- © 2007 by American Society of Clinical Oncology
Bortezomib Is Active in Patients With Untreated or Relapsed Waldenström's Macroglobulinemia: A Phase II Study of the National Cancer Institute of Canada Clinical Trials Group
- Christine I. Chen,
- C. Tom Kouroukis,
- Darrell White,
- Michael Voralia,
- Edward Stadtmauer,
- A. Keith Stewart,
- John J. Wright,
- Jean Powers,
- Wendy Walsh and
- Elizabeth Eisenhauer
- From the Princess Margaret Hospital, Toronto; Juravinski Cancer Centre, Hamilton; National Cancer Institute of Canada Clinical Trials Group, Kingston, Ontario; Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia; Saskatoon Cancer Centre, Saskatoon, Saskatchewan, Canada; Eastern Cooperative Oncology Group, Philadelphia, PA; Cancer Therapy Evaluation Program, National Cancer Institute, National Institute of Health, Bethesda, MD
- Address reprint requests to Christine I. Chen, MD, Princess Margaret Hospital, 610 University Ave, Suite 5-220, Toronto, ON, Canada M5G 2M9; e-mail: christine.chen{at}uhn.on.ca
Abstract
Purpose To evaluate the efficacy and toxicity of single-agent bortezomib in Waldenström's macroglobulinemia (WM).
Patients and Methods Symptomatic WM patients, untreated or previously treated, received bortezomib 1.3 mg/m2 intravenously days 1, 4, 8, and 11 on a 21-day cycle until two cycles past complete response (CR), stable disease (SD) attained, progression (PD), or unacceptable toxicity. Responses were based on both paraprotein levels and bidimensional disease measurements.
Results Twenty-seven patients were enrolled. A median of six cycles (range, two to 39) of bortezomib were administered. Twenty-one patients had a decrease in immunoglobulin M (IgM) of at least 25%, with 12 patients (44%) reaching at least 50% IgM reduction. Using both IgM and bidimensional criteria, responses included seven partial responses (PRs; 26%), 19 SDs (70%), and one PD (4%). Total response rate was 26%. IgM reductions were prompt, with nodal responses lagging. Hemoglobin levels increased by at least 10 g/L in 18 patients (66%). Most nonhematologic toxicities were grade 1 to 2, but 20 patients (74%) developed new or worsening peripheral neuropathy (five patients with grade 3, no grade 4), a common cause for dose reduction. Onset of neuropathy was within two to four cycles and reversible in the majority. Hematologic toxicities included grade 3 to 4 thrombocytopenia in eight patients (29.6%) and neutropenia in five (19%). Toxicity led to treatment discontinuation in 12 patients (44%), most commonly because of neuropathy.
Conclusion Bortezomib has efficacy in WM, but neurotoxicity can be dose limiting. The slower response in nodal disease may require prolonged therapy, perhaps with a less intensive dosing schedule to avoid early discontinuation because of toxicity. Future studies of bortezomib in combination with other agents are warranted.
INTRODUCTION
Waldenström's macroglobulinemia (WM) is a B-cell lymphoproliferative disorder characterized by a serum monoclonal immunoglobulin M (IgM) paraprotein and bone marrow infiltration of lymphoplasmacytoid cells. As with other indolent lymphomas, treatment is not warranted until symptoms arise. First-line treatment has traditionally included single-agent alkylating agents (response rate [RR], 50% to 80%) or purine analogs (RR, 70% to 80%).1-6 For relapsed or refractory disease, purine analogs lead to responses in approximately 30% of patients.3,7,8 Rituximab, thalidomide, and combination chemotherapy are also active in this disease7,9-13; however, regardless of regimen used, responses tend to be partial (few complete responses [CRs]) and transient. New agents such as bortezomib with novel, biologically based mechanisms of action are, therefore, interesting to evaluate in this disease.
Bortezomib is a potent, reversible proteasome inhibitor. Proteasomes catabolize a wide array of regulatory proteins, leading to inhibition of certain pathways (eg, cyclins) and activation of others (eg, nuclear factor kappa [NF-κ] B–dependent pathways). Inhibition of proteasomes, therefore, may inhibit cell proliferation through multiple mechanisms. Bortezomib has been shown to induce cellular arrest and apoptosis of both a WM cell line model (WM-WSU) and tumor cells isolated from WM patients.14 As well, suppression of NF-κB activity in WM-WSU cells and decreased expression of growth and survival-related kinases is seen in response to bortezomib in vitro.14 Clinical response in a WM patient was noted in an early phase I study of bortezomib in various refractory hematologic malignancies.15 Bortezomib is currently approved for use in relapsed myeloma, where response rates range from 28% to 38%.16,17 The clinical and laboratory similarities of WM to myeloma and other lymphoma subtypes, as well as the in vitro activity of bortezomib in this disease, make this agent of interest to evaluate further in WM. Here, we report results from a phase II multicenter, National Cancer Institute of Canada Clinical Trials Group (NCIC CTG) trial conducted in collaboration with selected centers from the Eastern Cooperative Oncology Group (ECOG) assessing single-agent bortezomib in WM.
PATIENTS AND METHODS
Eligibility
Study patients were at least 18 years old with histologically confirmed WM, either untreated (with a minimum IgM ≥ 20 g/L) or relapsed (minimum IgM ≥5 g/L). Bone marrow studies were not routinely performed at study entry, provided that they had been previously performed to confirm the diagnosis. Relapsed patients were not eligible if refractory to their most recent regimen as defined by progression (PD) during treatment or within 4 weeks of last dose of most recent regimen, or 12 weeks for rituximab. Patients could have received a maximum of two prior chemotherapy regimens, with single-agent rituximab permitted as a third regimen. Patients who had undergone prior stem-cell transplantation and/or radioactive monoclonal antibody therapy were excluded. All patients were symptomatic and requiring treatment (at least one of the following: symptomatic lymphadenopathy; hepatomegaly and/or splenomegaly; anemia [Hb < 110 g/L]; or clinical hyperviscosity syndrome). An ECOG performance status of 2 or lower was required. Patients were excluded for any of the following baseline laboratory values: absolute neutrophil count (ANC) less than 1.0 × 109/L, platelets less than 50 × 109/L, serum creatinine or bilirubin more than 1.5× the upper limit of normal, and AST or ALT more than 2.5× the upper limit of normal. Patients with preexisting neurotoxicity (sensory or motor) grade 2 or higher (National Cancer Institute [NCI] Common Toxicity Criteria version 2.0), pregnant or lactating women, and those with a history of other malignancies (including WM transformed to an aggressive lymphoma) were also excluded. All patients gave written informed consent according to local institutional and/or university human experimentation committee requirements.
Study Design and Treatment
Bortezomib was administered as an outpatient bolus intravenous injection at a dose of 1.3 mg/m2/d on days 1, 4, 8, and 11 in a 21-day cycle. Patients were treated until PD or until two cycles after complete response (CR). The study protocol did not specify a maximum number of cycles. Patients with partial response (PR) could continue until PD or until two cycles after stable PR. Therapy was continued for a minimum of four cycles (12 weeks) for those with stable disease (SD), after which patients continued therapy or discontinued the study at investigator's discretion. Treatment was withheld with grade 3 or worse nonhematologic toxicities until grade 1 or lower, and then bortezomib was resumed at 1.1 mg/m2/d. One further dose reduction to 0.9 mg/m2/d was allowed. Grade 3 neuropathy mandated discontinuation of study protocol with no dose reductions allowed. diarrhea grade 2 or lower was managed empirically with loperamide without dose reductions. For hematologic toxicities, treatment was withheld with grade 3 or worse neutropenia and thrombocytopenia and resumed only when ANC was at least 1.0 × 109/L and platelets were at least 50 × 109/L. Doses were held for the remainder of the cycle if day-8 ANC was less than 0.75 × 109/L or platelets were less than 50 × 109/L. Nadir values of ANC less than 0.5 × 109/L or platelets less than 25 × 109/L required a reduction by two dose levels (to 0.9 mg/m2/d) in the subsequent cycle.
Use of granulocyte colony-stimulating factor and erythropoietin was allowed. Concomitant treatment with corticosteroids, other cytotoxic therapy, or investigational anticancer agents was not permitted.
Assessment of Response and Toxicity
The primary end point of the study was the overall rate of objective response to bortezomib (CRs + PRs). Secondary end points were toxicity, time to PD, SD duration, and response duration. All patients underwent baseline evaluations, including a detailed physical exam with clinical tumor measurements (if applicable), blood counts, liver and renal function tests, serum albumin, serum protein electrophoresis (SPEP) or quantitative immunoglobulins (QIG), urine pregnancy test (for women of child-bearing potential), chest x-ray (CXR), and computed tomography (CT) of the abdomen and pelvis (± chest CT if indicated by CXR findings). Clinical and laboratory assessments (including tumor measurements) were repeated on day 1 of each cycle. Either SPEP or QIG were acceptable to assess serum monoclonal protein (IgM) levels as long as the same method as at baseline was used throughout the study. CXR and CT scans were repeated every two cycles (6 weeks) to assess bidimensional disease.
All patients who had received at least one cycle (3 weeks) of therapy were considered assessable for response. Evaluation of the paraprotein was performed on day 1 of each cycle, with imaging for bidimensional lesions every two cycles. Composite response criteria of both monoclonal protein and bidimensional disease (the latter as per Cheson criteria18) were used. CR was defined as disappearance of the monoclonal protein, confirmed by immunofixation (repeated at least 6 weeks later), plus regression of lymph nodes and spleen to normal size, and a normal bone marrow biopsy with less than 20% lymphocytes. bone marrow aspirates were performed only routinely to confirm CR. PR was defined as a serum monoclonal protein reduction to 50% or less of baseline, confirmed at least 6 weeks later, plus at least a 50% reduction in the sum of the products of the diameters (SPD) of dominant nodes/masses. SD was defined as stable monoclonal protein levels (< 50% increase or decrease from baseline) for at least 6 weeks with no new lesions or sites of disease. PD was defined as any of the following: at least 25% increase in serum IgM, with minimum absolute increase of 5 g/L from baseline; at least 50% increase in the SPD of the dominant nodal and non-nodal sites as compared to baseline; or appearance of a new involved site or lesion. For patients who achieve a CR on treatment, relapse was defined as recurrence of the paraprotein confirmed by immunofixation or at least 25% increase of tumor infiltrate at any previous site, or appearance of a new involved site or lesion.
All patients were assessable for toxicity from time of their first bortezomib dose. Adverse events were assessed at each visit and graded according to the NCI Common Toxicity Criteria (version 2.0).
After protocol therapy was completed, patients were assessed at 4 weeks. Thereafter, for patients off study for PD, follow-up was not required except to document ongoing or late toxicities and death. For patients off study with CR, PR, or SD ongoing, follow-up was every 3 months until relapse or death.
Statistical Considerations
A two-stage design was employed to minimize the number of patients treated should the investigational therapy prove inactive.19 A true response rate of at least 30% was considered of interest, with less than 10% uninteresting. In the first stage, 15 patients were accrued, with at least one response required to proceed to the second stage (additional 10 patients). A significance level (α error) of .10 and power 0.9 were used. Survival was measured from the day of registration to the day of death, and time to PD was from the day of registration to the date PD criteria were first met. Response duration was measured from the time CR or PR criteria were first met until disease relapse or PD. SD duration was measured for those patients who did not go on to alternative salvage treatment, from the time of start of therapy until disease PD.
RESULTS
Patients and Treatment
Twenty-seven patients were enrolled from nine NCIC and two ECOG centers between January 2003 and March 2005. Patient characteristics are listed in Table 1. The median age was 65 years (range, 46 to 87 years) with a male-to-female ratio of 14:13. At study entry, the median serum IgM was 37.6 g/L (range, 13.9 to 79.8 g/L); 21 patients had paraprotein levels more than 30 g/L. The median baseline hemoglobin was 108 g/L (range, 63 to 125 g/L), and 12 patients (44%) had hemoglobin levels less than 100 g/L. Eighteen patients (66%) had measurable bidimensional disease (lymphadenopathy and/or other tumor masses) on baseline CT scans, and the other nine patients had paraprotein elevations only. Fifteen patients (56%) had received prior chemotherapy: six patients received one prior regimen, nine received two prior regimens. Prior treatment included single-agent chlorambucil,11 purine analogs (fludarabine or cladribine),7 CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone)/CVP (cyclophosphamide, vincristine, prednisone),2 and rituximab.5 Nine patients had required previous plasmapheresis, with three receiving regular plasmapheresis for disease control at the time of study entry.
Baseline Characteristics (N = 27)
A median of six cycles (range, two to 39 cycles) were administered; 85% of patients received at least four cycles, and 33% received at least eight cycles. Of the 27 patients enrolled, 12 (44.4%) were discontinued from study for toxicity, 11 (40.7%) for completion of treatment, one (3.7%) for PD, one (3.7%) for symptomatic progression (objective criteria for SD reached), and one (3.7%) for multifactorial debilitation. Dose reductions were required in eight patients (10 cycles), most commonly because of nonhematologic toxicity (neuropathy in four patients; various toxicities in others). Missed doses occurred in 19 patients (30 cycles), most commonly because of hematologic toxicity (neutropenia/thrombocytopenia). Fifty-nine percent of patients received at least 90% of the planned dose intensity.
Efficacy
Response rates are outlined in Table 2. Of the 27 patients enrolled, 21 patients (78%) had at least a 25% decrease in their monoclonal protein on treatment; 12 patients (44%) had at least a confirmed 50% decrease. In one patient, the monoclonal protein became undetectable by electrophoresis, but immunofixation was unavailable to confirm CR. The median reduction in IgM levels was 52 g/L (range, 11 to 98 g/L). Of the three patients who required regular plasmapheresis before study entry, all became plasmapheresis-independent while on study. Of the 18 patients with measurable disease on baseline CT, five patients (28%) had at least a 50% reduction in disease volume (including one with complete resolution of lymphadenopathy and hepatosplenomegaly). With both IgM and bidimensional criteria applied as per study protocol, seven of 27 patients attained a major response (all PR) for an overall objective response rate of 26% (95% CI, 11% to 46%). Response rates were similar between patients who had no prior therapy (three of 12 patients; 25%) and those previously treated (four of 15; 27%). IgM reductions of at least 25% were reached at a median of two cycles (range, one to eight cycles), whereas nodal responses appeared more slowly, with onset of nodal reduction seen in the majority of patients with nodal disease, but PR criteria reached in only five patients at a median of four cycles (range, two to 13 cycles). Hematologic responses (hemoglobin increases of at least 10 g/L) were seen in 18 (66%) of 27 patients.
Responses to Bortezomib (N = 27)
Median progression-free survival for all 27 patients was 16.3 months (95% CI, 14.2 to ∞; Fig 1). Median duration of response (PR) was 10 months (range, 1.4 to 14.9 months), and SD duration was 14.3 months (range, 1.2 to 28.5 months).
Progression-free survival (n = 27; median for all patients, 16.3 months; 95% CI, 14.2 to ∞).
Toxicity
Bortezomib was generally well tolerated, as indicated in Table 3, which lists common and grade 3 to 4 toxicities. The most common nonhematologic toxicities were fatigue, nausea, neuropathy, myalgias, infections (non-neutropenic), diarrhea, and constipation. Twenty patients (74%) developed new or worsening neuropathy while on study: 12 sensory (tingling and/or numbness), two neuropathic pain in legs/feet, and six with mixed sensory and painful neuropathy. Typical onset of neuropathy was between two to four cycles of therapy (range, one to nine cycles), with two patients developing worsening neuropathy 2 and 3 months after stopping therapy. Of all 27 study patients, nine had evidence of baseline neuropathy at study entry (relation to WM not confirmed). However, only four of the 20 neuropathy patients had baseline neuropathy, and only one patient had received prior neurotoxic agents (vincristine). Five patients developed grade 3 neuropathy (no grade 4), four of whom did not have pre-existing neuropathy. Improvement of neuropathy by at least one grade occurred in 15 (75%) of 20 patients, with complete resolution in 14 patients (70%). Time to at least one grade improvement ranged from 2 to 13 months after drug discontinuation in 12 patients, with three patients improving while still receiving bortezomib. Of the five patients with grade 3 neuropathy, two have completely resolved at 16 and 25 months after drug discontinuation, and the other three have improved to grade 2 but persist at 16, 20.6, and 25.2 months off study. neuropathy led to dose reductions in four patients and discontinuation of study drug in five patients in total.
Drug-Related Adverse Events Reported by at Least 15% of Patients and Grade 3 or 4 Events
Low-grade infections were common (n = 13), except one grade 3 non-neutropenic pneumonia requiring hospitalization. The most common infections were upper respiratory tract infections/sinusitis (n = 7) and herpes zoster (n = 3). Biochemical toxicities were uncommon and all grade 1 to 2. There were no thromboembolic events noted.
Drug-related toxicity led to discontinuation of bortezomib in 12 patients (44%). The most common causes were neuropathy and/or neuropathic pain (n = 5) and myalgias (n = 2). Other causes included rash (n = 1), fatigue (n = 1), edema/dyspnea from pleural effusions (n = 1), congestive heart failure (n = 1), and abdominal pain (n = 1).
DISCUSSION
Currently, there is no standard first-line or salvage treatment in WM. On the basis of clinical and preclinical data, the use of bortezomib is a rational choice for investigation in WM, and is attractive for use as a steroid-sparing approach because prolonged and high-dose steroid exposure can cause significant morbidity. In the current phase II study using single-agent bortezomib, we report an overall response rate of 26% when both a 50% reduction in the paraprotein IgM level and bidimensional disease measures are used. In contrast to the majority of other WM studies, which use paraprotein criteria alone in response assessment, we routinely monitored CT scans for bidimensional disease following recommendations of the Consensus Panel on Uniform Response Criteria of the Second International Workshop on WM.20 Our observation of a lag in nodal tumor disease reduction despite prompt IgM reductions contributes to the overall modest composite response rate of 26%. This lag in tumor mass reduction may reflect poor tumor penetration of bortezomib or, alternatively, may suggest that bortezomib acts first by altering protein production or clearance and then later affecting tumor burden.
In contrast to nodal/tumor disease change, paraprotein responses in our study were prompt, with at least a 25% reduction (minor response) seen usually within the first one to two cycles. Overall, 21 patients (78%) demonstrated at least a 25% decrease in paraprotein with treatment (44% with at least a 50% reduction confirmed). Response rates did not appear to differ between previously treated and untreated patients. These results are consistent with the only published report of bortezomib in WM to date.21 In this study, six of 10 WM patients treated with four cycles of bortezomib reached PR. A 50% reduction of the paraprotein was reached at a median of 1 month (range, 0.7 to 2 months). Although CXR and CT scans of the abdomen and pelvis were performed to assess tumor reduction in this cohort, it is not clear whether any responders had bidimensional disease or whether a lag behind paraprotein responses was noted. Treon et al22 reported in abstract form results from a phase II study of 27 WM patients treated with eight cycles of bortezomib. A paraprotein response rate of 84.6% (including minor responses of 25% to 49% reduction in monoclonal protein) was reported, but no evaluation of bidimensional disease with imaging was performed.
Unlike in the Dimopoulos21 and Treon22 studies, we did not restrict the number of cycles of bortezomib to a limited number. One third of our patients received at least eight cycles of therapy, with the majority (85%) receiving at least four cycles (median, six cycles). Although IgM responses are rapid, slower responses of nodal masses suggest that prolonged therapy with bortezomib may be needed to ensure maximal antitumor effect. Alternatively, late responses have been documented months after treatment discontinuation in WM populations treated with rituximab10,23 and fludarabine,24 suggesting that, if similar observations are noted with bortezomib, prolonged treatment may not be required. Our study could not address the optimal strategy; thus, further research using different durations of therapy and longer follow-up are needed.
Bortezomib was generally well-tolerated in our study. neuropathy, however, was more common and severe in this population (74% all grades; 18.5% grade 3) than reported in multiple myeloma (36% all grades; 8% grade 3 to 4)25 or other indolent lymphomas (5% to 8% grade 3 to 4).26,27 This is presumably a result of drug neurotoxicity's exacerbating pre-existing and/or subclinical WM-related neuropathy. There does not appear to be an association with prior neurotoxic chemotherapy, nor does pre-existing clinical neuropathy appear to predispose to a greater rate or severity of neuropathy. Despite the high incidence of neuropathy, the majority of cases were not severe (only five patients with grade 3; no grade 4) and were reversible. Adverse events, mostly neurotoxic, led to dose modifications in 30% and drug discontinuation in 37% of our patients. Because the bidimensional disease in WM is slow to respond, premature drug discontinuation may indeed be limiting drug efficacy in this population. This is in contrast to preliminary data in myeloma suggesting that there is no significant compromise in efficacy in analyses to date with neuropathy-related dose modifications.28 In WM, therefore, approaches to attenuate the neurotoxicity of bortezomib (eg, with concomitant use of neuroprotective agents or less intensive drug administration schedules), rather than routine dose modifications/discontinuations, may be preferable.
In summary, bortezomib as a single agent has efficacy in untreated and previously treated patients with WM, but neuropathy can be dose limiting. The slow nodal/bidimensional disease response observed suggests that prolonged therapy is needed, perhaps with a less intensive dosing schedule to avoid early discontinuation resulting from treatment emergent toxicity. Future investigation into combinations of bortezomib with other cytotoxic or biologic agents to enhance efficacy and decrease toxicity with attenuated dosing, with careful consideration of these limitations, is warranted.
Authors’ Disclosures of potential Conflicts of Interest
Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment: N/A Leadership: N/A Consultant: Christine I. Chen, Ortho Biotech Canada; C. Tom Kouroukis, Ortho Biotech Canada; Michael Voralia, Ortho Biotech Canada; A. Keith Stewart, Ortho Biotech Millenium Pharmaceuticals Stock: N/A Honoraria: Christine I. Chen, Ortho Biotech Canada; Darrell White, Ortho Biotech Canada; A. Keith Stewart, Ortho Biotech Millenium Pharmaceuticals Research Funds: Edward Stadtmauer, Millennium Testimony: N/A Other: N/A
AUTHOR CONTRIBUTIONS
Conception and design: Christine I. Chen, A. Keith Stewart, John J. Wright, Wendy Walsh, Elizabeth Eisenhauer
Administrative support: Jean Powers
Provision of study materials or patients: Christine I. Chen, C. Tom Kouroukis, Darrell White, Michael Voralia, Edward Stadtmauer, A. Keith Stewart
Collection and assembly of data: Christine I. Chen, C. Tom Kouroukis, Darrell White, Michael Voralia, Edward Stadtmauer, Jean Powers, Elizabeth Eisenhauer
Data analysis and interpretation: Christine I. Chen, Jean Powers, Wendy Walsh, Elizabeth Eisenhauer
Manuscript writing: Christine I. Chen
Final approval of manuscript: Christine I. Chen, C. Tom Kouroukis, Darrell White, Michael Voralia, Edward Stadtmauer, A. Keith Stewart, John J. Wright, Jean Powers, Wendy Walsh, Elizabeth Eisenhauer
Footnotes
-
published online ahead of print at www.jco.org on March 12, 2007.
-
Supported by grants from the National Cancer Institute of Canada and funds from the Canadian Cancer Society.
-
Presented in part at the Annual Meeting of the American Society of Hematology, December 4-7, 2004, San Diego, CA, and at the 42nd Annual Meeting of the American Society of Clinical Oncology, June 2-6 2006, Atlanta, GA.
-
Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received July 11, 2006.
- Accepted January 18, 2007.
