Efficacy and safety of lenalidomide in intermediate-2 or high-risk myelodysplastic syndromes with 5q deletion: results of a phase 2 study

Myelodysplastic syndromes (MDS) are clonal stem cell disorders predominating in the elderly, characterized by ineffective hematopoiesis leading to blood cytopenias, especially anemia, and often evolving to acute myeloid leukemia (AML).¹  Main prognostic factors of MDS are the number of cytopenias, percentage of marrow blasts, and bone marrow cytogenetics, which are combined in an International Prognostic Scoring System (IPSS) that distinguishes 4 subgroups with significantly different risk of progression to AML and survival: low, intermediate-1 (int-1, often grouped together as lower-risk MDS) and intermediate-2 (int-2), and high (often grouped together as higher-risk MDS).² 

del 5q is found in approximately 15% of MDS.³  Approximately two-thirds of MDS with del 5q have low or int-1 IPSS and relatively favorable prognosis but generally profound anemia.^4-6  The recent advent of lenalidomide has been a major breakthrough in those cases, as this drug leads to red blood cell (RBC) transfusion independence in two-thirds of them.^6-8  Kinetics of response to lenalidomide (ie, occurrence of neutropenia and thrombocytopenia, often profound, followed by hematopoietic recovery) and in vitro studies suggested a selective inhibitory effect of lenalidomide on the clone with 5q deletion.^8-11 

The remaining third of MDS with del 5q, however, have int-2 or high IPSS because of an increase in marrow blasts and/or chromosomal abnormalities, often complex, in addition to del 5q.³  Prognosis is poor in those patients, especially in case of complex karyotype, associated with poor response to classic anthracycline cytarabine (AraC) chemotherapy and a high risk of relapse after allogeneic stem cell transplantation.^5,12  In addition, although hypomethylating agents have demonstrated a clear benefit in higher-risk MDSs in general,¹³  especially in patients with −7/del 7q, preliminary results, including those from our group, suggest that their results as single agents may be less favorable in higher-risk MDSs with del 5q.^14-16 

The apparently exquisite sensitivity of clones with del 5q to lenalidomide makes this drug also of potential interest in higher-risk MDSs with del 5q, especially as del 5q generally appears to be an early genetic event in such cases, present in all clonal cells even if there is complex karyotype.¹⁷  Response to lenalidomide is, however, poorly known in this population. In the lenalidomide pivotal study in MDSs with del 5q (MDS-003 trial), only 8 patients were found on review to have high or int-2 IPSS, and 3 of them had significant hematologic improvement.⁸  Only a few other patients with higher-risk MDSs or AML with del 5q treated with lenalidomide have been reported.^18-22 

The Groupe Francophone des Myélodysplasies therefore conducted a phase 2 trial with lenalidomide in high-risk and int-2 risk MDS with del 5q.

Inclusion criteria were as follows: (1) age more than 18 years; (2) documented diagnosis of MDS according to French-American-British classification²³  and World Health Organization (WHO)²⁴  criteria, with IPSS intermediate-2 or high-risk,²  also including chronic myelomonocytic leukemia with white blood cell (WBC) count less than 13 000/mm³ and refractory anemia with excess blasts in transformation (RAEB-T); (3) del 5q(31) by conventional cytogenetics, with or without additional chromosomal changes; (4) platelet count more than 25 000/mm³ and absolute neutrophil count (ANC) more than 500/mm³; (5) patients must have signed an informed consent form; and (6) negative serum or urine pregnancy test in women of childbearing potential, in accordance with the Declaration of Helsinki. The trial had been approved by ethical committees at all participating institutions in France and in Belgium. The Groupe Francophone des Myelodysplasies was the sponsor of this trial, for which Celgene (Paris, France) provided the drug and a scientific grant.

Conventional cytogenetics were performed by analyzing G- and R-banded metaphase chromosomes in at least 20 mitoses, and results were interpreted using International System Cytogenetic Nomenclature.²⁵ 

The main objective of the trial was to assess response to lenalidomide, according to International Working Group (IWG) 2006 criteria.²⁶  Secondary objectives were to assess response duration, progression to AML, overall survival, and safety of lenalidomide.

Patients meeting all inclusion criteria received lenalidomide 10 mg once daily orally during 21 days every 4 weeks. Dose reductions were made in case of a grade of more than 2 side effects according to National Cancer Institute toxicity criteria (Common Terminology Criteria for Adverse Events, version 3.0), except for neutropenia and/or thrombocytopenia where dose reductions were made only in case of grade 4 toxicity.

In patients without response after 8 weeks and no grade more than 2 nonhematologic toxicity (or grade 4 on neutrophils and platelets), the lenalidomide dose could be increased to 15 mg/day 21 days every 28 days during an additional 8 weeks (or continued at 10 mg for an additional 8 weeks if the toxicity profile did not allow an increase to 15 mg). If no response was observed after those 8 additional weeks of treatment, the drug was to be discontinued.

In case of response, subsequent cycles were to be repeated every 28 days, with delays if ANC less than 500/mm³ or platelet count less than 25 000/mm³, until recovery above those thresholds. Study visits were scheduled every week during the first 8 weeks, then every other week and serial measurements of safety and efficacy were performed. Complete blood count was monitored at least weekly for the first 16 weeks of study treatment, then every 2 weeks for the next 36 weeks.

Bone marrow aspirate and cytogenetic studies were repeated at weeks 8, 16, 32, and 52, and when clinically indicated for assessment of disease progression. The use of granulocytic colony stimulating factor (G-CSF) was recommended in case of febrile neutropenia or grade 3 or 4 neutropenia (including subjects with grade 3 neutropenia at baseline), to avoid as much as possible delay or dose reduction in lenalidomide. Grade 3 and 4 neutropenia was defined by an ANC between 1 and 0.5 Giga per liter (G/L) and less than 0.5 G/L, respectively. Grade 3 and 4 thrombocytopenia was defined by a platelet count between 25 and 50 G/L and less than 25 G/L, respectively, according to Common Terminology Criteria for Adverse Events (version 3.0).

Physician transfusion thresholds prestudy and intrastudy were the following: for RBC transfusions hemoglobin less than 8 g/dL, or higher (9-10 g/dL) in case of severe infection, underlying cardiac or pulmonary disease, or severe symptoms of anemia; for prophylactic platelet transfusions, platelets less than 10 G/L or higher in case of fever, rapid platelet decrease, mucositis, concomitant coagulopathy.

All patients who achieved hematologic complete remission (CR), partial response (PR), marrow CR, or hematologic improvement on the erythrocytic lineage (HI-E) according to IWG 2006 criteria,²⁶  were considered responders and were to continue treatment until relapse. In agreement with IWG 2006 criteria,²⁶  complete cytogenetic response was defined by the disappearance of all chromosomal abnormalities (including del 5q and other additional abnormalities) without appearance of new ones and partial cytogenetic response by at least a 50% reduction of the number of mitoses with any chromosomal abnormality. In addition, in agreement with IWG 2006 recommendations,²⁶  response of RAEB-T patients was evaluated according to criteria that apply to MDS, but we also evaluated them according to response criteria designed for AML.²⁷ 

The study was started in November 2006 and closing date was May 1, 2008. Analysis was made on a modified intent-to-treat principle, excluding only diagnostic errors and withdrawals of consent. Prognostic factors of CR and PR achievement were analyzed by univariate and multivariate analysis (the latter using a logistic regression model). Characteristics were compared by nonparametric tests (Fisher exact test for qualitative variables). Censored endpoints were estimated by the nonparametric Kaplan-Meier method then compared between randomized groups by the log-rank test.

Because RAEB-T are now classified among AML, response, and its prognostic factors were also analyzed on WHO-defined MDS only (ie, excluding RAEB-T patients).

Between November 2006 and July 2007, 49 patients from 9 centers were included in the trial (Table 1). Two were not evaluable because of diagnosis error (n = 2, including overt AML with 60% blasts in one case and MDS with low IPSS in one case). The remaining 47 patients, who constituted the modified intent-to-treat population, included 24 males and 23 females, with a median age of 69 years (range, 36-84 years). Median time from diagnosis to treatment was 5.4 months (range, 0-117 months), 5.4 months and 6.9 months in the whole cohort, in RAEB-T/AML patients and MDS patients, respectively. A total of 25 patients had received previous treatment for MDS, including low-dose cytarabine (n = 3), intensive anthracycline-AraC chemotherapy (n = 3), arsenic trioxide (n = 3), erythropoietic stimulating agents (n = 14), and thalidomide (n = 2). At inclusion, according to WHO classification, 2 patients had pure RA, 1 had refractory cytopenia with multilineage dysplasia and ring sideroblasts (RCMD-RS), 6 RAEB-1, 20 RAEB-2, and 18 AML (RAEB-T according to French-American-British classification, with 20%-30% marrow blasts). A total of 28 (60%) patients had IPSS high and 19 (40%) had IPSS int-2. del 5q was isolated, with one additional and more than one additional abnormality in 9 (19%), 11 (23%), and 27 (58%) patients, respectively. In the last group, the median number of additional abnormalities was 5 (range, 2-16). All but 4 patients had transfusion-dependent anemia. Platelet count was less than 100 000/mm³ and ANC less than 1000/mm³ in 57% and 38% of the patients, respectively.

Four patients were excluded from the trial before completion of the first cycle by physician's decision because of cytopenias (after 5, 10, 12, and 15 days of treatment, respectively), whereas the other 43 patients received at least one complete cycle of treatment. The median number of complete cycles of lenalidomide was 2 (range, 0-16).

Thirteen of the 47 patients (27%) achieved response according to IWG 2006 criteria, including 7 (15%) CR, 2 marrow CR, and 4 erythroid hematologic improvement (HI-E; Table 2). The 2 cases of marrow CR included one patient who reached normal ANC and RBC transfusion independence, and complete cytogenetic response, and was still responding after 14.5 months but remained thrombocytopenic during the whole treatment period (patient 9 in Table 2), and a patient who remained cytopenic and RBC transfusion dependent, had no cytogenetic response, and relapsed after 3.5 months (patient 8). RBC transfusion independence was achieved in 12 patients (25%), including the 7 CR, one of the marrow CR, and the 4 HI-E.

Of the 7 patients who achieved CR, 4 obtained complete and 3 partial cytogenetic response, whereas one of the patients with marrow CR and one of the patients with HI-E also obtained complete and partial cytogenetic response, respectively (Table 2).

Twelve of the 13 responses, including 6 of the CR, were achieved after 2 cycles of lenalidomide, that is, the date of first evaluation, whereas only one later response (at 4 cycles) was observed in the 9 patients who were nonresponders after 2 cycles and continued treatment. No patient had an increase in the daily dose of lenalidomide to 15 mg, resulting from ANC and/or platelet grade 4 toxicity at lower dose.

In the 29 WHO-defined MDSs (ie, after exclusion of the 18 RAEB-T), 12 (41%) achieved response, including 6 (21%) CR, 2 (7%) mCR, and 4 (14%) HI-E. Using MDS IWG 2006 response criteria²⁶  or AML response criteria,²⁷  only 1 of the 18 patients with RAEB-T (WHO-AML) achieved response corresponding to CR using both systems.

Hematologic response duration is shown in Table 2. All responders continued lenalidomide until relapse. Median response duration was 6.5 months. Ten of the responders relapsed after 2 to 12 months, whereas 3 (including 2 CR and 1 marrow CR) were still responding after 11.5+, 14.5+, and 18+ months. Median hematologic CR duration was 11.5 months, and median duration of transfusion independence was 6.5 months (range, 2-18+ months).

With a median follow-up of 330 days, median overall survival was 272 days (Figure 1). Median survival was 169 days, 560 days, and not reached in patients who failed to respond, in patients who achieved hematologic response, and in patients who achieved hematologic CR, respectively (P < .01). Nineteen patients were still alive, including 7 of the hematologic responders.

Prognostic factors of CR were analyzed in the whole population and after exclusion of RAEB-T. In the whole cohort, CR was achieved in 6 of 9 patients (67%) with isolated del 5q, 1 of 11 (9%) with single additional abnormality, and none of the 27 patients with more than one additional abnormality (P < .001). Six of the 29 patients (21%) with less than 20% of bone marrow blasts achieved CR compared with only 1 of 18 (5%) patients with more than 20% blasts (P = .16). Finally, 7 of 20 (35%) with initial platelets more than 100 000/mm³ obtained CR, compared with none of the 27 with lower platelet counts less than 100 000/mm³ (P = .001; Table 3). Other pretreatment factors, including age, gender, time from diagnosis to treatment, hemoglobin level, and ANC, had no prognostic value for CR achievement. Using multivariate analysis, only isolated del (5q) (P = .001) and baseline platelet count more than 100 G/L (P = .009) were associated with CR achievement.

When the analysis was restricted to the 29 patients with WHO-defined MDS, CR was achieved in 6 of 9 patients (67%) with isolated del 5q, and none of the 20 patients with one or more additional abnormality (P < .001). Six of 13 (46%) with initial platelets more than 100 000/mm³ obtained CR, compared with none of the 16 with platelet counts less than 100 000/mm³ (P = .003), whereas the percentage of marrow blasts had no prognostic value for CR achievement.

Of the 6 patients who had received chemotherapy before inclusion, 2 responded (patients 1 and 7 in Table 2). Both had previously received low-dose AraC without response, and they achieved hematologic CR with lenalidomide, with complete and partial cytogenetic response, respectively.

The 9 patients with isolated del 5q all had RAEB-2 with a median bone marrow blast percentage of 15% (range, 12%-18%). Six of them achieved CR, all with cytogenetic response (3 complete and 3 partial).

Main side effects were cytopenias: 36 of the 47 patients (76%) developed grade 3 (n = 3) or grade 4 (n = 33) thrombocytopenia. Despite G-CSF administration, 23 of the 29 (79%) patients with baseline ANC more than 1000/mm³ developed grade 3 (n = 4) or grade 4 (n = 19) neutropenia, whereas 9 of the 18 (50%) with baseline ANC less than 1000/mm³ developed grade 4 neutropenia. Six of the 13 responders (46%) and 23 of the 30 nonresponders (77%) developed grade 4 thrombocytopenia, whereas 7 of the 13 responders (54%) and 17 of the nonresponders (57%) developed grade 4 neutropenia. The remaining 4 patients who received less than one cycle of treatment all experienced grade 4 thrombocytopenia and neutropenia. Serious adverse events secondary to cytopenias were sepsis (10 cases, including 1 pneumonia, 5 fever of unknown origin, 3 cellulitis, and 1 dental abscess, nonfatal central nervous system bleeding 1 case). Other severe adverse effects were thrombosis (2 cases, including 1 of ischemic colitis and 1 pulmonary embolism) and cardiac failure (2 cases). Hospitalization was required during treatment in 30 patients.

Transient treatment discontinuation followed by dose reduction to 5 mg/day was required in 34 patients, resulting from grade 4 thrombocytopenia (n = 33), grade 4 neutropenia despite G-CSF administration (n = 28), and/or severe rash (n = 1).

Fifteen patients received only one treatment cycle. Reasons for treatment discontinuation after only one cycle were death (n = 3), disease progression (n = 3), or investigator's decision resulting from severe pancytopenia (n = 6) or stable disease only (n = 3).

Twenty-eight patients had died: 19 from overt disease progression, 3 from hemorrhage, 4 from sepsis, 1 from suicide, and 1 from cardiac failure.

This is the first clinical trial assessing the therapeutic efficacy of lenalidomide in higher-risk MDSs with del 5q. The observed overall response rate was 27%, and 41% if the analysis was restricted to WHO-defined MDS patients (ie, excluding RAEB-T). This is clearly lower than in the MDS 003 study, in which 76% of the patients (112 of 148) responded.⁸  In addition, median duration of overall response was 6.5 months in our study (11.5 months in patients who achieved CR), whereas the median duration of transfusion independence was 26 months in the MDS 003 trial.⁹  However, the MDS 003 trial evaluated therapeutic efficacy of lenalidomide mainly in lower-risk MDS patients. Only 5% (n = 8) of patients in that trial had higher-risk MDS, and 3 of them became transfusion independent.

In 9 of the 11 responders in our study where cytogenetics were evaluated at response, complete or partial cytogenetic response was achieved. The MDS 003 trial also reported a close correlation between cytogenetic and hematologic improvement, as 61 of 62 patients with complete or partial cytogenetic response became RBC transfusion independent.^8,9  We also observed 7 (15%) hematologic CR (21% if the analysis was restricted to WHO-defined MDS), with a median duration of 11.5 months. In the MDS 003 study, 36% of the patients achieved histologic response, and 76% of the patients with a baseline excess of blasts normalized their marrow blast count.^8,9 

Grade 3 or 4 neutropenia and thrombocytopenia were seen in most patients, but we did not find, as in the MDS 003 trial, a correlation between the importance of treatment-induced cytopenias and subsequent response.^6,8,9  This was possibly the result of the high incidence of neutropenia and thrombocytopenia already present at baseline in the present study.

In our study, absence of cytogenetic abnormalities in addition to del 5q and baseline platelets more than 100 000/mm³ were significant predictors of CR by univariate and multivariate analysis, both in the whole patient population and when the analysis was restricted to WHO-defined MDS. The MDS 003 study had also identified pretreatment platelet counts more than 100 000/mm³ as a favorable prognostic factor for hematologic and cytogenetic response. Furthermore, in the MDS 003 study, although the difference was not statistically significant, the cytogenetic response in patients with isolated del 5q, one additional cytogenetic abnormality and complex karyotype was 77%, 67%, and 50% (3 of 6), respectively, whereas patients with isolated del 5q had more prolonged responses.^8,9 

In the present study, 6 of the 9 patients with isolated del 5q reached hematologic CR associated with complete or partial cytogenetic response, with a median duration of 11.5 months. Previous studies had shown a certain selectivity of action of lenalidomide on myeloid clones with del 5q in MDS with no or limited excess of marrow blasts.^8-11  Our results support a similar mechanism of action of lenalidomide in del 5q clones associated with a clear excess of marrow blasts in MDS. On the other hand, only 1 hematologic CR, 2 marrow CR, and 4 HI-E with RBC transfusion independence, generally of very short duration, were obtained in the 38 patients with additional chromosomal abnormalities, and evolution in those patients was generally rapid, with a median survival of only 5.5 months. It is unknown in those cases whether failure could be attributed to the presence of additional chromosomal abnormalities or to the mean higher percentage of marrow blasts (18 of those 38 patients had > 20% marrow blasts) compared with patients with isolated del 5q.

Another explanation of the limited efficacy of lenalidomide in patients with additional chromosomal abnormalities and/or marrow blasts more than 20% may have been drug dosing, which was relatively low in our study. There are ongoing experiences where lenalidomide is used at higher dose (up to 50 mg daily) and for prolonged periods in higher-risk MDS and AML with del 5q, with the aim of targeting the del 5q clone, and knowing that this regimen will be associated with prolonged cytopenias requiring lengthy hospital stays, but results of those trials are not yet available (Alan List, Tampa, FL and Gautam Borthakur, M. D. Anderson, Houston, TX, oral communication, June 2008). On the other hand, in the few reported cases of higher-risk MDSs and AML with del 5q successfully treated with lenalidomide reported so far outside of the MDS 003 trial, the drug was generally given at 10 mg/day, often reduced in case of major cytopenias, as in our study.^18,20-22 

Finally, a possible role of lenalidomide in higher-risk MDSs with del 5q has to be discussed in the light of the results observed with other treatments in this subset of MDSs. Higher-risk MDS with del 5q has poor prognosis, especially in case of complex karyotype: in a large German Austrian study, median survival of MDS with complex karyotype was 11 months in the absence of del 5q and 7 months in its presence.²⁸  CR rates observed with intensive chemotherapy in higher-risk MDS or AML after MDS with del 5q are lower than in other subtypes of AML/MDS, and median CR durations shorter.¹²  Hypomethylating agents appear to bring significant benefit in higher-risk MDS, especially after the results of a phase 3 trial showing a significant survival advantage with azacytidine over conventional treatments.¹³  Their results appear especially promising in patients with −7/del7q. On the other hand, results of hypomethylating drugs, at least as single agent, may not be as favorable in higher-risk MDSs with del 5q. Indeed, in higher-risk MDSs with del 5q treated with decitabine, Lubbert et al observed response in 4 of 5 patients¹⁴  but Kantarjian et al in only 2 of 16 patients.¹⁵  Finally, in our experience in MDSs with complex karyotype (n = 46) treated with azacytidiine, the overall response rate was 9% in the presence of del 5q, compared with 27% in its absence.¹⁶ 

In conclusion, the relatively high CR rate we observed with lenalidomide in MDSs with an excess of blasts but isolated del 5q further suggests that this drug may specifically target myeloid clones with del 5q and play a therapeutic role in such patients. By contrast, lenalidomide alone had limited efficacy in higher-risk MDS with chromosomal abnormalities in addition to del 5q, a situation where other treatments, including hypomethylating agents (at least when given alone), also may have limited efficacy. We are currently testing lenalidomide in combination with other drugs (chemotherapy or hypomethylating agents) in this patient population with very poor prognosis.

Contribution: L.A., S.B., and P.F. analyzed data and wrote the manuscript; L.A., P.F., and F.D. designed study; T.P., O.B.-R., L.L., C.R., F.D., A.S., M.P.C., J.D., G.L., N.V., S.B., and C.G. provided clinical care to patients, assisted in the analysis of data, and coauthored the paper; and R.-M.M. and N.H. collected data.

Conflict-of-interest disclosure: The Groupe Francophone des Myelodysplasies was the sponsor of this trial, for which Celgene provided the drug and a scientific grant.

A complete list of members of the Groupe Francophone des Myelodysplasies appears in Appendix S1 (available on the Blood website; see the Supplemental Materials link at the top of the online article).

Correspondence: Pierre Fenaux, Assistance Publique-Hôpitaux de Paris, Hôpital Avicenne, Service d'hématologie Clinique, Paris 13 University, 125 rue de Stalingrad, 93009 Bobigny, France, e-mail: pierre.fenaux@avc.aphp.fr.