M.A. Rodriguez and Jane Winter performed research, analyzed data, and wrote the paper. Robert Pytlik performed research for the paper. Biao Lu analyzed data and wrote the paper. Steven Deitcher analyzed the data.
Vincristine sulfate liposomes injection (Marqibo) in heavily pretreated patients with refractory aggressive non-Hodgkin lymphoma†
Report of the Pivotal Phase 2 Study
Article first published online: 17 JUN 2009
Copyright © 2009 American Cancer Society
Volume 115, Issue 15, pages 3475–3482, 1 August 2009
How to Cite
Rodriguez, M. A., Pytlik, R., Kozak, T., Chhanabhai, M., Gascoyne, R., Lu, B., Deitcher, S. R. and Winter, J. N. (2009), Vincristine sulfate liposomes injection (Marqibo) in heavily pretreated patients with refractory aggressive non-Hodgkin lymphoma. Cancer, 115: 3475–3482. doi: 10.1002/cncr.24359
- Issue published online: 20 JUL 2009
- Article first published online: 17 JUN 2009
- Manuscript Accepted: 12 DEC 2008
- Manuscript Revised: 19 SEP 2008
- Manuscript Received: 15 JUL 2008
- vincristine liposomes;
- vincristine sulfate;
Marqibo, a sphingosomal/cholesterol encapsulation of vincristine sulfate has targeted, increased, and sustained delivery of vincristine to tumor tissues. A phase 2, open-label, single-arm, and multinational study evaluated the efficacy and tolerability of Marqibo as a single agent in patients with multiply relapsed or refractory aggressive non-Hodgkin lymphoma (NHL).
Eligible patients had relapsed or refractory de novo or transformed aggressive NHL and prior treatment with at least 2 multiagent chemotherapy regimens. Marqibo was administered at 2 mg/m2, every 2 weeks, for a maximum of 12 cycles or until toxicity or disease progression.
One hundred and nineteen patients were enrolled and treated on trial. Ninety-six had histological confirmed de novo (N = 89) or transformed (N = 7) aggressive NHL. Median number of cycles was 4 (median dose/cycle 4 mg). Overall response (CR and complete response unconfirmed and PR) was 25% (95% confidence interval [CI], 17, 35), CR and complete response unconfirmed confirmed by external reviewers was 5%. Median overall survival was 6.6 months (Kaplan-Meier estimate, 95% CI, 4.7, 9.8). Grade 3 of 4 neurotoxicity occurred in 32% of patients. All patients had prior neurotoxic agents, and 85% had baseline residual neuropathy symptoms (grades 1-2) from prior treatment.
Marqibo is an active agent in patients with heavily pretreated aggressive NHL, and tolerated at approximately twice the dose intensity of standard vincristine. Its activity supports further investigation as a substitution for vincristine in combination treatment of lymphoid disorders. Cancer 2009. © 2009 American Cancer Society.
Although advances in combination chemotherapy, stem cell transplantation, and the advent of rituximab have improved the cure rates for patients with aggressive non-Hodgkin lymphoma (NHL), nearly 50%-60% of patients diagnosed with aggressive NHL will ultimately fail therapy and require palliation.1 Currently, there is no standard approach to patients who are no longer candidates for curative strategies. New treatment approaches are needed.
Vincristine, a vinca alkaloid, interacts with tubulin and leads to the disruption of microtubules comprising the mitotic spindle apparatus, resulting in metaphase arrest in dividing cells.2, 3 The length of exposure is critical to the efficacy of phase-specific agents such as vincristine. The cytotoxicity of vincristine has been shown to be related to both extracellular concentration and duration of exposure.4 A novel formulation of vincristine encapsulated in a liposomal vesicle composed of sphingomyelin and cholesterol (Marqibo, Hana Biosciences) was designed with the goal of increasing time of exposure to vincristine.5, 6 Marqibo has extended plasma circulation compared with free vincristine and passively targets vincristine to tumors by extravasation through the fenestrations that typify tumor neovasculature.5-7
In this report, we present the results of a phase 2, open-label, single-arm, and multinational study performed to evaluate the efficacy and tolerability of Marqibo as monotherapy in patients with multiply relapsed or refractory aggressive NHL. These data indicate that Marqibo is an active agent in these patients.
MATERIALS AND METHODS
One hundred nineteen patients from 42 centers were enrolled between May 2000 and September 2002. Adult patients (defined as persons 18 years of age or older) were considered for inclusion in this single-arm, open-label study if they had histologically confirmed de novo or transformed aggressive NHL as defined by the WHO.8 The following histologies were included in the definition of aggressive NHL: diffuse large B-cell lymphoma with sclerosis, including primary mediastinal large B-cell lymphoma, immunoblastic B-cell lymphoma, T-cell rich B-cell lymphoma, anaplastic large B-cell lymphoma; peripheral T-cell not otherwise specified; and anaplastic large null-/T-cell lymphoma. Written informed consent was obtained in all cases. Two independent pathologists retrospectively performed central pathology review on all patients. Eligible patients had to meet the following inclusion criteria: failed 2 or more previous combination chemotherapy regimens (at least 1 anthracycline-containing); had at least a minimal response to first-line therapy;an Eastern Cooperative Oncology Group (ECOG) performance status of less than or equal to 3; absolute neutrophil count (ANC) greater than or equal to .5 × 109 /L and platelets of at least 50 × 109 /L (unless decreased because of bone marrow involvement); total bilirubin less than 2 times upper limits of normal (ULN); alanine aminotransferase (ALT) and alkaline phosphatase less than or equal to 4 times ULN; and at least 1 measurable lesion greater than 2 cm in 1 dimension.
Major exclusion criteria were as follows: HIV-positive serology; prior allogenic bone marrow or stem cell transplantation; radiotherapy, chemotherapy, immunotherapy or corticosteroids within 4 weeks of study enrollment; active CNS lymphoma or AIDS-related lymphoma; neuromuscular impairment or grades 3 or 4 sensorimotor neuropathy related to previous chemotherapy.
Other exclusion criteria are as follows: patients receiving therapy with phenytoin or drugs known to inhibit or induce hepatic drug metabolism by cytochrome P450 isoenzymes in the CYP3A subfamily; patients with any previous malignancies with less than a 5-year complete remission interval (except curatively resected basal cell carcinoma or curatively resected in situ carcinoma of the uterine cervix); patients with history of neurologic disorders unrelated to chemotherapy; major surgery within 4 weeks before enrollment; and patients who were pregnant or lactating.
Patients received Marqibo at 2 mg/m2 as an infusion over 1 hour, every 2 weeks. There was no dose capping. Patients received up to 12 cycles of Marqibo in the absence of toxicity or progressive disease. Treatment did not exceed 12 cycles unless it was determined that a patient might obtain a benefit from continued treatment. Determination of potential patient benefit was based on the consensus opinion of the Sponsor, Medical Monitor, and the Investigator. Dose reductions for Marqibo-associated grade 3 and grade 4 toxicities were carried out in .2 mg/m2 decrements. A maximum of 3 10% dose reductions were allowed.
The primary efficacy endpoint of this study was the objective response rate, defined as the percentage of patients whose best response was complete response (CR), complete response unconfirmed, or partial response (PR). Secondary endpoints include duration of response and overall survival time.
Evaluation of Response
Tumor assessments were made by imaging studies and physical exams, performed at baseline (defined as within 3 weeks of initiation of treatment), after every 4 cycles of treatment, and at 2, 8, 16, and 24 weeks after the last cycle of Marqibo. The investigators determined patient responses, and then an independent review panel reviewed each case. The patient's best response was determined by the independent review panel according to International Workshop Response Criteria.9 The independent review panel members were blinded to investigator response evaluation. After completion of the study, patients were followed on an ongoing basis to document survival. The last date of patient follow-up was June 24, 2004.
The primary efficacy endpoint would be summarized as the proportion of patients who achieved overall response (CR, complete response unconfirmed, or PR) based on the independent review panel review in the intention-to-treat (ITT) population. An exact 2-sided 95% confidence interval (CI) was to be constructed around the estimated proportion. The ITT population included all patients who received at least 1 dose of Marqibo (n = 119). Of these, 96 had histologically de novo or transformed aggressive NHL confirmed by the Central Pathology Review and constituted the aggressive NHL population. This histologic group was the primary group of interest and for which the protocol was designed. Patients without response data (missing data) would be treated as failures to achieve overall responses.
As a secondary analysis, the overall response rate and its 2-sided exact 95% CI would be summarized for the per protocol (PP) and aggressive NHL population. The PP population comprised histologically eligible patients who received the study drug and met all prospectively defined inclusion criteria. Median duration of response (defined as the time from the first documentation of response to first documentation of relapse or progression) and overall survival (defined as the time from the date of the first study dose to the date of death) as well as their associated 95% CIs were estimated with the Kaplan-Meier method, respectively, for the ITT, aggressive NHL, and PP populations. Standard descriptive statistics were used to analyze baseline demographics and disease characteristics and presented for both populations.
All patients who received at least 1 dose of the study drug were included in the safety analysis, which evaluated the events that occurred from the initiation of the study drug through 30 days after the discontinuation of study drug. The severity of adverse events and laboratory abnormality was graded in accordance with the National Cancer Institute toxicity criteria grading scale. In addition, the Kaplan-Meier method was also used to estimate the median dose to grade 3 or 4 neuropathy.
Sample Size Determination
The study sample size was determined to ensure that the half-width of the 95% CI to be less than 10% and required a maximum of 100 patients. With a sample size of 100 patients, the half-width of the 95% CI would be no larger than 9.8% no matter what response rate is observed. This is based on the finding that for a given confidence level, the maximum width of the CI occurs when the response rate is 50%.
One hundred and nineteen patients were enrolled and treated with Marqibo. An independent central pathology panel subsequently reviewed all cases, and 96 patients (81%) were determined to have histologically eligible large cell lymphoma de novo or transformed. Histologically ineligible patients (n = 23) were predominantly patients with follicular cell lymphoma or small cell lymphoma that did not meet the criteria for transformation according to the reviewers. An additional 19 patients were ineligible because they failed to meet all prospectively defined inclusion criteria, and, thus, 77 patients met all protocol criteria for eligibility. Patient ineligibility for each criterion were not mutually exclusive and patients were excluded for: inappropriate number of prior combination chemotherapy regimens (n = 5); inappropriate prior anthracycline based therapy (n = 4); lack of minor response to first chemotherapy regimen (n = 3); disease without at least 1 measurable lesion greater than 2 cm in 1 dimension (n = 8); prior radiotherapy, chemotherapy, immunotherapy, or alternative cancer treatment within 4 weeks before study entry (n = 1); missing baseline; and at least 1 postbaseline tumor assessment (n = 6). Therefore, 77 histological eligible patients met all prospectively defined inclusion criteria and are evaluable. These patients comprised the per protocol (PP) population.
Patient characteristics at study entry are listed in Table 1 for all 3 populations defined above (ITT, aggressive NHL, and PP). Similar baseline demographic and baseline disease characteristics were observed for the 3 populations. Patients had widespread disease with high tumor burden at baseline. Forty-eight percent of the ITT population had advanced disease (high-intermediate to high International Prognostic Index scores) at baseline. The patients in this study had experienced multiple relapses and had received extensive prior treatments with multidrug chemotherapeutic regimens (median of 3 prior lines of treatments). Most patients had disease that was resistant to their last treatment. Approximately one third of patients relapsed after autologous stem cell transplantation. At baseline, 78% of patients were anemic, 40% had thrombocytopenia, 15% had neutropenia, and 33% had ANCs <1.5 × 109 /L or platelets <100 × 109 /L. In general, patients in this study had few treatment options available and many were unable to tolerate further treatment with myelosuppressive drugs.
|PP* N = 77 No. (%)||ANHL† N = 96 No. (%)||ITT‡ N = 119 No. (%)|
|Median range [interquartile]||62 [49-71]||60 [47-70]||60 [48-70]|
|Men||42 (55)||52 (54)||64 (54)|
|Women||35 (45)||44 (46)||55 (46)|
|ECOG performance status|
|0-1||60 (78)||73 (76)||94 (79)|
|2-3||16 (21)||22 (23)||24 (20)|
|Unknown||1 (1)||1 (1)||1 (1)|
|Histologic type by central pathology review|
|Diffuse large B-cell lymphoma||53 (69)||68 (71)||68 (57)|
|Primary mediastinal large B-cell lymphoma with sclerosis||4 (5)||5 (5)||5 (4)|
|Immunoblastic B-cell lymphomas||1 (1)||1 (1)||1 (1)|
|T-cell rich B-cell lymphomas||2 (3)||2 (2)||2 (2)|
|Peripheral T-cell lymphoma||1 (1)||1 (1)||1 (1)|
|Anaplastic large null-/T-cell lymphoma||2 (3)||2 (2)||2 (2)|
|Composite lymphoma, DLBCL+||6 (8)||7 (7)||7 (6)|
|Large cell lymphoma, FNA||6 (8)||7 (7)||7 (6)|
|Other, intermediate grade B-cell and large B-cell, follicular & small lymphocytic lymphomas||2 (3)||3 (3)||3 (3)|
|0 (0)||0 (0)||23 (19) (3)|
|Ann Arbor stage|
|I-II||17 (22)||24 (25)||29 (24)|
|III-IV||60 (77)||72 (75)||90 (76)|
|Low risk, 0-1||20 (26)||29 (30)||35 (29)|
|Low-intermediate risk, 2||16 (21)||20 (21)||24 (20)|
|High-intermediate risk, 3||23 (30)||27 (28)||38 (32)|
|High risk, 4-5||17 (22)||19 (20)||20 (17)|
|Unknown||1 (1)||1 (1)||2 (2)|
|No. of prior chemotherapies and/or immunotherapies|
|Median [range] interquartile||3 [2-4]||3 [3-4]||3 [3-4]|
|No. of prior ABMTs||26 (33.8)||31 (32.3)|
|Prior neurotoxic agents|
|Any prior neurotoxic agent||77 (100)||96 (100)||119 (100)|
|Vinca alkaloids||74 (96)||93 (97)||116 (97)|
|Paclitaxel||9 (12)||15 (16)||17 (14)|
|cisplatinum and/or carboplatin||51 (66)||64 (67)||80 (67)|
The primary efficacy endpoint in this study was objective response rate (Table 2). The objective response rate for the ITT population was 25% (95% CI, 18, 34) and was approximately 2.3-fold greater in patients with sensitive disease (objective response rate = 41% [95% CI, 26, 58]) compared with those who were resistant (objective response rate = 18% [95% CI, 10, 28]) to their last treatment regimen. The response by histology was as follows: 27% (24 of 89) for de novo large cell lymphoma; 0% (0 of 7) for transformed lymphoma, and 26% (6 of 23) for indolent (follicular + small lymphocytic) lymphoma.
|PP N = 77 n/N (%)||ANHL N = 96 n/N (%)||ITT N = 119 n/N (%)|
|Objective response [95% CI]||21/77 (27) [18, 39]||24/96 (25) [17, 35]||30/119 (25) [18, 34]|
|Complete response||1/77 (1)||2/96 (2)||4/119 (3)|
|Complete response, unconfirmed||3/77 (4)||3/96 (3)||1/119 (3)|
|Partial response||17/77 (22)||19/96 (20)||22/119 (18)|
|Stable disease||22/77 (29)||26/96 (27)||31 (26)|
|Objective response by disease sensitivity|
|Sensitive* [95% CI]||11/21 (52) [30, 74]||12/27 (44) [25, 65]||16/39 (41) [26, 58]|
|Resistant† [95% CI]||10/56 (18) [9, 30]||12/69 (17) [9, 28]||14/80 (18) [10, 28]|
Response rates were similar in all 3 populations (ITT, aggressive NHL, and PP) in patients with sensitive disease, compared with those who were resistant to their last treatment regimen. Response rates also appeared to be higher in patients that had received fewer previous treatment regimens in all groups. Objective response rate did not differ based on age, sex, or prior autologous bone marrow transplant (ABMT) in the ITT population.
Based upon the independent review panel review, approximately 2 of 3 of patients did not have evidence of progression and their data were censored in the analysis at the last contact. The reason the patients are censored is that they proceeded to subsequent treatment, including additional investigational drugs or stem cell transplantation. Thus, the median duration of response cannot be adequately assessed. However, at the time of the last documented progression, which occurred at 85 days after initiation of the study drug, the Kaplan-Meier estimate for the probability of remaining in response was 52%. This suggests that the median duration of response is approximately 12 weeks, although this analysis is limited by the high proportion of responders that went on to other treatments.
All patients were followed for survival. As of the last survival update, 61% and 64% of the patients died in the ITT population and PP population, respectively. The estimated median survivals were similar between the 2 populations: 6.9 months (95% CI, 4.8, 11.7) for the ITT population and 6.6 months (95% CI, 4.8, 13.1) for PP population, based on the Kaplan-Meier analysis (Fig. 1).
One-year overall survival rates were 33% and 61%, respectively, for patients who had not and had achieved overall response in the ITT population (Fig. 2). Patients who had disease that was sensitive to the last qualifying therapy had a median survival of 13.1 months, versus 5.1 months for patients with resistant disease in the ITT population (Fig. 3).
All patients who received at least 1 study dose in this study (n = 119) are eligible for safety evaluation. The median number of cycles received was 4 (range, 1, 20). Sixty-nine (58%) patients received 4 or more cycles of therapy, including 11 (9%) patients who received 10 or more cycles of Marqibo. The median total dose (as the sum of study dose across all cycles) delivered was 7.9 mg/m2 (range, 1.9, 39.6).
No treatment-related deaths occurred on study. Adverse events were typically grade 1 and 2. Treatment withdrawal due to adverse events occurred in only 15% of patients. Dose modifications occurred in 24% of patients, and dose delays were required by 20% of patients. Of the 24 patients who had dose delays, 18 were discontinued from treatment without receiving another dose of Marqibo. Treatment was discontinued because of progressive disease in 8 patients and neuropathy in 7 patients. Neuropathy was the main reason for discontinuation or modification of Marqibo treatment. Common grade 3 and 4 adverse events are listed in Table 3.
|Adverse event||Percentage of Patients, N = 119|
|Grade 3||Grade 4|
|Any neurologic symptom||29||3|
Grade 3 neuropathy occurred in 29% of patients, all of whom had baseline neuropathy grade 2. Grade 4 neuropathy was seen only in 3% of patients, and they also had grade 2 neuropathy at baseline. In general, patients on trial experienced neurotoxicity that was characterized as 1 grade worse than their baseline assessment. The projected median cumulative dose of Marqibo resulting in grade 3 of 4 neurotoxicity was 21 mg/m2 (95% CI, 10, 31), which equates to approximately 11 cycles of Marqibo. Most patients who responded to Marqibo showed clinically evident tumor responses after 1 dose of Marqibo.
The incidence of grade 3 and 4 hematologic toxicity is noted in Table 3. Grade 4 neutropenia was seen in 9 patients who at baseline are normal (6 pts) and already had grade 1 (2 pts) and grade 3 (1 pt) neutropenia. Bone marrow involvement at baseline and prior ABMT increased the risk of developing myelosuppression. The use of hematologic growth factor and transfusion support while receiving Marqibo was less than 22% (filgrastim 13%, erythropoietin 22%, transfusions 19%). Febrile neutropenia was reported in 3% of patients.
Marqibo is the encapsulation of vincristine in lipid vesicles composed of sphingomyelin and cholesterol called optisomes.5 It is the unique composition of optisomes that allows them to effectively contain lipophilic drugs like vincristine. In preclinical studies, Marqibo provided a targeted, increased, and sustained delivery of vincristine to tumor tissue.5, 6 This delivery resulted in prolonged availability and activity of vincristine and greater anticancer properties than free vincristine in preclinical solid tumor and lymphoma models.
The results of this study indicate that single-agent Marqibo is an active agent in patients with relapsed or refractory aggressive NHL. An objective response rate of 25% was observed in this study population that comprised heavily pretreated patients with few therapeutic options and with multiply relapsed or refractory disease. Moreover, 5% of patients receiving Marqibo had complete responses (CR and complete response unconfirmed). Greater objective response rate was observed in patients with sensitive disease. Patients with sensitive disease had an approximately 3-fold greater objective response rate than those who were refractory or resistant to their last qualifying therapy.
When Marqibo was administered at approximately twice the dose intensity of standard free vincristine (2 mg/m2 over 1 hour every 2 weeks with no upper dose limit), patients in this study exhibited a similar toxicity profile to standard vincristine.2, 10 Neurotoxicity was the most common adverse event, with severe (grade 3 or 4) neurotoxicity noted in only 32% of patients. The projected median cumulative dose of Marqibo resulting in grade 3 of 4 neurotoxicity, however, was 21 mg/m2 (95% CI, 10, 31), which equates to approximately 11 cycles of Marqibo. Marqibo thus has a good tolerance profile, given that all patients had prior exposure to neurotoxic drugs (eg, vincristine), that 85% of patients at study entry had residual neurologic abnormalities from their prior treatments, and that Marqibo was administered without dose limitation. In fact, 86% of patients had received between 2 and 5 prior chemotherapy regimens containing known neurotoxins. Constipation was commensurate with vincristine exposure and was clinically managed with stool softeners. Severe myelotoxicity, gastrointestinal toxicity, and alopecia were rare.
This phase 2 trial used a rigorous external review design and a larger number of patients than other studies of single-agent therapies in patients with aggressive NHL.1 The use of an independent review panel in this unblinded single-arm study provided an objective assessment of response9 and alleviated any potential investigator bias. The independent review panel also ensured that response criteria were consistently applied in the individual institutions in this multicenter trial. In all cases, the concordance between independent review panel and investigator observations was excellent (>85%).
To address the complex histology of aggressive NHL and variability between study sites, a retrospective central pathology review of patient histological diagnoses was performed.
Investigators determined histological eligibility for entry into the study based upon review of patient pathology reports on site. All cases were subsequently reviewed by 1 of 2 independent pathologists after patient enrollment in the study. The rate of histologic eligibility (81%) reflects the complex histology of aggressive NHL.
This study demonstrates clinically relevant activity of Marqibo in patients with relapsed aggressive NHL and with poor prognoses. Furthermore, Marqibo has a safety profile comparable to free vincristine when administered at approximately twice the dose intensity. Although it remains to be seen what effect Marqibo will have in other types of cancers and in combination with other agents, data suggest that objective response rate greater than 95% are achieved when Marqibo is used in CHOP and R-CHOP regimens in place of generic vincristine in previously untreated patients with aggressive NHL.11-13 Overall the data suggest that Marqibo has potential as a nonmyelosuppressive option in patients with aggressive NHL for whom other therapeutic options have been exhausted.
Conflict of Interest Disclosures
The authors made no disclosures.
- 10Antimicrotubule agents. In: DeVitaJrVT, HellmanS, RosenbergSA, eds. Cancer: Principles & Practice of Oncology. 6th ed. Philadelphia: Lippincott Williams & Wilkins; 2001: 431-452., .
- 11A phase II study of liposomal vincristine in CHOP with rituximab for the elderly patients with untreated aggressive B-cell non-Hodgkin's lymphoma (NHL) [abstract]. Proc Am Soc Clin Oncol. 2002; 21: 284a. Abstract 1132., , , et al.
- 12A phase II study of liposomal vincristine in CHOP with rituximab for patients with untreated aggressive B-cell non-Hodgkin's lymphoma (NHL): a safe and effective combination [abstract]. Blood. 2002; 100: 92a. Abstract 338., , , et al.
- 13Sphingosomal vincristine in CHOP is a promising new treatment for elderly, as well as poor prognosis patients with aggressive non-Hodgkin's lymphoma (NHL): follow-up results of a phase II study [abstract]. Presented at 2004 ASCO Annual Meeting; June 5, 2004; New Orleans, LA. Abstract 8080., , , et al.