ն100 ϫ 103 cells/mL; (3) total bilirubin equal to or preclinical data suggest that bortezomib carries antineoplastic below the upper limit of normal; (4) aspartate aminotransfer- activity in this setting. Fujita and others examined the gastric ase Յ3 times the upper limit of normal; (5) the alkaline cancer cells lines AZ521, MKN45, and NUGC3 and observed phosphatase Յ2 times the upper limit of normal; and (6) that exposure to bortezomib resulted in cell growth inhibi- serum creatinine Յ1.5 times the upper limit of normal. tion.6 Synergy occurred when bortezomib was combined with Patients were not allowed to enter the trial in the event a conventional chemotherapy agent, such as paclitaxel, pre- of any one of the following: (1) pregnant, nursing, or of sumably from greater down-regulation of NF kappa B and child-bearing potential and unwilling to use effective contra- increasing certain cell cycle regulatory proteins.5 Although ception; (2) previous radiation to greater than 25% of the this study focused on gastric cancer cell lines, not esophageal, marrow cavity; (3) open abdominal surgery in the previous 4 the adjacency of these two organs coupled with established weeks or a laparoscopic procedure within the previous 2 similar treatment approaches suggest that these data are in weeks; (4) an uncontrolled infection or a chronic debilitating fact relevant even to the latter. Third, favorable, mounting illness; (5) known central nervous system metastases; (6) clinical evidence in patients with other solid tumor malignan- peripheral neuropathy of grade 2 or worse; (7) a previous cies suggests that this agent merits further study in patients allergic reaction to either carboplatin or paclitaxel; (8) a prior with esophageal cancer.7–9 Ocean and others treated 36 gas- malignancy, except for adequately treated basal cell or squa- tric cancer patients with bortezomib alone versus bortezomib mous cell carcinoma of the skin or any other cancer for which plus irinotecan.7 The latter regimen yielded a 44% response the patient has been cancer-free for 5 years or longer; or (9) rate, and the former a 9% response rate. These findings led prior chemotherapy, radiation, immunotherapy, or biologic these investigators to conclude that bortezomib is “active” in therapy for recurrent or metastatic disease. For clarification, patients with metastatic gastric cancer. Hence, the foregoing prior radiotherapy and/or chemotherapy was permissible if suggests a rationale for further testing bortezomib in patients administered in the neoadjuvant or adjuvant setting after with metastatic esophageal cancer. complete resection of the original tumor. Moreover, previous Thus, the present study was undertaken. The study team combination chemotherapy and radiation for locally ad- used a phase II study design to test a three-drug regimen of vanced disease given with curative intent were also permis- bortezomib, paclitaxel, and carboplatin. This three-drug reg- sible if a complete clinical or pathologic response had been imen was chosen based on the following three factors that achieved with this therapy and if the last chemotherapy or summarize rationale and feasibility: (1) the preclinical and radiation was Ն6 months before enrollment. clinical synergistic antitumor effects, as described earlier; (2) Pretreatment Evaluation. All patients underwent a history the fact that the regimen paclitaxel and carboplatin has and physical examination, which included a brief neurologic demonstrated antineoplastic activity in this setting10–12; and examination, within 14 days of study registration. In addition, (3) the fact that Ma et al.5 had generated the prerequisite a hemogram, a total bilirubin, a serum creatinine, an alkaline phase I data for the appropriate dosing of this combination. phosphatase, an aspartate aminotransferase, and a chest ra- These phase I data defined the optimal dosing of all three of diograph were also obtained. A computerized tomography these agents based on sequencing of bortezomib with respect scan or a magnetic resonance imaging study that included a to chemotherapy and dose-escalation of both bortezomib and site of measurable tumor was required within 30 days of paclitaxel. study registration. Treatment. Patients were treated with bortezomib 1.2 mg/m2 intravenously on days 1, 4, and 8 of a 21-day chemo- PATIENTS AND METHODS 2 Overview. This phase II study was designed and con- therapy cycle. Additionally, paclitaxel 175 mg/m was ad- ducted within the North Central Cancer Treatment Group, ministered intravenously over 3 hours on day 2. Finally, and the Institutional Review Boards at each study site had carboplatin with an area under the curve (AUC) of 6 was also approved the study protocol. All patients were required to administered intravenously on day 2. The protocol specified provide written consent before enrollment. appropriate premedication with dexamethasone, diphenhy- Eligibility. Eligibility criteria consisted of the following: (1) dramine, and ranitidine or an equivalent medication in an patient age of Ն18 years; (2) histologic or cytologic evidence of effort to prevent drug reactions. adenocarcinoma of the esophagus, esophagogastric junction, Patients were to continue on the regimen indefinitely, until cancer progression, or until unacceptable toxicity oc- or gastric cardia; (3) no curative therapeutic options; (4) curred. The patient’s decision to stop therapy with no appar- 13 measurable disease, as defined by the RECIST criteria ; (5) ent explanation was also considered acceptable and was Eastern Cooperative Oncology Group performance status of 2 specified in the protocol. or better; (6) a negative pregnancy test 7 days before regis- Dose modifications were specified for each agent. For tration or earlier; and (7) patient’s estimated life expectancy bortezomib, an interval absolute neutrophil count of less than of 12 weeks or longer. 1.0 ϫ 103 cells/mL and/or a platelet count of less than 50 ϫ

103 cells/mL required that this agent be omitted on day #8 Secondary endpoints included descriptive summaries and then reduced by 20% for subsequent cycles. A similar of time-to-cancer progression, adverse events, and overall approach was specified in the event of febrile neutropenia, survival. Time-to-cancer progression is the time from enroll- defined as the above neutrophil count with a concurrent fever. ment on study to the time of evidence of cancer progression. Similarly, the above parameters at any point during a treat- Adverse events are presented in part in tabular form. Overall ment cycle prompted a 20% dose reduction of both paclitaxel survival is defined as the time from study registration to death and carboplatin. Grade 3 or worse gastrointestinal toxicity or from any cause. Time-to-event distributions are estimated other nonhematologic toxicity also required dose reductions with the Kaplan Meier method.15 for both these drugs. Finally, at the beginning of each treatment cycle, all cancer treatment was to be held until the RESULTS absolute neutrophil count and platelet count exceeded 1.5 ϫ Demographics. Thirty-five eligible patients were enrolled 3 3 10 cells/mL and 100 ϫ 10 cells/mL, respectively. Treat- and treated between November 2005 and June 2006. One ment was also to be held in the event of grade 3 or worse patient cancelled and did not receive any of the drugs, and nonhematologic toxicity with resumption of chemotherapy another was subsequently found to be ineligible; thus, these when toxicities either returned to baseline or dropped to less two patients are not included in the tumor response and than grade 1. Patients with grade 2 neuropathy could proceed survival analyses shown below. The median age of this with chemotherapy with a 20% dose reduction of all three drugs, but patients with neuropathy of grade 3 or worse were 35-patient cohort at study entry was 59 years with a range to discontinue treatment on protocol. In the event that che- from 36 to 78 years. Eleven percent were women. Eastern motherapy needed to be held for more than 3 weeks, patients Cooperative Oncology Group performance scores of 0, 1, and were again to discontinue treatment on protocol. 2 were observed in 60%, 34%, and 6% of the cohort, The protocol stated that full supportive care measures respectively (Table 1). were to be used while patients received chemotherapy on Treatment Administration. Patients completed a median of study. The use of blood products, antibiotics, nutrition sup- four cycles of chemotherapy (range, 1–13). Reasons for port, and antiemetics were allowed as clinically appropriate. discontinuation are as follows: cancer progression (n ϭ 23), The use of neutrophil growth factor support was allowed in adverse events or patient declined further,10 or other.3 the event of severe neutropenic complications. Although intermittent dose reductions were required for Follow-up Evaluations. Visits with an oncologist were medical reasons, most patients nonetheless received the ma- required every 3 weeks, but patients were otherwise moni- jority of the originally intended chemotherapy dosing. For tored with a weekly hemogram and other weekly blood bortezomib, the range of the median percentage of targeted testing as deemed clinically appropriate by their healthcare dose prescribed was 67 to 100% over the first 6 cycles of providers. Tumor measurements were to be obtained imme- chemotherapy; and, for paclitaxel, it was 72 to 100%. For diately before the anticipated third cycle, and every other carboplatin, the percentage of patients who maintained an cycle thereafter unless more frequent assessments were re- AUC of 6 over the first 6 cycles ranged from 7% for cycle 6 quired to confirm a tumor response. RECIST criteria were to 100% for cycle 1 (median 45%). For the first 3 cycles, the used to define a confirmed tumor response.13 Adverse events were recorded with the National Cancer Institute’s Common TABLE 1. Baseline Characteristics (n ϭ 35) Toxicity Criteria, version 3.0 and tracked throughout the a study period. Patient (%) Statistical Analyses. The primary objective of this trial was to Age, median (range) 59 (36, 78) assess the proportion of patients with a confirmed tumor Gender response treated as per the regimen outlined in this protocol. Female 4 (11) The largest success proportion whereby the proposed treat- Male 31 (89) Performance score ment regimen would be considered unworthy of further study 0 21 (60) was 30%, based on earlier data that tested carboplatin and 1 12 (34) paclitaxel alone, and the smallest success proportion that 2 2 (6) would warrant further study was 50%. These parameters were Prior radiation therapy conservative based on previously reported response rates with Yes 5 (14) carboplatin and paclitaxel in similar settings.8–10 A one-stage No 30 (86) Fleming design with a minimum of 25 patients and a maxi- Prior cancer surgery mum of 60 patients were to be accrued.14 If 8 or fewer Yes 19 (54) confirmed responses were observed in the first 25 patients, it No 16 (46) was thought appropriate to terminate the study. Accrual was Prior chemotherapy Yes 6 (17) not terminated during the interim analysis. A confidence No 29 (83) interval for the percentage of patients with a confirmed a response was calculated using the exact binomial method. Unless otherwise specified.

518 Copyright © 2008 by the International Association for the Study of Lung Cancer Journal of Thoracic Oncology • Volume 3, Number 5, May 2008 Metastatic Esophageal, Gastric, and Gastroesophageal Cancer majority of patients did receive carboplatin at an AUC of 6. TABLE 2. Maximum Severity Adverse Eventsa During the first 2 cycles of therapy, 17 patients (47%) (36 ؍ received approximately the full dose of all 3 drugs. Patients (%)b (N Response, Progression, and Survival Data. At the time of Grade 2 Grade 3 Grade 4 this report, within this 35-patient cohort, eight patients had Hematologic manifested a confirmed partial tumor response, thus provid- Neutropenia 1 (3) 8 (22) 21 (58) ing a response rate of 23% (95% confidence interval: 10%, Thrombocytopenia 5 (14) 9 (25) 0 40%). There were no complete responses. The median dura- Anemia 6 (17) 0 1 (3) tion of response was 4.7 months (95% confidence interval: Gastrointestinal 2.8, 9.0). Nausea 13 (36) 3 (8) 0 The median survival for the cohort was 8.9 months Vomiting 7 (19) 5 (14) 0 (95% confidence interval: 5.9, 12.8) (Figure 1). The median Diarrhea 4 (11) 6 (17) 0 time-to-cancer progression was 4.2 months (95% confidence Abdominal pain 1 (3) 2 (6) 0 interval: 2.5, 5.1) (Figure 2). At the time of this report, 97% Other of patients have shown evidence of cancer progression, and Infection 0 2 (6) 0 86% have died. For the five patients who remain alive, the Hyponatremia 0 2 (6) 0 median follow-up has been 16.8 months (range, 14.7, 21.6). Neurologic 9 (25) 8 (22) 0 Adverse Events. Adverse event data pertain to all 36 pa- Myalgias 4 (11) 3 (8) 0 tients who received any study drug; the one ineligible patient Dyspnea or evidence of hypoxia 4 (11) 0 1 (3) is included. Only grade 2 or worse events are reported. Of the Pleural effusion 0 0 1 (3) 36 patients treated on this trial, all suffered at least one grade Fatigue 13 (36) 6 (17) 0 Dehydration 1 (3) 2 (6) 0 3 or worse event, and 25 (69%) suffered at least one grade 4 event. Grade 3 events included the following: neutropenia a Grade 1 events are not reported. b If a patient experienced repeated events of the same severity, only one is counted. (22%), thrombocytopenia (25%), vomiting (14%), diarrhea

(17%), neurologic symptoms (22%), and fatigue (17%) with parenthetical percentages referring to the percentage of patients who sustained at least one such event. Grade 4 events included primarily neutropenia (58%). There were no grade 5 events (Table 2).

DISCUSSION This study was undertaken to explore the role of a three-drug regimen that included bortezomib in patients with metastatic esophageal cancer. The agents bortezomib, paclitaxel, and carboplatin were tested for this purpose. The tumor response rate of 23% (95% confidence interval: 10%, 40%) was well below the anticipated threshold to suggest promise, FIGURE 1. The median survival for the cohort was 8.9 and the study was therefore terminated prematurely. In effect, months (95% confidence interval: 5.9, 12.8 months). previously reported tumor response rates in the range of 43% with paclitaxel and carboplatin alone were not observed in the present study.8 One might argue that an end point, other than tumor response rate, should have been used in this trial, particularly because bortezomib may exert only static effects. In reply, it is important to point out that other outcome data from this study also did not suggest that bortezomib was providing clinical benefit. The median survival in previous studies that evaluated the doublet of paclitaxel and carboplatin was 9 months8; and, in the present study, it was 8.9 months— roughly comparable. Thus, based on these overall survival data, this three-drug combination does not appear to be providing a favorable clinical impact over and above what would be expected from paclitaxel and carboplatin alone. FIGURE 2. The median time-to-cancer progression for the This finding underscores the need to evaluate other regimens cohort was 4.2 months (95% confidence interval: 2.5, 5.1 for the treatment of patients with metastatic cancer of the month). esophagus, stomach, and gastroesophageal junction.

REFERENCES into the treatment of lung cancer. Clin Cancer Res 2007;13(15 Suppl): 1. Adams J. The proteasome: a suitable antineoplastic target. Nat Rev 4647s–4651s. ϩ Cancer 2004;4:349–360. 9. Davies AM, McCoy J, Lara PN, et al. Bortezomib gemcitabine/ 2. Richardson PG, Sonneveld P, Schuster MW, et al. Bortezomib or carboplatin results in encouraging survival in advanced non-small cell lung cancer: results of a phase II Southwest Oncology Group trial high-dose dexamethasone for relapsed multiple myeloma. N Engl J Med (S0339). ASCO Annual Proceedings Part I. J Clin Oncol 2006;24:7017. 2005;352:2487–2498. 10. El-Rayes BF, Shields A, Zalupski M, et al. A phase II study of 3. Dispenzieri A. Bortezomib for myeloma – much ado about something. carboplatin and paclitaxel in esophageal cancer. Ann Oncol 2004;15: N Engl J Med 2005;352:2546–2548. 960–965. 4. Valverde CM, Macarulla T, Casado E, et al. Novel targets in gastric and 11. Philip PA, Zalupski MM, Gadgeel S, Hussain M, Shields A. A phase II esophageal cancer. Crit Rev Oncol/Hematol 2006;59:128–138. stud y of carboplatin and paclitaxel in the treatment of patients with 5. Ma C, Mandrekar SJ, Alberts SR, et al. A phase I and pharmacologic advanced esophageal and gastric cancer. Semin Oncol 1997;24(6 Suppl study of sequences of the proteasome inhibitor, bortezomib, in combi- 19):S19–S86. nation with paclitaxel and carboplatin in patients with advanced malig- 12. Keresztes RS, Port JL, Pasmantier MW, et al. Preoperative chemother- nancies. Cancer Chemother Pharmacol 2007;59:207–215. apy for esophageal cancer with paclitaxel and carboplatin: results from 6. Fujita T, Doihara H, Washio K, et al. Antitumor effects and drug a phase II trial. J Thorac Cardiovasc Surg 2003;126:1603–1608. interactions of the proteasome inhibitor bortezomib (PS341) in gastric 13. Therasse P, Arbuck SG, Eishenhauer EA, et al. New guidelines to cancer cells. Anti-Cancer Drugs 2007;18:677–686. evaluate the response to treatment in solid tumors. J Natl Cancer Inst 7. Ocean AJ, Schnoll-Sussman F, Chen XE, et al. Recent results of phase 2000;92:205–216. II study of PS-341 with or without irinotecan in patients with advanced 14. Fleming TR. One sample multiple testing procedures for phase II clinical gastric adenocarcinoma. Proceedings of the American Society of Clin- trials. Biometrics 1982;43:143–151. ical Oncology, abstract #45, Gastrointestinal Cancers Symposium, 2007. 15. Kaplan E, Meier P. Nonparametric estimation from incomplete obser- 8. Davies AM, Lara PN, Mack PC, Gandara DR. Incorporating bortezomib vation. J Am Stat Assoc 1958;53:457–481.