Blood Journal
Leading the way in experimental and clinical research in hematology

Clinical factors predictive of outcome with bortezomib in patients with relapsed, refractory multiple myeloma

  1. Paul Gerard Guy Richardson,
  2. Bart Barlogie,
  3. James Berenson,
  4. Seema Singhal,
  5. Sundar Jagannath,
  6. David Irwin,
  7. S. Vincent Rajkumar,
  8. Teru Hideshima,
  9. Hugh Xiao,
  10. Dixie Esseltine,
  11. David Schenkein,
  12. Kenneth C. Anderson, and
  13. for the SUMMIT Investigators
  1. From the Dana-Farber Cancer Institute, Boston, MA; University of Arkansas for Medical Sciences, Little Rock, AR; Cedars-Sinai Medical Center, Los Angeles, CA; Northwestern Memorial Hospital, Chicago, IL; St Vincent's Comprehensive Cancer Center, New York, NY; Alta Bates Cancer Center, Berkeley, CA; Mayo Clinic, Rochester, MN; and Millenium Pharmaceuticals, Inc., Cambridge, MA.

Abstract

Bortezomib, a potent and reversible proteasome inhibitor, affects the myeloma cell and its microenvironment, resulting in down-regulation of growth and survival signaling pathways and durable responses in patients with relapsed and refractory myeloma. Potential associations between baseline parameters and outcomes with bortezomib were explored in 202 patients who received bortezomib 1.3 mg/m2 twice weekly for 2 weeks every 3 weeks for up to 8 cycles in a phase 2 trial. Using European Group for Blood and Marrow Transplantation criteria, the response rate (complete or partial response) to bortezomib alone was 27% and was not associated with sex, race, performance status, isotype, chromosome 13 deletion, number or type of previous therapies, or concentration of hemoglobin or β2-microglobulin. By multivariate analysis, factors associated with lower response were being age 65 or older versus younger than 65 (19% vs 32%; P < .05) and plasma-cell infiltration in bone marrow greater than 50% versus 50% or less (20% vs 35%; P < .05). Factors that may be indicative of tumor burden (bone marrow plasma-cell infiltration greater than 50%, hypoalbuminemia, thrombocytopenia) were predictive of overall survival. Chromosome 13 deletion and elevated β2-microglobulin, generally considered poor prognostic factors, were not predictive of poor outcome with bortezomib in this study.

Introduction

Prognostic factors may be useful to help select patients with multiple myeloma who are more likely to benefit from certain types of therapy. Factors that have been identified as predictive of a poor outcome in patients with multiple myeloma at presentation are listed in Table 1.1-12 At relapse, adverse prognostic factors have included increased age,13,14 elevated creatinine (177 μM; 2 mg/dL),14 chromosome 13 deletion or other cytogenetic abnormalities,13-15 increased plasma-cell labeling index,15 and elevated levels of β2-microglobulin and C-reactive protein.13,15

View this table:
Table 1.

Factors previously reported to be associated with poor outcomes in patients with MM at presentation

Bortezomib, a potent, specific, and reversible inhibitor of the proteasome, offers a novel approach to the treatment of multiple myeloma.16 In an open-label, phase 2 clinical trial (Study of Uncontrolled Myeloma Managed with Proteasome Inhibition Therapy [SUMMIT]), 202 patients with relapsed and refractory myeloma were treated with bortezomib.17 Durable responses were reported in this heavily pretreated population who had received a median of 6 previous therapies. The response rate (complete response [CR] + partial response [PR]) was 27%, and the median time to response (TTR) was 1.3 months (approximately 2 cycles).17 With extended follow-up, the median duration of response (DOR) was 12.7 months, median time to progression (TTP) among all patients was 7.0 months, and median overall survival (OS) was 17.0 months.18

New therapies for myeloma with novel mechanisms of action may have the ability to change the course of the disease or to be active in patients in whom other therapies fail. Importantly, their use should be prioritized in patients most likely to benefit from them. The objective of this study was to perform an analysis of baseline demographic and disease parameters to determine their value in predicting outcome for patients with relapsed and refractory multiple myeloma being treated with bortezomib. Univariate analyses were used to identify single factors that appeared to be associated with outcome, then multivariate analyses were used to eliminate factors that may confound univariate analysis.19

Patients, materials, and methods

Patients

Patients with multiple myeloma who had relapsed after initial chemotherapy and who had been refractory to their most recent salvage chemotherapy were enrolled into the SUMMIT trial, the methodology of which has been previously published.17 Main inclusion criteria were Karnofsky performance score (KPS) of 60 or greater, hepatic transaminases within 3 times the upper limit of normal (ULN), serum total bilirubin concentration within 2 times ULN, creatinine clearance greater than or equal to 0.5 mL/s (exceptions were allowed for levels < 0.5 but ≥ 0.7 mL/s if caused by significant myelomatous involvement of the kidney), platelet count greater than or equal to ×109/L (or 30 × 109/L if extensive bone marrow infiltration by myeloma was present), hemoglobin level greater than or equal to 80 g/L (8 g/dL), and absolute neutrophil count greater than or equal to 1.0 × 109/L (except where the bone marrow was extensively infiltrated and absolute neutrophil count ≥ 0.5 × 109/L was allowed). All patients provided written informed consent before entering the study, which was performed in accordance with the Declaration of Helsinki and was approved by the institutional review board at each participating center.

Treatment and evaluation of outcome

Bortezomib 1.3 mg/m2 was administered by intravenous bolus over 3 to 5 seconds on days 1, 4, 8, and 11 of a 21-day cycle for up to 8 cycles. Patients with a suboptimal response, defined as progressive disease after 2 cycles or no change after the first 4 cycles of bortezomib, could receive oral dexamethasone 20 mg on the day of and day after bortezomib treatment.

Responses were categorized according to the criteria of the European Group for Blood and Marrow Transplantation,20 with the addition of a near CR category (stable bone disease, normal serum calcium levels, and 100% disappearance of M-protein by electrophoresis but immunofixation positive).

The response rate (CR + PR) to bortezomib alone was examined in 193 evaluable patients because 9 of the 202 patients enrolled into the SUMMIT trial were excluded from the intent-to-treat population for efficacy analyses. These patients were excluded because they were enrolled in the study with nonmeasurable disease, and a priori the statistical analysis plan specified that only patients with measurable disease would be included in the intent-to-treat population given that the internal review committee could not assess response. DOR was assessed only in patients who responded to bortezomib alone (n = 53). TTR and TTP were assessed for patients receiving bortezomib alone, whereas OS was assessed for all patients, regardless of the addition of dexamethasone. For each of these outcomes, the following factors were assessed: age (younger than 65 years, 65 years or older), sex, race (white, black, other), body surface area (≤ 2 m2, > 2 m2), KPS (≤ 70, 80, ≥ 90), number of previous therapies (≤ 5, ≥ 6), previous treatment with thalidomide (yes, no), previous stem-cell transplantation (yes, no), percentage of plasma-cells in bone marrow biopsy (≤ 50%, > 50%), abnormal cytogenetics, chromosome 13 deletion by conventional cytogenetics, hemoglobin concentration (< 105 g/L [10.5 g/dL], ≥ 105 g/L [10.5 g/dL]), and M-protein isotype (immunoglobulin G [IgG], IgA, light chain). Serum levels of β2-microglobulin, albumin, platelets, and C-reactive protein were assessed as continuous variables without specific cutoff values. The effect of disease stage on outcome according to the new International Staging System was assessed in the univariate and multivariate settings using the following classification: stage I, β2-microglobulin less than 3.5 mg/L and serum albumin more than or equal to 3.5 g/dL; stage II, β2-microglobulin less than 3.5 mg/L and albumin less than 3.5 g/dL, or β2-microglobulin 3.5 to less than 5.5 mg/L; stage III, β2-microglobulin more than or equal to 5.5 mg/L.19

Statistical analyses

In the univariate analysis assessing response rate (CR + PR), Fisher exact test was performed for categoric factors, and logistic regression was performed for continuous variables (β2-microglobulin level, serum albumin level, platelet count, and C-reactive protein level). In the univariate analysis assessing TTR, DOR, TTP, and OS, the Cox proportional hazards model was applied to calculate a hazard ratio to determine the relative risk of an event at any point in time associated with a 1-unit change in the variable. A hazard ratio greater than 1 indicated increased risk, whereas a ratio less than 1 indicated decreased risk. In the Cox regression analyses assessing TTR, TTP, and OS, a backward elimination method was used to derive the final analysis model. In the analysis of DOR (n = 53), a forward stepwise model selection method was used.

In the multivariate analysis assessing response rate, logistic regression was performed for all factors to determine the relative contribution of each in predicting response. All factors were analyzed in a stepwise model selection procedure to determine a minimal set of prognostic factors. In the multivariate analysis assessing TTR, DOR, TTP, and OS, the Cox proportional hazards model was used to analyze all factors to determine the relative contribution of each factor. All factors were analyzed through a model selection procedure to determine a minimal set of prognostic factors. P less than .05 was considered statistically significant, and P greater than or equal to .05 but less than .10 was considered marginally significant.

Results

The 202 patients enrolled in SUMMIT had been heavily pretreated with a median of 6 previous therapies (range, 2-15), and 91% were refractory to their last therapy before bortezomib. Previous therapy included steroids in 99.5% of patients, alkylating agents in 92%, thalidomide in 83%, anthracyclines in 81%, and stem-cell transplantation in 64%. Ninety-two percent of the patients were treated with 3 or more therapies, excluding stem-cell transplantation.17

Response rate

Of 202 patients, 193 patients had measurable disease and were included in the analyses of response to bortezomib monotherapy. The response rate (CR + PR) was 27% (53 of 193 patients).17 In univariate analysis, the only factors that were associated with a significantly lower response rate to bortezomib were bone marrow plasma-cell infiltration greater than 50% compared with 50% or less (20% vs 35%; P = .030) and abnormal cytogenetics compared with normal cytogenetics (19% vs 35%; P = .047), but not chromosome 13 deletion. Although there was no significant association between response rate and light-versus heavy-chain disease (9 [33%] of 27 vs 44 [27%] of 163; P = .49), 8 of the 9 responses in patients with light-chain disease were CRs, whereas 11 of the 44 responses in patients with heavy-chain disease were CRs. Marginal significance was noted for age 65 (19% vs 32%; P = .064) and white race (24% white vs 48% black and 33% other; P = .064). The response rate to bortezomib was not significantly affected by sex, body surface area, KPS, number or type of previous therapy, myeloma type, chromosome 13 deletion, and hemoglobin or β2-microglobulin concentration (Table 2). By multivariate analysis (Table 3), only 2 factors—age greater than or equal to 65 years (P = .03) and bone marrow plasma-cell infiltration greater than 50% (P = .03)—were significantly associated with a lower response rate to bortezomib after stepwise selection analysis.

View this table:
Table 2.

Effect of demographic variables, disease characteristics, and previous treatment on response (CR + PR) to bortezomib in univariate analysis

View this table:
Table 3.

Factors significantly predictive of outcome with bortezomib in multivariate analysis

TTR, DOR, TTP, and OS

By univariate analysis, the only factors associated with longer TTR were older age (P = .042) and abnormal cytogenetics (P = .047) but not chromosome 13 deletion (Table 4). By multivariate analysis, however, abnormal cytogenetics dropped out, whereas older age (P = .009) and bone marrow plasma-cell infiltration greater than 50% (P = .021) were found to be associated with longer TTR. An increased C-reactive protein concentration was marginally significant by multivariate analysis (P = .050).

View this table:
Table 4.

Factors significantly associated with TTR, DOR, TTP, and OS in univariate analysis

Factors were assessed to determine their association with DOR in this patient population. Previous treatment with thalidomide (P = .033), hypoalbuminemia (P = .001), low KPS (P = .004), and abnormal cytogenetics (P = .035), but not chromosome 13 deletion, were associated with shorter DOR in the univariate analysis. Factors that remained significantly associated with shorter DOR in the multivariate assessment were hypoalbuminemia (P = .003) and low KPS (P = .01).

Similarly, factors associated with shortened TTP were assessed by univariate analysis. Patients with bone marrow plasma-cell infiltration greater than 50% (P = .009), increased β2-microglobulin concentration (P = .02), thrombocytopenia (P = .025), increased C-reactive protein concentration (P = .037), anemia (P = .005), abnormal cytogenetics (P < .001), higher stage (P = .024), and chromosome 13 deletion (P = .013) had shorter TTP. Factors associated with shortened TTP that remained significant when assessed by multivariate analysis were increased C-reactive protein concentration (P = .011) and abnormal cytogenetics (P = .02), but not chromosome 13 deletion.

Although many factors were found to be associated with shorter OS by univariate analysis, only 4 were significantly associated with shorter survival in the multivariate analysis: bone marrow plasma-cell infiltration greater than 50% (P = .028), hypoalbuminemia (P < .001), thrombocytopenia (P = .024), and low KPS (P = .034).

Discussion

Bortezomib can induce objective responses in patients with relapsed and refractory multiple myeloma. Results of the multivariate analysis confirmed that the following baseline factors did not influence clinical response rate with bortezomib: number or type of previous therapies, sex, body surface area, chromosome 13 deletion, and hemoglobin and β2-microglobulin concentrations. Although the overall response rate was similar in patients with light- and heavy-chain disease, most of the responses in patients with light-chain disease were CRs. It is important to note that the detection of free light chains in urine depends on renal function and on the sensitivity of the method used. Therefore, although it appears that patients with light-chain disease achieved a higher CR rate, this conclusion is difficult to confirm. Conversely, the lack of effect of chromosome 13 abnormalities on response to bortezomib is noteworthy because this cytogenetic abnormality has been shown to be an important negative prognostic factor of response in relapsed myeloma treated with certain other agents.15

Advanced age has been associated with poor response to treatment in patients with relapsed multiple myeloma who were treated with thalidomide.13 Age was also predictive of response to bortezomib but did not affect DOR, TTP, or OS. These results suggest that age alone should not represent a barrier to treatment with bortezomib. Although older patients may take longer to respond, the data suggest that elderly patients who do respond may achieve DOR, TTP, and OS benefits similar to those of younger patients.

The other significant factor in the multivariate analysis of response rate was greater than 50% plasma-cell infiltration in the bone marrow. In general, factors that correlate with high tumor burden (ie, hypoalbuminemia, bone marrow plasma-cell infiltration greater than 50%, and thrombocytopenia) were associated with shorter DOR and OS with bortezomib.

Clinical outcome with bortezomib was independent of the number of previous lines of treatment. Based on the Mayo Clinic experience in patients with relapsed myeloma, irrespective of the type of treatment, DOR steadily decreases with each successive treatment regimen, ranging from 10 months with the first therapy to approximately 3 months with 6 or more previous therapies.2 Patients who received bortezomib in the SUMMIT trial had received a median of 6 previous lines of therapy and had a DOR of approximately 13 months, but neither the type nor the number of previous treatments influenced DOR. Indeed, TTP with bortezomib was 2 to 4 times longer in duration than that reported with the last therapy before enrolling patients in the study.17

The importance of prognostic factors in the pretreatment evaluation of myeloma has gained even more significance with the creation of the new International Staging System (ISS).19 The new system uses only 2 factors for staging patients at diagnosis: serum albumin as a correlate of rapid myeloma growth, and β2-microglobulin as a marker of tumor burden. In the univariate analysis, shorter DOR and OS were individually associated with lower albumin, and shorter TTP and OS were associated with higher β2-microglobulin, but response rate, TTR, and DOR with bortezomib treatment were independent of stage as defined by the ISS using albumin and β2-microglobulin. In the univariate analysis, stage was predictive of TTP and OS.

The ability to generally predict response, TTR, DOR, TTP, and OS is meaningful for the physician and the patient in making treatment decisions. In a phase 3 randomized trial, Assessment of Proteasome Inhibition for Extending Remissions (APEX), in which bortezomib therapy was compared with high-dose dexamethasone in patients who had received 1 to 3 previous lines of therapy, bortezomib demonstrated significantly greater benefit in terms of TTP and OS at the preplanned interim analysis, resulting in an independent committee decision to recommend early closure of the dexamethasone arm and to allow all patients access to bortezomib.21 Prognostic factors and the prognostic value of certain genetic markers will be further explored in APEX.

A fascinating research focus derived from this work has been the exploration of genetic markers that may aid in identifying patients with a high potential for response to specific therapies.22 Genetic markers from pretreatment bone marrow aspirates of myeloma cells are being studied by microarray analysis in subsets of patients treated with bortezomib in phase 2 and 3 studies to validate this methodology as a promising tool in targeted therapy. Initial research from SUMMIT has shown the feasibility of this approach, and preliminary data suggest that a highly significant gene expression signature can distinguish responders from nonresponders.23 These data may provide yet another valuable method to identify patients with the potential to obtain clinical benefit from bortezomib.

Positive clinical outcomes in patients with relapsed and refractory multiple myeloma treated with bortezomib appeared to be independent of many patient and disease factors historically considered to be adverse prognostic factors in myeloma. Importantly, the type or number of previous therapies and an increased baseline β2-microglobulin were not associated with poor response to bortezomib or to unfavorable changes in time-to-event parameters. Although older patients had a lower response rate to bortezomib in this study, their duration of response and survival were not negatively affected.

It is recognized that the analyses and models used for each of the several dependent variables (response rate, TTP, OS, TTR, DOR) should be considered exploratory, and the conclusions are subject to confirmation by further clinical studies of bortezomib in this patient population. Ongoing studies will provide further understanding regarding the use of prognostic factors to identify patients most likely to benefit from bortezomib therapy. Pharmacogenomics in conjunction with these clinical factors may provide insight not only into response to bortezomib but also into rational designs for combinations.

Appendix

The names and affiliations of all individuals comprising the SUMMIT Investigators are as follows:

Paul Gerard Guy Richardson, MD, Dana-Farber Cancer Institute, Boston, MA; Bart Barlogie, MD, PhD, University of Arkansas for Medical Sciences, Little Rock, AR; James Berenson, MD, Cedars-Sinai Medical Center, Los Angeles, CA; Seema Singhal, MD, Northwestern Memorial Hospital, Chicago, IL; Sundar Jagannath, MD, St Vincent's Comprehensive Cancer Center, New York, NY; David Irwin, MD, Alta Bates Cancer Center, Berkeley, CA; S. Vincent Rajkumar, MD, Mayo Clinic, Rochester, MN; Gordon Srkalovic, MD, PhD, Cleveland Clinic Foundation, Cleveland, OH; Melissa Alsina, MD, H. Lee Moffitt Cancer Center, Tampa, FL; Raymond Alexanian, MD, M.D. Anderson Cancer Center, Houston, TX; David Siegel, MD, Carol G. Simon Cancer Center, Morristown, NJ; Robert Z. Orlowski, MD, PhD, University of North Carolina, Chapel Hill, NC; David Kuter, MD, Massachusetts General Hospital, Boston, MA; Steven Limentani, MD, Charlotte Medical Clinic, Charlotte, NC; Teru Hideshima, MD, PhD, Dana-Farber Cancer Institute, Boston, MA; Hugh Xiao, PhD, Millennium Pharmaceuticals, Inc., Cambridge, MA; Dixie Esseltine, MD, Millennium Pharmaceuticals, Inc., Cambridge, MA; David Schenkein, MD, Millennium Pharmaceuticals, Inc., Cambridge, MA; and Kenneth C. Anderson, MD, Dana-Farber Cancer Institute, Boston, MA.

Footnotes

  • Reprints:

    Paul G. G. Richardson, Department of Adult Oncology, Dana-Farber Cancer Institute, 44 Binney St, Dana 1B02, Boston, MA 02115; e-mail: paul_richardson{at}dfci.harvard.edu.
  • Prepublished online as Blood First Edition Paper, July 14, 2005; DOI 10.1182/blood-2005-02-0691.

  • A complete list of the members of the Study of Uncontrolled Myeloma Managed with Proteasome Inhibition Therapy (SUMMIT) Investigators appears in the “Appendix.”

  • Supported by Millennium Pharmaceuticals, Inc.

  • Several of the authors have declared a financial interest in and/or are employed by Millenium Pharmaceuticals, whose product was studied in the present work. P.G.G.R. is on the Speakers' Bureau and the Advisory Board for Millennium Pharmaceuticals, Inc. J.B. receives grants and honoraria from Speakers' Bureaus and consultant fees from Millennium Pharmaceuticals, Inc., and is on the Advisory Board of Millennium Pharmaceuticals, Inc. S.J. is a consultant on the Speakers' Bureau for Millennium Pharmaceuticals, Inc. S.V.R. has received grant support from Millennium Pharmaceuticals, Inc., to conduct clinical trials in myeloma at the Mayo Clinic. H.X. is a contractor-biostatistician working with Millennium Pharmaceuticals, Inc. D.E. is a full-time employee of and owns stock in Millennium Pharmaceuticals, Inc. D.S. is a full-time employee of and owns stock in Millennium Pharmaceuticals, Inc. K.C.A. receives grant support from Millennium Pharmaceuticals, Inc.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.

  • Submitted February 18, 2005.
  • Accepted June 10, 2005.

References

View Abstract