|Home | About | Journals | Submit | Contact Us | Français|
Though lenalidomide maintenance therapy has demonstrated improved outcomes after autologous hematopoietic stem cell transplantation (auto-HCT) in patients with multiple myeloma (MM), the impact of the duration of this therapy is not clearly known.
We retrospectively analyzed all MM patients who were placed on maintenance lenalidomide after auto-HCT between January 2007 and December 2013. Progression-free survival (PFS) and overall survival (OS) were analyzed in multivariable Cox proportional hazards regression models where duration of maintenance was included as time-dependent covariate.
Of the 464 patients identified, 46% initiated therapy early (<4 months after auto-HCT). The median PFS and OS were 38 months and 78 months, respectively. Improvements in PFS (HR 0.13; 95% CI 0.04-0.38; p<0.0001) and OS (HR 0.09; 95% CI 0.03-0.26, p<0.0001) were seen for those on maintenance for >2 years versus ≤2 years. For those on maintenance for >3 years versus ≤3 years, this trend continued with improvements seen in PFS (HR 0.02; 95% CI 0.00-0.44, p=0.012) and OS (HR 0.05; 95% CI 0.00-0.83; p=0.037). The incidence of SPMs in the entire cohort was 3%. No differences were seen in survival between early and late initiators of maintenance lenalidomide.
Longer duration of maintenance therapy was associated with longer survival. The incidence of SPMs was low and not associated with the duration of maintenance. The timing of initiation of maintenance had no effect on survival.
Nearly all patients with multiple myeloma (MM) who undergo autologous hematopoietic stem cell transplantation (auto-HCT) ultimately experience disease relapse. It has been well established that post-transplant maintenance therapy can delay the relapse and potentially prolong survival1-3. Among various drugs used in this setting, immunomodulatory agents, particularly lenalidomide, has been most extensively studied in recent large prospective randomized trials4-6. These trials showed that the patients who received maintenance therapy with lenalidomide after auto-HCT had significantly longer progression-free survival (PFS)4-6 and overall survival (OS)5 compared to the patients who did not receive maintenance therapy. Therefore, maintenance with an immunomodulatory agent after auto-HCT is commonly recommended in the absence of contraindications7.
Despite promising survival data many clinicians remain apprehensive about the long-term use of lenalidomide maintenance due to unclear association between duration of maintenance and survival, increased risk of toxicity, and a higher incidence of second primary malignancies (SPM)4-6. Currently there is no uniform consensus about the optimal duration of the maintenance therapy and it remains to be seen if increased incidence of SPMs is associated with duration of maintenance therapy. A meta-analysis of individual patient data suggested that the prolonged duration of lenalidomide slightly increased the risk of developing second primary hematological malignancies8. Furthermore, in the study conducted by the Intergroupe Francophone du Myélome (IFM), lenalidomide was stopped after a median of 2 years due to concerns regarding SPMs4. Nevertheless, it remains to be seen whether the longer duration of lenalidomide maintenance improves survival or is associated with unacceptably high incidence of SPMs.
In this analysis we sought to assess the impact of duration of maintenance lenalidomide therapy after auto-HCT on PFS and OS. Additionally, we determined the incidence of SPMs and association between timing of maintenance therapy initiation (early vs. late) and survival in a large cohort of patients treated at a single center.
We retrospectively identified all patients with MM who received high-dose chemotherapy and auto-HCT at our center between January 2007 and December 2013 and were initiated on maintenance lenalidomide after transplant. Patients who received maintenance other than lenalidomide were excluded from this analysis. All patients met the eligibility criteria for transplant, which included having adequate organ function and free of any active infection. This analysis was approved by the Institutional Review Board at MD Anderson Cancer Center.
The conditioning regimen for the majority of patients (85%) was single agent melphalan. Other conditioning regimens included: busulfan/melphalan, gemcitabine/busulfan/melphalan, and melphalan/lenalidomide, which were used in 15% of patients. Lenalidomide maintenance was administered at varying doses ranging from 5 mg/day to 15 mg/day based on the treating physician's decision.
Response, disease progression and relapse were defined according to the International Myeloma Working Group (IMWG) response critera 9, 10. Complete remission (CR), included those classified as achieving near (n)CR, and stringent (s)CR.
Descriptive statistics were produced for patient and clinical characteristics for all patients. Pearson's chi-square test was used to assess associations between CR and high-risk patients as well as between patients with any improvement and incidence of death and disease progression.
OS was computed from date of auto-HCT to last known follow-up. Patients alive at the last follow-up date were censored. PFS was computed from date of auto-HCT to date of disease progression or death (if died without disease progression) or the last evaluation date. Patients who were alive and did not experience progression of disease at the last follow-up date were censored. Multivariable Cox proportional hazards regression models for OS and PFS were used to assess the associations between CR and duration of maintenance therapy, adjusting for patient's cytogenetic risk, creatinine (< 2, ≥ 2 mg/dL), hemoglobin (< 10, ≥ 10 mg/dL), B2-microglobulin (≤ 3.5, > 3.5 mg/L), International Staging System (ISS) stage at diagnosis, disease status at auto-HCT (newly diagnosed versus relapsed/refractory myeloma), and response prior to auto-HCT, where possible. Since CR and duration of maintenance therapy occurred after auto-HCT, these measures were included in the multivariable model as time-dependent covariates. In addition, patients who discontinued maintenance treatment due to disease progression were excluded from the association between PFS and duration of maintenance therapy assessment.
We identified 464 patients who underwent auto-HCT and fit the above-mentioned criteria (Table 1). Fifty-six percent of the patients were male with a median age of 60.1 (range: 30.9 - 80.2) years. IgG was the most common type of myeloma (59%) followed by IgA (20%). Eighty-three patients (18%) had high-risk cytogenetic features, defined as having one or more of the following chromosomal abnormalities detected by conventional cytogenetics or fluorescence in situ hybridization (FISH) at any time before auto-HCT: t(14,16), t (4,14), deletion of 17p, deletion of 1p, and 1q gain. Deletion of 13q was included if detected on cytogenetics only. Forty-two percent of patients had ISS stage I at the time of diagnosis with a median B2-microglobulin of 3.5 mg/L. Seventy-eight patients (17%) had history of relapsed and/or refractory disease at the time of auto-HCT. The median time from diagnosis to auto-HCT was 7.2 (range: 1.7 – 170.2) months. The overall response rate at auto-HCT [≥ partial response (PR)] was 90% (n=419) with 8% (n=39) in CR. The median duration between auto-HCT and maintenance therapy initiation was 4.2 (range: 0.9 – 61.7) months for all patients. Maintenance therapy was initiated early (< 4 months after auto-HCT) in 213 (46%) patients (median duration: 3.2 months) and late (≥ 4 months after auto-HCT) in 251 (54%) of the patients (median duration: 6.2 months). The median follow-up time for all patients was 26.6 (range: 4.3 – 78.7) months.
Twenty-two patients had missing response information after maintenance therapy. Overall response rate after auto-HCT for the 442 patients was 97%. An improvement in response was seen in 36% (n=158) patients during maintenance therapy, increasing the overall response rate to (98%). Nineteen percent (n=86) of patients who were not in CR at the time of starting maintenance achieved CR during maintenance therapy. This increased the overall CR rate from 40% (n=175) after auto-HCT to 59% (n=261) with maintenance therapy. The median time to achieve CR from the start of maintenance was 8.6 (range: 0.9-46.0) months. The CR rate after auto-HCT was higher in patients who were transplanted for newly diagnosed multiple myeloma versus those with history of relapse and/or refractory disease: 42% (n=156) versus 29% (n=26), respectively (p=0.025). However, the CR rate after auto-HCT was similar in patients with or without high-risk chromosome abnormalities: 42% (n=35) versus 39% (n=145), respectively (p=0.62). Similar trend was seen during maintenance therapy, albeit with higher CR rates: newly diagnosed multiple myeloma 60% (n=217) versus relapse and/or refractory disease 45% (n=36), p=0.014 and patients with high-risk cytogenetics 53% (n=39) versus patients without high-risk cytogenetics 58% (n=209), p=0.38.
The median PFS and OS in the entire cohort were 38 months and 78 months, respectively. Median PFS and OS for patients with high-risk cytogenetics were 24.6 months and 67.7 months, respectively, compared with median PFS of 39.6 months and median OS not reached for patients without high-risk cytogenetics (p<0.001). The time of initiation of maintenance therapy did not impact PFS (early=31.9 months versus late=31.5 months; p=0.82) and OS (early=64.0 months versus late=not reached since less than half of the patients died in this group; p=0.45). In order to exclude the patients who initiated maintenance very late after auto-HCT and thus had a guaranteed survival time by virtue of being in remission for many months prior to starting maintenance, we specified the maximum time frame of starting maintenance after auto-HCT to be 18 months. We then compared the early (who started maintenance within 4 months after auto-HCT; N=213) versus late (those who started maintenance between 4-18 months after auto-HCT; N=237) group. No difference in PFS (early=31.9 months versus late=31.5 months; p=0.88) and OS (early=64.0 months versus late=49.6 months; p=0.69) was seen. We further conducted a landmark analysis at 18 months including patients who received maintenance therapy within 18 months after transplantation, who were still alive and had not progressed at 18 months after auto-HCT. There was no difference in PFS (early=48.5 months versus late=41.6 months; p=0.23) or OS (early=not reachable versus late=not reachable; p=0.21) between the two groups in this patient cohort. No significant difference in PFS and OS was observed in patients who were in CR after auto-HCT compared with those who were not in CR after auto-HCT (PFS: HR=1.17; p=0.41; OS: HR=1.28; p=0.43). Similarly, no significant difference in PFS and OS was noted in patients who achieved CR with maintenance therapy compared with those who did not achieve CR with maintenance therapy (PFS: HR=0.85; p=0.63; OS: HR=0.38; p=0.17).
Effect of duration of maintenance therapy was assessed in multivariable analysis (Table 2). Patients who received maintenance for >2 years experienced a significantly reduced risk in progression and/or death as compared to the patients who received maintenance for ≤2 years: HR for PFS 0.13, p<0.001 and HR of OS 0.09, p<0.001. Similarly, patients who received maintenance for >3 years experienced a significantly lower risk in progression and/or death as compared to the patients who received maintenance for ≤3 years: HR for PFS 0.02, p=0.012 and HR of OS 0.05, p=0.037. Additionally, multivariable analysis was performed on patients with newly diagnosed multiple myeloma only, yielding similar findings to those of the overall patient sample studied (Table 2b)
The most common reason for discontinuation of maintenance was disease progression, observed in n=143 (31%) patients. In addition, 91 (20%) patients experienced treatment related toxicities, which caused them to stop maintenance therapy. Of these, the most common was cytopenias, experienced by 29 (32%) patients. Other common reasons for discontinuation of therapy included the development of dermatological toxicities in 13 (14%) patients, gastrointestinal toxicities in 5 (5%) and fatigue in 8 (9%) patients. Of note, the median duration from time of discontinuation to either disease progression, death or loss of follow-up for the 91 patients was 10.0 months with 64% of them remaining in our study 6 months after maintenance therapy discontinuation.
Only 12 cases of SPMs (3%) were noted in this study. These included myelodysplastic syndrome (4 patients), melanoma (2 patients), sarcoma (2 patients), acute myeloid leukemia (1 patient), squamous cell carcinoma of the skin (1 patient), hodgkin's lymphoma (1 patient), and lung cancer (1 patient). The median time to SPM development from auto-HCT was 2.2 years (range 0.6 - 5.5 years). There was no association between the duration of maintenance therapy and development of SPMs. The cumulative incidence of SPMs is presented in Figure 1.
In this analysis we have pursued the key question of the optimal duration of maintenance therapy with lenalidomide after auto-HCT that would provide the maximum survival benefit without excessively high incidence of SPMs. The median PFS and OS in our study were 38 months and 78 months, respectively, figures largely similar to those seen in other studies involving maintenance lenalidomide therapy4-6. However, our study shows that a longer duration of maintenance lenalidomide may be associated with improved PFS and OS. Furthermore, in the patients who remained on therapy for more than 3 years, a 95% reduction in the risk for disease progression or death was seen compared with patients who remained on therapy for less than 3 years. While this analysis does not answer how long to continue therapy beyond 3 years, the overall data is suggestive to continue maintenance until disease progression. This conclusion is in line with the other reports suggesting superior survival outcomes with continuous lenalidomide maintenance in the transplant11 and non-transplant setting12.
An additional 19% of patients achieved CR with maintenance therapy, similar to what is previously reported4. However, the PFS and OS did not differ in patients who were in CR at the start of maintenance or achieved CR during maintenance compared to the patients not in CR at these time points. This is in concordance with the results reported by Attal et al. where the hazard ratio for progression or death was similar in patients with ≥ VGPR or those with <VGPR at the time of randomization to lenalidomide or placebo (p=0.44)4. Similarly, the time to progression and OS were not significantly different in patients who were or were not in CR at randomization; p=0.38 and p=0.64, respectively in the study by McCarthy et al. 5. While it is well established that the patients achieving CR after auto-HCT have better survival compared to the patients with lesser responses13, perhaps, the patients who fail to achieve CR after auto-HCT derive more benefit from post-transplant maintenance therapy thus yielding equivalent survival. Furthermore, in the recent years the detection of minimal residual disease (MRD) has emerged as a strong prognostic factor14. We intended to further characterize the depth of response by detecting MRD using multiparameter flow cytometry, however, we did not have sufficient data to draw meaningful conclusions (results not presented).
In our cohort, 20% of patients discontinued the lenalidomide due to adverse events. This is similar to that reported in the large trials, where the discontinuation rate due to adverse events ranged from 5-27%4-6. SPMs were diagnosed in 3% of patients; myelodysplastic syndrome being the most common. Although the pattern of SPMs is similar to that reported in the literature, the overall incidence was somewhat lower than expected. For instance, in a meta-analysis involving 3,254 patients the cumulative incidence of SPMs at 5-years was 6.9%8. Similarly, the incidence of MDS type cytogenetic abnormalities reported in the total therapy 2 and 3 studies was 11% with clinical MDS/acute leukemia developing in 3% of patients 15. In another report from the Arkansas group, the overall incidence of SPMs in the total therapy 2 study arm that used thalidomide maintenance was 9% 16. The relatively lower incidence of SPMs seen in our study is likely attributable to a relatively shorter follow-up and the actual incidence may increase over time. We also note that there was no association between the duration of maintenance therapy and incidence of SPMs in our study. In the meta-analysis by Palumbo et al. the authors discussed that the prolonged duration of lenalidomide-treatment slightly increased the risk of hematological SPMs: 1% in patients who received lenalidomide for less than 24 months and 2% in patients who received it for longer than 24 months8. However, no statistical comparison was noted and an incidence of 2% in 502 patients with more then 2 years of maintenance does not appear excessive. Nevertheless, a longer follow-up with a greater number of patients is likely needed to adequately address this question. Moreover, although the risk of SPMs may in fact be increased with post-transplant lenalidomide maintenance compared to no maintenance, the risk of death from SPM is still quite minute when compared to that of death from MM relapse, therefore risks must be weighed against benefits when considering maintenance treatment17, 18.
We acknowledge several limitations of our study including retrospective nature of analysis, missing data, and lack of standardization for screening of SPMs, which could result in under reporting. In addition, we did not study the impact of induction regimens on post-transplant clinical outcomes. However, we have conducted an exhaustive analysis of impact of duration of lenalidomide maintenance on survival and incidence of SPMs and conclude that conceivably the most practical approach is to continue maintenance lenalidomide until unacceptable toxicity or disease progression after auto-HCT.
Funding source: This research is supported in part by the National Institutes of Health through MD Anderson's Cancer Center Support Grant (CA016672).
Significant conclusions of the manuscript: Maintenance lenalidomide after autologous hematopoietic stem cell transplantation for multiple myeloma should be continued until unacceptable toxicity or disease progression.