Historical Perspective of High-Dose Therapy Followed by Autologous Stem Cell Transplantation in Multiple Myeloma

High-dose therapy (HDT) followed by an autologous stem cell transplantation (ASCT) is considered part of the standard of care for eligible patients with newly diagnosed multiple myeloma (NDMM), after induction therapy with novel agents [1‒3]. Since its introduction in the 1980s, this practice has been developed and studied extensively. In this review, we wish to summarize the important milestones in the history of ASCT and highlight the issues currently under debate in the field.

HDT with melphalan was introduced in 1983 by McElwain and Powles [4], who treated 9 patients, of which four achieved complete response (CR). Neutrophil recovery was prolonged, lasting 20–56 days, longer in previously treated patients. In 1986, Selby et al. [5] reported on 58 patients under 63 years old treated with melphalan 140 mg/m² (MEL140). Treatment caused severe myelosuppression; the median time to leukocyte recovery was 28 days in 41 previously untreated patients and 42 days in 15 refractory patients. The overall response rate (ORR) was 78% for previously untreated patients, and CR was 27%. There was a high rate of early death, with 17% (10 patients) dying within 2 months. The authors conducted another study of 22 previously untreated patients who received MEL140 along with high-dose methylprednisolone. The response rate was similar, but the early death rate was reduced to 5% (only 1 patient) which was attributed by the authors to better supportive treatment.

The concept of stem cell reconstitution was introduced in the 1990s, aiming to shorten the time to hematopoiesis recovery. Several studies [6‒10] compared HDT followed by ASCT with standard-dose therapy (SDT) for patients under 65 years old, the results of which are listed in Table 1.

Fermand et al. [6] treated 63 patients with HDT followed by ASCT. Mobilization was done with a cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP)-like regimen, bridged to HDT with three courses of a vincristine, doxorubicin, and dexamethasone (VAD)-like regimen. HDT included total body irradiation (TBI) and MEL140, as well as carmustine, etoposide, and cyclophosphamide. Seven patients died of early toxicity. Hematopoiesis was recovered within 2 weeks. Twenty percent of the patients achieved CR, and the 5-year overall survival (OS) was 60%.

Next, 167 patients were randomized to either SDT or HDT with ASCT. Patients under 55 received a combination of melphalan, carmustine, etoposide, and cyclophosphamide as HDT, and SDT included vincristine, cyclophosphamide, melphalan, and prednisone (VCMP). The HDT arm included an ASCT as rescue. Patients 55–65 years old received melphalan 200 mg/m2 (MEL200) as HDT. Survival at 2 years was 82% for HDT compared to 67% for SDT; event-free survival was 77% for HDT and 47% for SDT.

Attal et al. [7] randomized 200 previously untreated patients under the age of 65 to SDT or HDT with ASCT. SDT included VCMP alternating with vincristine, carmustine, doxorubicin and prednisone for 12 months, with interferon (IFN) added in cycle nine and given until progression. HDT included 4–6 cycles of vincristine, carmustine, doxorubicin and prednisone (VCMP) followed by MEL140 with TBI, consolidating with IFN-alpha. ORR was 81% for HDT compared with 57% for SDT, and EFS at 5 years was 28% for HDT compared with 10% for SDT.

Child et al. [8] randomized 407 patients to SDT or HDT that included either MEL200 followed by peripheral blood stem cell infusion or MEL140 with TBI and a bone marrow autograft. Patients received high-dose methylprednisolone for 4 days after melphalan. Both SDT and HDT arms received IFN maintenance. OS and progression-free survival (PFS) were improved with HDT. Median OS was 54 months for HDT compared with 42 months for SDT. Median PFS was 31 months for HDT compared with 19 months for SDT. There were more infections in the HDT arm.

Two more recent studies did not, however, show a survival benefit. Fermand et al. [9] randomized 190 patients aged 55–65 years with NDMM to either SDT or HDT with ASCT. SDT included monthly VCMP. HDT included either MEL200 or MEL140 with busulfan 16 mg/kg orally. This study had a median follow-up of approximately 10 years. The study showed a longer median EFS of 25 months for HDT versus 19 months for SDT, but this difference did not reach statistical significance (p = 0.09). Median OS was 47.8 months in HDT and 47.6 months in SDT. Twenty-two percent of patients in the SDT group received salvage therapy with HDT, which could explain the comparable survival rates between the groups.

Barlogie et al. [10] randomized 516 patients to SDT versus HDT with ASCT. SDT included vincristine, carmustine, melphalan, cyclophosphamide, and prednisone. HDT included MEL140 with TBI. No difference was observed in response rates, PFS, or OS. However, in this study as well, 55% of patients who received SDT and relapsed received HDT. Patients who continued with SDT had a shorter median survival of 23 months compared with 30 months of those salvaged with HDT, but the difference was not statistically significant.

A meta-analysis by Koreth et al. [11] included nine prospective, randomized-controlled trials comparing HDT to SDT. Different HDT and SDT regimens were used, and studies had significant heterogeneity. The authors concluded that HDT had a PFS benefit, but not an OS benefit, and that the treatment-related mortality (TRM) rate was substantial (OR 3, 95% confidence interval [CI]: 1.64–5.5).

Different conditioning regimens were examined in several studies. Different doses of melphalan were compared in two studies. A retrospective study done by the European Society for Blood and Marrow Transplantation (EBMT) [12] reported 1,964 patients, of which 1,719 received MEL200 and 245 received MEL140. Notably, the groups differed in pre-transplant treatment regimens; patients in the MEL140 group were older (median age 64 vs. 59, p = 0.001); more patients in the MEL140 group had a Karnofsky score <90 (38% vs. 28%, p = 0.002), an estimated GFR <50 mL/min (37% vs. 4%, p = 0.001), and received novel agents (49% vs. 40%, p = 0.001). MEL200 had an OS benefit for patients in less than partial response, while for patients in VGPR or CR, MEL140 showed a survival benefit. The authors suggested that in patients with more chemoresistant plasma cells, a higher dose of melphalan was needed. Presumably, the patients with more chemosensitive cells benefitted from reduced toxicity.

A study prospectively followed 459 consecutive MM patients who received different melphalan doses as conditioning for ASCT [13]. Sixty-nine patients received melphalan 150 mg/m² (MEL150), and 390 received MEL200. Patients in the MEL150 group were older (29% were over 60 vs. 15.9% in the MEL200, p = 0.009) and more frequently had ISS stage III (62% vs. 31%, p = 0.001), extramedullary disease (32% vs. 20%, p = 0.001), and eGFR <40 (50% vs. 19%, p = 0.001). MEL200 had more patients with high-risk cytogenetics (HRCG), but this difference was not statistically significant (26% vs. 5%, p = 0.36; only 144 patients had data). Seventy-nine percent of patients received induction with novel agents, which did not differ significantly between the groups. Although patients in the MEL200 group had a higher ORR (93% vs. 85%, p = 0.02), PFS and OS did not differ; median PFS was 53 months versus 60 months (p = 0.746), and median OS was 100 months versus 102 months (p = 0.917) for the MEL150 and MEL200 cohorts, respectively. TRM was comparable between the groups (7.2% in MEL150 vs. 4.1% in MEL200, p = 0.194), and toxicity was similar. The authors concluded that MEL150 is a safe option for older patients and suggested that MEL200 did not show a survival advantage because of the high proportion of patients who had HRCG.

Another study examined the role of TBI, prospectively comparing 140 patients treated with MEL140 + TBI to 142 patients treated with MEL200 [14]. In this study, patients in the MEL140 + TBI had a higher rate of severe mucositis and a lower rate of survival at 45 months (45.5% vs. 65.8%, p = 0.05). A phase 3 study examined the addition of busulfan to melphalan (Bu-Mel) [15]. Patients were randomized to receive either MEL200 (n = 98) or Bu-Mel (n = 104). The Bu-Mel regimen consisted of busulfan on days −7,−6,−5, and −4 and MEL140. At a median follow-up of 22 months, patients who received Bu-Mel had a higher rate of grade II–III mucositis (74% vs. 14%) but superior PFS (64.7 months vs. 43.5 months, p = 0.022). However, the study did not demonstrate an OS benefit. In a subgroup analysis of 62 patients with HRCG [16], patients in the Bu-Mel group had a longer median PFS (44.7 vs. 25.7 months, p = 0.044). Again, there was no difference in OS.

A recent phase 1/2 study examined adding carfilzomib to MEL200 in patients undergoing ASCT for MM in patients treated with two or fewer prior lines of therapy [17]. The dose escalation cohort included 14 patients, and the phase 2 cohort included 35 patients. Only 4 patients received carfilzomib as a part of their induction. None of the patients received an anti-CD38 antibody. At 1-year post-transplant, 22% of patients achieved CR or better as their best response, and 77% achieved VGPR or better. Notably, 16% of the patients had grade 3–4 cardiac toxicity. The CR rates were lower than those in the IFM-2009 study [18] with comparable VGPR rates; the authors attribute this to the lack of maintenance therapy.

A new formulation of melphalan, propylene-free melphalan (or EVOMELA), is more stable, less likely to form precipitates in solution than melphalan, and propylene glycol was shown to be associated with various toxicities. A large meta-analysis demonstrated propylene-free melphalan to be safe and tolerable, with comparable toxicity to melphalan, with some studies demonstrating lower rates of mucositis. Response rates are also similar [19].

With the emergence of novel agents as induction regimens, the role of HDT with ASCT was evaluated in several studies [18, 20‒23], which are summarized in Table 2. The studies differed in induction regimens; two used Revlimid and Dexamethasone [21, 22], one used bortezomib with melphalan and prednisone [20], and two used a combination of Revlimid, Velcade, and Dexamethasone (RVD) [18, 23]. SDT regimens used the same novel agents as those used for induction in the study. HDT regimens all used MEL200, but two of the studies used tandem transplantation [21, 22] and one included both a single ASCT (Dutch cohort) and tandem ASCT (German cohort), per the local policy of the center. All patients received lenalidomide maintenance. HDT improved PFS in all the abovementioned studies, and in a meta-analysis [24], the combined hazard ratio (HR) was 0.55 (95% CI: 0.41–0.74; p = 0.004). Only three studies reported OS. Two studies showed an improved OS [21, 22], both of which administered Revlimid and Dexamethasone induction and a tandem ASCT. In a study by Gay et al. [21], of 389 patients enrolled, only 256 were eligible for consolidation therapy. A total of 127 patients received tandem transplantation, and 129 patients received cyclophosphamide, prednisone, and lenalidomide. At 4 years, transplanted patients had a higher rate of OS, 86% (95% CI: 79–92) versus 73% (95% CI: 65–82); HR 2.40, 95% CI: 1.32–4.38; p = 0.004. However, in the entire study population, there was no significant difference in OS between the treatment arms. Palumbo et al. [22] randomized 273 patients younger than 65 years to tandem transplantation versus melphalan, prednisone, and lenalidomide. At 4 years, OS was superior in the transplantation group, 81.6% versus 65.3%; HR for death was 0.55 (95% CI: 0.32–0.93; p = 0.02). Attal et al. [18] used RVD induction and only one transplant, and this study did not show an OS benefit. In a meta-analysis of these studies [24], the combined HR of 0.76 for OS was not statistically significant. In the phase 3 DETERMINATION study, 722 patients under 65 years received RVD induction (3 cycles) [23] and were then randomized to either five additional RVD cycles (n = 357) or MEL200 + ASCT (n = 365). Transplanted patients had a superior PFS, with a median PFS of 67.5 months versus 46.2 months (p < 0.001). The difference was more pronounced in the HRCG group (defined as 17p deletion, t(4;14) translocation, or t(14;16) translocation), with a median PFS of 55.5 months versus 17.1 months. With a median follow-up of over 6 years, there was no OS benefit for ASCT. The lack of an OS benefit is probably associated with the highly efficacious treatments available after first-line therapy.

In the phase 2 FORTE study [25], 474 patients with NDMM were randomized to either carfilzomib, lenalidomide, and dexamethasone for 12 cycles (KRd12), KRd for 4 cycles, followed by ASCT, and then consolidation with 4 cycles of KRd (KRd + ASCT), and carfilzomib, cyclophosphamide and dexamethasone (KCd), followed by ASCT, and then consolidation with 4 cycles of KCd (KCd + ASCT). There was no significant difference in the risk of progression or death between the KRd12 groups compared with KCd + ASCT (HR 0.88, 95% CI: 0.64–1.22, p = 0.45); however, KRd + ASCT did show a survival advantage over KRd12 (HR = 0.61, 95% CI: 0.43–0.88, p = 0.0084).

Sufficient collection of cells for transplantation is crucial for the success of ASCT. The International Myeloma Working Group (IMWG) recommends a threshold of 4–6 ×10⁶ CD34-positive cells for a single ASCT [26]. A minimum of 2 ×10⁶ CD34-positive cells is recommended for adequate engraftment [27, 28]. Granulocyte colony-stimulating factor (GCSF) is commonly used for mobilization; it was first used alongside cyclophosphamide and is now usually used alone [29], as it was shown that this method yields a sufficient amount of cells for successful engraftment and less toxicity. Chemotherapy mobilization can improve yields over growth factor mobilization alone [30]. Plerixafor, a selective reversible inhibitor of the CXCR4 chemokine receptor, was shown to reduce mobilization failure (3.8% for plerixafor + GCSF cohort vs. 12.8% for GCSF alone, p < 0.014) when used in a risk-adapted manner, i.e., for patients with less than 20 CD34+ cells/μl after 4 days of GCSF treatment [31]. There are concerns that novel agents might hinder the ability to collect stem cells; however, data from the GRIFFIN study, in which patients received daratumumab, lenalidomide, bortezomib, and dexamethasone (D-VRD) as induction, and the MASTER study, in which patients received daratumumab, lenalidomide, carfilzomib, and dexamethasone, showed that 2–7% of these patients had collection failure after the first attempt. Plerixafor was used either upfront or as a rescue strategy, and ultimately over 98% of patients in all groups were able to collect enough stem cells and undergo successful ASCT [32]. A real-world retrospective study demonstrated similar findings, with no collection failures after D-VRD; notably, patients received plerixafor both preemptively and as a rescue strategy [33]. In the phase 3 PERSEUS study, stem cell yield was higher in the VRD group compared to D-VRD (7.4 ×10⁶/kg vs. 5.5 ×10⁶/kg), but the two groups did not differ in the percentage of patients who proceeded to ASCT. Median time to engraftment was also similar (median 14 days in both groups) [34]. A new CXCR4 inhibitor, motixafortide, greatly improved stem cell collection rates when added to GCSF compared to placebo in a phase 3 study [35].

The choice of single versus tandem ASCT is debatable. Tandem transplantation was shown to have a clear OS benefit over a single transplantation in a randomized-controlled trial of 399 patients under the age of 60, especially for patients not achieving VGPR after the first transplant [36]. However, this study was conducted before the era of novel agents (patients were treated with VAD). Another prospective study of 321 patients showed a PFS benefit for tandem transplantation versus single transplantation, after induction which included VAD chemotherapy only, but not an OS benefit [37]. A possible explanation is that patients relapsing in this study were given salvage treatment with a second transplantation or with novel agents (thalidomide and bortezomib). Another phase 3 study of 358 patients receiving chemotherapy-based induction with a long follow-up (median 11 years) also showed no OS benefit for tandem transplantation; most patients received novel agents at first relapse in this study as well [38]. In a study designed to examine the addition of different consolidation therapies after induction, 247 patients received tandem transplantation. Compared to single ASCT + lenalidomide maintenance, tandem ASCT did not improve PFS or OS, not for the whole cohort or for the subset of high-risk patients [39].

Improved supportive care in recent decades has made ASCT a safe procedure with low rates of toxicity and mortality. However, HDT, followed by ASCT, still carries a higher risk than standard therapy for adverse events, mainly hematologic adverse events, infections, and mucositis. Moreover, ASCT reduces quality of life for approximately 3 months. Reported mortality rates following ASCT are 1% [40, 41].

In a study comparing high-dose melphalan followed by ASCT to standard-dose melphalan following induction with lenalidomide and dexamethasone, patients receiving HDT had higher rates of grade 3 or 4 neutropenia (94.3% vs. 51.5%, p < 0.001) and thrombocytopenia (93.6% vs. 8.3%, p < 0.001), as well as higher rate of gastrointestinal events (18.4% vs. 0%, p < 0.001) and infections (16.3% vs. 0.8%, p < 0.001) [22]. In a study comparing RVD alone to RVD followed by ASCT, TRM rates were 0.6% versus 1.7% [18]. The rate of secondary malignancies including acute myeloid leukemia was numerically higher in the transplantation group, but not statistically significant. In another study comparing these regimens [23], transplanted patients had a 9% rate of febrile neutropenia (vs. 4.2% in non-transplanted) and a 5.2% rate of oral mucositis (vs. none in the non-transplanted patients). TRM was 1.6% in the transplanted patients versus 0.3% in the non-transplanted patients. Of 6 patients who died, 5 died during maintenance with lenalidomide, and 1 died after transplantation. In this study, at a median follow-up of 76 months, secondary malignancy rates were comparable between groups; however, the rate of acute myeloid leukemia or myelodysplastic syndrome was significantly higher in the transplantation group (2.7% vs. none, p = 0.002).

Several studies have demonstrated that with an appropriate program, ASCT can be performed safely in an outpatient setting for selected patients with MM [40, 42, 43]. In these studies, the 100-day mortality rate was very low at 0–1%. In a study by Holbro et al. [42], 84% of patients were hospitalized in the 100 days post-transplant. Of 91 patients in the study, 71 were hospitalized for febrile neutropenia and 5 for mucositis. In this study, age >60 and the stage of the disease were found to be predictive of the need for hospitalization. In a study by Paul et al. [43], of 82 patients deemed eligible for outpatient monitoring, 67% required readmission before day 100, but these patients still spent fewer days in the hospital compared to inpatients. In a study by Gertz et al. [40], of 716 patients undergoing ASCT, 278 (39%) completed the procedure as outpatients. Patients older than 65 were less likely to complete the ASCT as outpatients (only 29%), and those hospitalized had a significantly longer length of stay (median 7 days). Elevated serum creatinine (>1.5 mg/dL) also predicted hospitalization and a longer length of stay (median 9 days, compared to 4 days for patients with normal creatinine, p < 0.001).

The important role of lenalidomide maintenance after ASCT for NDMM was examined in a large meta-analysis which included 1,208 patients. Patients receiving lenalidomide maintenance had a significantly longer median PFS (52.8 months vs. 23.5 months, HR 0.48, 95% CI: 0.41–0.55) and median OS (not reached vs. 86 months, HR 0.75, 95% CI: 0.63–0.9). This study demonstrated an increased rate of second primary malignancies with lenalidomide [44]. In the FORTE trial, 356 patients were randomized to carfilzomib plus lenalidomide maintenance versus lenalidomide alone; PFS at 3 years was 75% versus 65% (HR 0.64, 95% CI: 0.44–0.94, p = 0.023) [25]. Another study randomized 180 patients after different induction regimens and ASCT to maintenance with either lenalidomide alone or KRd. Median PFS was longer with KRd, 59 versus 41 months (HR 0.51, 95% CI: 0.31–0.86, p = 0.012) [45]. Bortezomib maintenance was compared to thalidomide maintenance in a study that included 827 patients [46]. Patients were randomized to either induction with VAD or with bortezomib, doxorubicin, and dexamethasone (PAD). Most patients (84% and 85%, respectively) in both treatment arms proceeded to HDT with melphalan and ASCT, following which 65% of patients in the VAD arm received thalidomide maintenance, and 55% patients in the PAD arm received bortezomib maintenance. PFS was superior in the PAD arm (35 months vs. 28 months, HR 0.75, 95% CI: 0.6–0.90, p = 0.002) as well as OS (HR 0.77; 95% CI: 0.60–1.00; p = 0.049). In a study that randomized 271 patients to different maintenance strategies after HDT with melphalan and ASCT, the combination of bortezomib and thalidomide maintenance was shown to prolong PFS when compared to maintenance with thalidomide alone or IFN (50.6 vs. 40.3 vs. 32.5 months, p = 0.03) [47]. No difference in OS was seen.

The role of maintenance therapy with daratumumab in addition to lenalidomide has recently been explored in the PERSEUS study [34], in which 709 patients were randomized to either induction with bortezomib, lenalidomide, and dexamethasone (VRD) or D-VRD. Post-ASCT maintenance included lenalidomide for all patients, and daratumumab was added in the D-VRD group for 24 cycles. Patients in the D-VRD group had superior PFS at 48 months – 84.3% versus 67.7% (HR 0.42; 95% CI, 0.30–0.59; p < 0.001). However, since all the patients in D-VRD group received daratumumab in induction as well, its role in maintenance still needs to be elucidated. The ongoing AURIGA study (NCT03901963) will randomize patients to maintenance with either lenalidomide alone or combination of subcutaneous daratumumab and lenalidomide, with rate of negative minimal residual disease (MRD) as its primary outcome.

The role of gut microbiota in ASCT for myeloma has been the subject of several studies recently. High-dose chemotherapy and exposure to broad-spectrum antibiotics cause a significant gut injury that persists up to 90 days [48]. A recent study showed that loss of diversity in the gut microbiome at engraftment was associated with an inferior response to therapy [49]. This could be a potential direction for intervention to improve ASCT outcomes. Microbiome diversity was also found to be associated with sustained MRD negativity in patients receiving lenalidomide maintenance, 40% of whom underwent ASCT [50]. The study also showed an association between stool butyrate, which is higher in individuals on a plant-based diet, to MRD negativity, highlighting another potential for intervention.

Chimeric antigen receptor (CAR) T-cell therapy and bispecific T-cell engagers (BiTEs) were demonstrated to be very effective in the relapse/refractory setting in MM [51‒54]. Whether these treatments will replace ASCT as the standard of care for NDMM remains to be seen. Several clinical trials are accruing to answer this question. The KarMMA-4 (NCT04196491) is a phase I trial studying ide-cell in the first line for high-risk patients with MM. Cilta-cell is explored in CARTITUDE-6 (NCT05257083), a phase 3 clinical trial randomizing transplant-eligible patients with NDMM to either ASCT or cilta-cell following D-VRD induction. The role of BiTEs for NDMM is also being explored, both for maintenance after ASCT and as an alternative to ASCT after induction therapy. In the MajesTEC-4 study (NCT05243797), patients are randomized to either maintenance with lenalidomide alone or the combination of lenalidomide and teclistamab, a BCMA-directed BiTE. In a phase 2 study (NCT05849610) currently recruiting, NDMM patients who are transplant-eligible will receive induction therapy with D-VRD, followed by intensification treatment with either teclistamab with daratumumab or talquetamab (a GPRC5D-directed BiTE) with daratumumab, following which treatment strategy will be determined by MRD status.

HDT with melphalan followed by ASCT is currently a part of recommended first-line therapy after induction with novel agents [1‒3], as it prolongs PFS considerably. The procedure is safe with manageable toxicity and low mortality rates. Stem cell mobilization has remained feasible after induction with novel agents, including new regimens with daratumumab. With the use of anti-CD38 antibodies at first line, and the emergence of novel approaches with CAR T cells and bispecific antibodies, whether ASCT will preserve its role as an integral part of first-line therapy remains to be seen.

I.C. and I.V. have no conflicts of interest to declare. M.A.G. served as a consultant for Millennium Pharmaceuticals and received honoraria from Celgene, Millennium Pharmaceuticals, Onyx Pharmaceuticals, Novartis, GlaxoSmithKline, Prothena, Ionis Pharmaceuticals, and Amge.

This study was not supported by any sponsor or funder.

I.C., I.V., and M.A.G. all contributed to writing the manuscript.