Durable Remission With Salvage Second Autotransplants in Patients with Multiple Myeloma

Abstract

Background

High dose chemotherapy with autologous hematopoietic cell transplant (auto-HCT) has been shown to improve survival in patients with newly diagnosed multiple myeloma. However, the role of salvage auto-HCT for relapsed patients, particularly in the era of novel therapeutics, is not well-defined.

Methods

We performed a retrospective analysis of all 44 myeloma patients (24 males, 20 females) who received a second auto-HCT as salvage between 1/3/1992 and 11/4/2008 at M.D. Anderson Cancer Center.

Results

Median interval between the first and salvage auto HCT was 30 months (range 2–78). Median age at salvage HCT was 54 years (38–73) and median number of salvage treatment regimens was 2 (range 0–5). Eleven (25%) patients had high-risk chromosomal abnormalities on conventional cytogenetic studies between diagnosis and salvage auto. Ten patients (23%) experienced grade 3 or higher non-hematologic toxicity after the salvage auto-HCT. One patient died within 100 days for a treatment-related mortality of 2%. Best responses after salvage chemotherapy + salvage auto-HCT were as follows: CR+ near CR 11%, PR 79%, with an overall response rate of 90%. Eighteen (41%) patients received post auto HCT maintenance therapy. Median follow-up from salvage HCT was 41 months. Kaplan-Meier estimates of median progression-free survival and overall survival (OS) from time of salvage auto-HCT were 12.3 and 31.7 months, respectively. Median OS from the time of diagnosis was 75 months. In a fitted Bayesian multivariate model, shorter time to progression (TTP) after first auto HCT, greater number of prior therapies, African-American race, and IgG subtype were significantly associated with worse OS.

Conclusions

In selected myeloma patients, a second auto-HCT for salvage therapy is well tolerated with acceptable toxicity. The ORR and PFS are comparable to other salvage regimens.

Background

Multiple myeloma is the second most common hematologic malignancy in adults, with over 20,000 new cases and 10,000 deaths annually in the United States.¹ Several studies have supported the use of autologous hematopoietic cell transplant (auto-HCT) in the front-line setting for patients under 65 with good renal function.^2–4 However, this procedure is not curative and the majority of patients inevitably relapse.

The management of patients who relapse after initial treatment with a single auto-HCT is still being debated. Novel agents such as thalidomide, lenalidomide and bortezomib are active in the salvage setting; however prolonged treatment can result in significant toxicities and many current patients have already been treated with these agents in the induction phase. Numerous promising agents are in development, including newer generation immunomodulatory drugs and proteosome inhibitors, histone deacetylase inhibitors, and plasma cell-specific antibodies. However, the use of these is often limited to those patients who can participate in a clinical trial. Despite the use of approved novel agents, the PFS in a majority of patients with persistent or refractory disease is only 6–14 months.^{5, 6} These numbers highlight the need for more effective therapy, especially in patients with a good performance status. Several studies suggest that salvage auto-HCT is reasonably safe for selected patients and may grant additional progression-free survival (PFS).^7–10 In this study, we performed a retrospective review of all 44 patients who have undergone a second, salvage auto-HCT for multiple myeloma at M.D. Anderson Cancer Center (MDACC) through 2008.

Materials and Methods

Patients

Forty-four patients received a salvage auto-HCT between January 1992 and November 2008. The information from these transplants was prospectively collected in our database and used for the current analysis. Eligible patients had a diagnosis of myeloma by International Myeloma Working Group Criteria (IMWG) criteria and had evidence of relapse (by IMWG criteria) after having undergone an auto-HCT. Patients who underwent a second transplant as part of a planned tandem regimen were not included in this study. In general, patients were eligible to receive the second transplant if they had an ECOG performance status of < 2 and had adequate renal (Cr≤2.0), cardiac (left ventricular ejection fraction > 45%), pulmonary (diffusing capacity of the lung for carbon monoxide > 50%), and hepatic (bilirubin, transaminases < 2 times upper limit of normal) function.

Hematopoietic Stem Cell Mobilization and Collection

Bone marrow or G-CSF-primed peripheral blood progenitor cells were collected using standard mobilization protocols and apheresis techniques.¹¹ Forty-three patients received a peripheral blood stem cell autograft while 1 patient received a bone marrow-derived graft. Thirty-three patients received stem cells that had been collected prior to their first auto-HCT while 11 patients underwent a second collection before their salvage auto-HCT. All patients signed written informed consent according to our institutional and the National Marrow Donor Program guidelines. The study was reviewed and approved by the Institutional Review Board at MDACC.

Preparative Regimens and Supportive Care

Preparative regimen for the salvage transplant was high-dose melphalan alone (200 mg/m², n=21) or a melphalan-based combination (21). Two patients received a combination of thiotepa (750 mg/m²), busulfan (varying doses) and cyclophosphamide (120 mg/m²). Melphalan-based combinations used were as follows: 1) melphalan (140 mg/m²) with busulfan (520 mg/m²) (n=2), 2) melphalan (140 mg/m²) with topotecan (17.5 mg/m2) and cyclophosphamide (3 g/m²) (n=6) 3) melphalan (200 mg/m²) with arsenic trioxide (varying doses) (n=9), 4) melphalan (200 mg/m²) with arsenic trioxide (varying doses) and bortezomib (1 mg/m² × 3 doses) (n=2), 5) melphalan (120 mg/m²) with busulfan (420 mg/m²) and gemcitabine (150 mg/m²) (n=1) or 6) melphalan (140 mg/m²) with TBI (n=1). Patients received infection prophylaxis with levofloxacin or ciprofloxacin, fluconazole and acyclovir or valacyclovir. Filgrastim 5μg/kg was administered subcutaneously daily from day+1 after auto-HCT until the recovery of absolute neutrophil count (ANC) to > 0.5 × 10⁹/L for 3 days. Blood products were irradiated and filtered to remove leukocytes prior to transfusion. After recovery of neutrophil count, patients received infection prophylaxis with sulfamethoxazole-trimethoprim or pentamidine and acyclovir or valacyclovir.

Engraftment and Response

Responses were graded according to IMWG criteria.¹² Initial responses were assessed at approximately Day 100 and final, best responses were determined at last follow-up. Neutrophil engraftment was defined as the first of 3 consecutive days with an ANC ≥ 0.5 × 10⁹/L. Failure to engraft by day 30 was considered primary graft failure. Platelet engraftment was defined as the first of 7 consecutive days with a platelet of ≥ 20 × 10⁹/L without transfusion support. Adverse effects were graded according to current National Cancer Institute Common Toxicity Criteria.

Statistical Analysis

Primary endpoints were Kaplan-Meier’s estimates of OS and PFS. Secondary endpoints were response rate, treatment-related mortality (TRM) and toxicity rates. OS was measured from the day of autologous stem cell infusion (day 0) to death from any cause, with censoring performed at date of last contact. PFS was determined from the day of stem cell infusion to the day of documented relapse or progression. Death from any cause other than relapse before day 100 was classified as TRM. Patient characteristics were summarized using the median (range) for numerical variables or frequencies (percentages) for categorical variables. Differences in the distributions of patient characteristics between groups were assessed using Wilcoxon, Kruskal-Wallis or generalized Fisher exact tests^{13, 14}. Unadjusted probabilities of event times were estimated using the method of Kaplan and Meier ¹⁵. The log-rank test¹⁶ was used to compare unadjusted OS and PFS between subgroups. Bayesian exponential regression models were used to assess the joint effects of patient covariates on overall survival, with the best fitting model chosen from the exponential, gamma, inverse gamma, Weibull, and log normal using the Bayesian Information Criterion (BIC),¹⁷. In each multivariate regression model, the covariates included age, gender, race, log(CD34+ cell dose), time to progression after first therapy sequence, number of prior therapies before salvage auto-HCT, ISS stage, immunoglobulin subtype, and date of transplant (before or after January 1, 2003). The interaction terms [race=African-American]*[histology=IgG] and [race=African-American]*log(time from first to second SCT) were also included into the model first, but were dropped from the model since they were not significant. The exponential model for OS assumes that the mean OS takes the form μ = exp(m) where m is a linear combination of covariate effects and interactions. We assumed that each parameter in the linear term followed a non-informative normal prior with mean 0 and variance 1000, and a non-informative gamma prior for r with mean 1 and variance 1000. All statistical analyses were carried out in Splus 6.1¹⁸ and, for the Bayesian model fits, in WinBugs 1.4 (Cambridge, U.K.).

Results

Patient characteristics

Patient characteristics are shown in Table 1. Of the 44 patients, 24 were male and 20 were female. The median age at salvage transplant was 55 years (range 38–73) and the median time between the first auto-HCT and the salvage auto-HCT was 30 months (range 2–78). Median number of therapies received before salvage auto-HCT was 2 (range 0–5). Eleven patients (25%) had high-risk chromosomal abnormalities on conventional cytogenetic studies at any time between diagnosis and salvage auto HCT ¹⁹. Patients received a mean CD34+ cell dose of 4.1 ×10⁶/kg (SD 1.8 ×10⁶/kg). Thirty-seven (84%) patients received thalidomide, lenalidomide or bortezomib in the induction setting while 43 (98%) patients received those agents in the relapsed setting, before the salvage transplant. Fifteen patients (34%) received thalidomide, lenalidomide or bortezomib maintenance therapy after the 2^nd transplant. Eighteen (41%) patients received any type of maintenance therapy.

Table 1.

Patient Characteristics.

Variable		Number of Patients (%)
All patients		N=44 (100%)
Gender
	Female	20 (45.5%)
	Male	24 (54.5%)
Race
	African-American	8 (18.2%)
	Other	8 (18.2%)
	Caucasian	28 (63.6%)
Age at salvage HCT (years)	Median (range)	54.5 (37.5–72.9)
Time between first and salvage HCT (months)	Median (range)	30 (2–78)
No. of therapies before salvage HCT (after 1st HCT)	Median (range)	2 (0–6)
High-risk Cytogenetics		11 (25.0%)
Durie Salmon stage
	I	3 (7%)
	II	13 (26%)
	IIi	26 (59%)
International staging system stage
	I	10 (22.7%)
	II	10 (22.7%)
	II	8 (18.2%)
CD34	Mean (SD)	4.6 (1.8)
Histology (Myeloma subtypes)
	IgG	28 (63%)
	IgA	8 (18%)
	κ or λ Light Chain	6 (11.4%)
	Non-secretory	1 (2.3%)
Maintenance With Novel Agent After Salvage HCT		15 (34%)

Engraftment and Toxicity

Median time to neutrophil and platelet engraftment after the salvage auto-HCT was 10 days (range 8–14) and 11 days (range 6–398), respectively. Of 44 patients, 1 died within 100 days of the salvage transplant, due to supraventricular tachycardia and possible pneumonia, for a TRM of 2%. Ten (23%) patients experienced a grade 3 or higher non-hematologic toxicity, with symptom distributions detailed in Table 2.

Table 2.

Grade 3 or higher non-hematologic toxicities during salvage auto-HCT

Organ system	Number of Patients (%)	Details
All organs	10 (23)
Cardiovascular	5 (11)	2 tachycardia, 1 hypotension, 1 bradycardia, 1 other
Constitutional	0 (0)
Gastrointestinal	3 (7)	2 diarrhea, 1 mucositis
Infectious	5 (11)	5 neutropenic infections
Pulmonary	2 (5)	1 effusion, 1 pneumonia
Skin	1 (2)	1 rash
Other	0 (0)

Response and Survival

Five patients (11%) eventually achieved CR or VGPR and the best overall response rate (ORR) from the time of relapse was 90%. Two patients (5%) had stable disease and 1 patient (2%) had progressive disease. Median follow-up time from salvage auto-HCT in surviving patients was 41 months. After salvage auto-HCT the median PFS was 12.3 months (range 2–98 months) and the median OS was 31.7 months (0–96 months) (Figure 1). The median OS from time of diagnosis was 74.9 months (25–165 months). In all, 8 patients (18%) were able to have an improvement in their disease status from salvage chemotherapy after the salvage transplant. At last follow up, 16 patients were alive while 8 patients were alive and had not progressed.

Figure 1.

Kaplan-Meier plots for overall survival (OS) and progression-free survival (PFS): median OS is 31.7 months (95%CI: 21 – 69 months); median PFS = 12.3 months (95%CI: 8.3 –19.9 months)

Prognostic Factors

We studied the impact of age, (log)CD34+ cell dose, time to progression (TTP) after first therapy sequence, number of prior therapies before salvage auto-HCT, gender, race, ISS stage, immunoglobulin subtype, and date of transplant (before or after January 1, 2003), on OS after salvage auto HCT. The fitted Bayesian multivariate regression model for OS showed that shorter TTP after first transplant, larger number of prior therapies, race being African-American, and IgG subtype were significantly associated with worse OS, where a “significant” effect in the fitted Bayesian model was defined as one having a posterior probability of a harmful effect either > .95 or <.05. (Table 3, Figure 2). There was no significant independent impact of any other factors on the OS, which may be a reflection of a heterogeneous population and relatively small number of events in each group. Detection of high-risk chromosomal abnormalities showed a trend towards a shorter OS (p=0.07, data not shown).

Table 3. Multivariate analysis.

Bayesian multivariate analysis of overall survival time (starting from 2nd SCT) based on the exponential distribution (best fitting model for OS). The prior for all covariate parameters is N(0,100).

	Posterior Quantities
	Beta	SE	Prob(harmful effect)
Age	−0.033	0.031	0.132
log(CD34)	−0.256	0.551	0.33
TTP after 1st	−0.357	0.214	0.038
No. Prior. Therapy	0.385	0.145	0.995
Female	0.1	0.461	0.588
Black	1.022	0.517	0.968
High risk	0.137	0.651	0.596
IgG	1.058	0.605	0.971
Pts before 2003	−0.184	0.447	0.343

	Posterior Quantities
	Beta	SE	Prob(harmful effect)
Age	−0.033	0.031	0.132
log(CD34)	−0.256	0.551	0.33
TTP after 1st auto HCT	−0.357	0.214	0.038
No. Prior Therapy	0.385	0.145	0.995
Female	0.1	0.461	0.588
Black	1.022	0.517	0.968
High risk	0.137	0.651	0.596
IgG	1.058	0.605	0.971
Patients before 2003	−0.184	0.447	0.343

Figure 2.

Kaplan-Meier plots for overall survival by Ig subtype and Race: median survival (95%CI) is 8.3 months (6.3, 100) for African-American with non-IgG, 12.2 months (8.1, 100) for African-American with IgG, 36.7 months (21.2, 100) for Non African-American with IgG, and not reached (>100 months) for Non African-American with non-IgG.

Discussion

Despite numerous advances, multiple myeloma remains an incurable disease. While novel agents and high dose chemotherapy have improved response rates, patients eventually relapse and succumb to their disease. Given the rapid development in the field in recent years, we are now faced with the challenge of systematically determining how to best use our tools to improve survival and quality of life.

The idea of using a second, salvage transplant is appropriate, given the data with tandem transplants. In addition, patients who had a durable response to a first transplant may fare best after the salvage transplant.^{7, 20} For many patients the issue of a second transplant at relapse is an important one. For these patients, the results of this study suggest two important conclusions: 1) a salvage transplant done at the time of relapse is safe, and 2) the benefit of the salvage transplant is comparable (or potentially superior) to other salvage regimens. In addition, our survival and toxicity data are comparable with published and preliminary reports from several other institutions with similar sample sizes.^{9, 21–23} Combining the results of our study with others’, it appears that the salvage transplant is a viable option for select patients.

Another pressing issue facing myeloma physicians is having a long-term treatment plan as our patients are living longer with their disease. In this setting it is crucial to offer patients therapies that strike a balance between intensity and duration, such that toxicities can be limited. While the salvage transplant does have an initial risk of significant toxicity, the potential of having over a year or longer without treatment or with low dose maintenance treatment may be appealing to many patients. In addition, other preparative regimens that build on melphalan may offer an opportunity to enhance the benefit of a salvage auto-HCT. Phase 2 studies combining melphalan with other agents in this patient population are ongoing at our institution.

In interpreting the results from this retrospective analysis, there are certainly limitations. The most obvious of these is a selection bias: more favorable patients were likely recommended for salvage auto-HCT and thus, these results may not be applicable to all patients. In addition, the sample size is small and distributed over more than 15 years. However, most of the patients did receive novel therapeutics at some point, making these results more pertinent to the contemporary patient population.

Given the relatively small sample size, it is not surprising that most of the prognostic factors studied did not affect OS significantly. The negative impact of shorter TTP and greater number of therapies is consistent with the known biology of the disease. However, the association between African-American race and shorter OS is an interesting one. While it is known that African-Americans are disproportionately affected by multiple myeloma in incidence, ²⁴ there is debate over whether there is a difference in outcome in the era of auto-SCT. Preliminary analysis at our institution of all myeloma patients (including those not undergoing auto-SCT) in the era of novel therapeutics suggests that African-Americans have a similar OS in comparison to Caucasian patients, but that fewer African-Americans undergo auto-SCT²⁵. Though this is only a small analysis, one could hypothesize that an as yet undetermined factor is linked both to accessibility and success of auto-SCT, which might explain our current and previous findings in African-American patients.

Our finding that IgG subtype was associated with a worse OS was also surprising. Though there are few studied examining the prognostic significance of immunoglobulin subtype, at least one retrospective analysis suggests that IgG subtype is not a poor prognostic indicator in myeloma patients as a whole ²⁶. The relevance of subtype in the setting of second, salvage auto-HCT is less clear and may be an artifact of small sample size.

In deciding whether a salvage transplant for relapsed myeloma is the treatment choice in comparison with other salvage regimens, the answer remains unclear. The overall response rate of the second auto-HCT in our study (90%) is comparable to that reported with lenalidomide/dexamethasone ^{27, 28} or bortezomib-containing regimens ^29–36 for relapsed patients – though the chance of obtaining a CR or VGPR is likely greater with lenalidomide or bortezomib-chemotherapy combinations. Our ORR also suggests that salvage pharmacotherapy and salvage auto-HCT are not mutually exclusive. In fact, the use of novel agents in combination with salvage auto-HCT may offer patients the best chance of response, since 98% of patients received a novel agent as part of their salvage regimen. The PFS in our study is also comparable to those achieved in the lenalidomide- and bortezomib-combination studies. However, it is hazardous to compare results between all of these studies as there is marked heterogeneity in how many patients had previously been treated with thalidomide, lenalidomide or bortezomib. The ideal study to answer the question of salvage chemotherapy versus salvage auto-HCT would randomize patients to receive either combination chemotherapy only (for example bortezomib/dexamethasone/oral cyclophosphamide) versus combination chemotherapy and a second, salvage transplant. In the absence of prospective, multi-center studies, we can turn to smaller, single-center studies. One British group study found no difference in outcomes whether relapsed patients received salvage chemotherapy or salvage auto-HCT.²⁰ However, only a minority of these patients received thalidomide or bortezomib (none received lenalidomide) as part of their salvage therapy.

In conclusion, a second, salvage auto-HCT appears to be a safe and feasible treatment option for select patients with relapsed myeloma. This intervention can yield a durable remission and survival. Thus, patients with relapsed myeloma after a single auto-HCT should be considered for a salvage auto-HCT in an effort to provide an opportunity for a therapy-free interval.