https://orcid.org/0000-0001-7336-3003
TO THE EDITOR:
Smoldering multiple myeloma (SMM) is an asymptomatic condition that precedes newly diagnosed multiple myeloma (ND MM) and is present in ∼1 of 200 individuals (0.5%) aged >40 years.1 The optimal management of patients with high-risk SMM (HR-SMM) remains controversial. Among proponents of early intervention in HR-SMM, there are predominantly 2 strategies that have been tested in clinical trials: 1 with low-intensity regimens, such as lenalidomide and dexamethasone,2,3 aimed at preventing morbidites such as fractures and renal failure, and another strategy which instead uses intensive regimens with the aim of cure. Intensive approaches include multidrug combination induction therapy with or without high-dose melphalan and autologous hematopoietic cell transplantation (HDM-AHCT).4-6 An important distinction between these 2 strategies is that although the former has been subjected to randomized controlled trials (RCTs) against observation, the latter, to date, has only been tested in single-arm studies.7
The pursuit of fixed-duration intensive treatment strategies for HR-SMM seeks to potentially cure a subset of patients; based on the assumption that, if a curative treatment exists, a patient with a lower burden of disease (ie, SMM) may have a higher chance of being cured by it than a patient with symptomatic myeloma. Curability is, however, challenging to assess in uncontrolled trials, especially when the standard-of-care treatment for ND MM, imaging modalities to detect disease,8 and the diagnostic criteria for ND MM and for precursor states9 are all constantly evolving. Furthermore, single-arm trials can potentially mask signals of harm, which are of particular importance in patients who are asymptomatic. This is especially pertinent because, even among patients deemed to be at a high-risk by contemporary models, only 50% will progress to myeloma in the 2 years following diagnosis.10 Furthermore, the different risk-stratification models for SMM that are used to select patients at high-risk for early intervention have poor concordance with each other.11,12 In this viewpoint, we discuss the potential perils of pursuing intensive treatment strategies in HR-SMM highlighted by the GEM-CESAR and ASCENT trials (Table 1).
Salient features of the GEM-CESAR and ASCENT trials testing aggressive intervention in high-risk SMM
| . | GEM-CESAR . | ASCENT . |
|---|---|---|
| Inclusion criteria | High risk defined by the Spanish/PETHEMA model | High risk defined by the Mayo 20/2/20 model or IMWG risk-stratification score ≥ 9 |
| Mandatory MRI at screening | No∗ | No |
| Treatment schema | KRD × 4 → HDM-AHCT → KRD × 2 → R maintenance for 2 y | Dara-KRD × 24 cycles |
| Primary end point | Sustained MRD-negativity at 4 and 5 y by MFC (10−5) | sCR rate |
| . | GEM-CESAR . | ASCENT . |
|---|---|---|
| Inclusion criteria | High risk defined by the Spanish/PETHEMA model | High risk defined by the Mayo 20/2/20 model or IMWG risk-stratification score ≥ 9 |
| Mandatory MRI at screening | No∗ | No |
| Treatment schema | KRD × 4 → HDM-AHCT → KRD × 2 → R maintenance for 2 y | Dara-KRD × 24 cycles |
| Primary end point | Sustained MRD-negativity at 4 and 5 y by MFC (10−5) | sCR rate |
Dara-KRD, daratumumab-carfilzomib-lenalidomide-dexamethasone; MFC, multiparameter flow cytometry; R, lenalidomide.
Based on the presentation at the 2022 Annual Meeting of the American Society of Hematology,4 bone disease was ruled out by low-dose whole-body computed tomography or positron emission tomography/computed tomography. The complete protocol is not yet available.
Firstly, because of the rapidly expanding diagnostic and therapeutic landscape of plasma cell disorders, the patient populations enrolled in early intervention trials of HR-SMM in the past 2 decades were highly heterogeneous. This is further exacerbated by heterogeneity of inclusion criteria in currently enrolling HR-SMM clinical trials.13 A change in the diagnostic criteria of MM in 2014 led to patients who are asymptomatic with myeloma-defining biomarkers (≥60% plasma cells, a serum-free light chain involved-to-uninvolved ratio of ≥100 and/or >1 focal lesion detected via magnetic resonance imaging [MRI]) being classified as having MM despite the absence of end-organ damage.9 These patients would have previously been considered to have ultrahigh-risk SMM.14 Therefore, risk-stratification models from before 2014 may lack validity today. One such risk-stratification model is the Spanish or PETHEMA model, which included patients diagnosed with SMM between 1996 and 2003.15 Because advanced imaging (positron emission tomography and/or whole-body MRI) was not part of the diagnostic workup for SMM in that era, the cohort would likely have included patients who would now be classified as having myeloma. It is unclear how the model would perform if patients with lytic or focal lesions, detected via advanced imaging, were excluded. Although the Mayo 20/2/20 risk-stratification model included a more contemporary cohort than the Spanish model, a limitation of this model is the lack of sensitive imaging (eg, MRI at diagnosis was missing in the majority of patients (∼70%).16 Notably, the GEM-CESAR trial used the Spanish/PETHEMA model for inclusion of patients with HR-SMM, with 52% of patients meeting inclusion criteria only based on the Spanish model and not the Mayo 20/2/20 model.4 Furthermore, although the trial started accruing from June 2015, 33% of patients met the biomarker-driven criteria for symptomatic MM that was published in November 2014. Hence, the external validity of the long-term data from the GEM-CESAR trial for patients with HR-SMM per the current criteria will remain unclear. The cohort in the ASCENT trial was more representative of patients with contemporary HR-SMM, because only patients meeting the Mayo 20/2/20 criteria for high-risk or with International Myeloma Working Group (IMWG) risk scores ≥9 were included.5 However, MRI, the gold-standard for imaging in SMM per IMWG criteria because of its higher sensitivity,17-19 was not mandated at study entry in either the GEM-CESAR or ASCENT trial. Some patients with focal lesions on MRI may thus also have been included in the ASCENT trial.
Secondly, the risk of harm (including treatment-related deaths) after intensive treatment in an asymptomatic population that may not progress to MM ought to be considered. In the GEM-CESAR trial, 4.4% (4 of 90) of patients died because of reasons unrelated to myeloma progression, namely ischemic stroke during induction, cardiac arrest, lung cancer, and myelodysplastic syndrome (1 each).4 In the ASCENT trial, which had a shorter follow-up, 3 out of 87 patients (3.4%) died during treatment due to COVID-19 pneumonia (2 patients) and respiratory syncytial virus (1 patient).5 Furthermore, 51% of the patients experienced nonhematological grade ≥3 toxicities, including venous thromboembolism and cardiovascular complications. Without a control arm, ideally patients in whom the natural history of disease and competing causes of mortality can be assessed in parallel, a treatment-related death signal can be hidden in single-arm studies. As an example, the increased mortality signal with the use of pembrolizumab in myeloma was not apparent in early-phase trials, until 2 RCTs demonstrated worse overall survival with pembrolizumab when used in combination regimens with immunomodulatory drugs and steroids.20,21 Hence, we propose that future RCTs testing treatment strategies in SMM should have active surveillance as a control arm to ensure that we are accurately capturing both potential benefits and harms of early intervention.
Thirdly, an important objective of these HR-SMM trials is to assess the curative potential of early intensive intervention. However, the primary end point of GEM-CESAR was measurable residual disease (MRD)-negativity rate at a sensitivity of 10−5 via multiparameter flow cytometry at 4 and 5 years after AHCT, and that of the ASCENT trial was stringent complete response (sCR) rate after treatment. Although the ASCENT trial had a prespecified null hypothesis of sCR rate ≤ 65% and the alternative hypothesis of sCR rate ≥ 80%, to our knowledge, the GEM-CESAR trial has not yet reported its null hypothesis for sustained MRD-negativity rates. The MRD-negativity rate based on the intention-to-treat was 62% (56 out of 90 patients) at 3 months after AHCT and 28% (25 out of 90 patients) at 4 years after AHCT in GEM-CESAR. With only 28% of patients sustaining MRD-negativity at 4 years of follow-up, this raises substantial questions about whether patients are actually being cured with these approaches.
Furthermore, the MRD-negativity rates observed in this population with SMM do not surpass what was observed with the use of the same regimen (carfilzomib-lenalidomide-dexamethasone followed by HDM-AHCT) in ND MM in the FORTE trial, which had post-AHCT and 1-year sustained MRD-negativity rates of 62% and 47% respectively.22 Similarly, the posttreatment MRD-negativity rate (10−5; multiparameter flow cytometry) in ASCENT was 84%, comparable with the MRD-negativity rate with daratumumab-carfilzomib-lenalidomide-dexamethasone without HDM-AHCT in ND MM.23 Notably, the primary end point of ASCENT, that is, the sCR rate, was 43%, suggesting that the study might have failed to meet its primary end point. Long-term data from ASCENT are needed to assess sustained MRD-negativity after patients are off treatment. In summary, data from the GEM-CESAR and ASCENT trials suggest that depth of response with early intervention is not superior to what we can achieve by treating ND MM with these regimens. Furthermore, the 3-year progression-free survival with intensive treatment strategies in GEM-CESAR and ASCENT is similar to that with lenalidomide in the E3A06 trial (∼90%),3-5 although the intrinsic value of progression-free survival for an asymptomatic condition such as SMM is questionable. These findings challenge the hypothesis that intensive treatment of patients with SMM should lead to higher rates of cure compared with the rates of cure among patients with IMWG-defined myeloma. It is also plausible that the indolent biology and growth kinetics of SMM might make it less curable than ND MM.
Fourthly, quality of life (QoL) assessments are extremely important,24 particularly when studying asymptomatic diseases such as SMM. Neither GEM-CESAR nor ASCENT had QoL measures as primary or secondary end points. Focusing on disease response in asymptomatic patients can underestimate both short- and long-term treatment consequences that include disruption of normal daily activities, time toxicity,25 adverse events leading to hospitalization, patient and caregiver stress, and financial toxicity for patients and caregivers.26 For example, 1 patient in the GEM-CESAR trial developed graft failure after AHCT, which may result in long-term transfusion dependence. Furthermore, patients in the GEM-CESAR trial were offered daratumumab-pomalidomide-dexamethasone combination therapy for biochemical progression; such patients may, thus, rapidly exhaust most active antimyeloma treatment options, with additional uncertain long-term clinical implications on risk of secondary malignancies.27
The cumulative data from these studies should instill skepticism toward the hypothesis that early intervention with intensive treatment strategies in SMM is the optimal path to curing myeloma. Furthermore, these data should prompt us to challenge the utility and acceptance of uncontrolled studies of asymptomatic SMM. This is especially true in the era of advanced diagnostics such as whole-body diffusion-weighted MRI that can be used longitudinally to identify focal lesions before the appearance of lytic lesions or fractures. One prospective study that undertook active surveillance of SMM with serial MRIs and laboratory tests demonstrated that the vast majority of patients who progressed did so without morbid events such as fractures or renal failure.28 Dynamic risk scores that depend on multiple time-point tests will be needed to improve current risk-stratification systems and guide treatment initiation.12,29 Furthermore, the use of genomics that might predict, with greater accuracy, which patients are likely to progress to MM, will also aid in future decision making30-32 and pave the way for RCTs testing early intervention strategies in patients with genomically defined HR-SMM. Although the ongoing DETER-SMM trial (NCT03937635) will answer an important question, whether time-limited intensive treatment (daratumumab-lenalidomide-dexamethasone) improves OS when compared with low-intensity treatment (lenalidomide-dexamethasone) in patients with HR-SMM, it will remain unknown whether early therapeutic intervention improves OS compared with active surveillance in the modern era (ie, patients with contemporary HR-SMM followed with serial imaging). We, therefore, propose that an active surveillance strategy with advanced imaging should still be used as the control arm for RCTs testing early intervention in HR-SMM. In smaller hypothesis-generating trials in which having a control arm may not be feasible, prespecified boundaries for treatment-related toxicity and death as well as QoL assessments should be incorporated. In addition, end points that account for death and morbid or symptomatic progression events such as pathologic fracture or irreversible renal failure should be captured in clinical trials to better identify therapeutic benefit of early intervention.
Contribution: R.C. conceived the idea and wrote the first draft of the manuscript; and S.A.H., E.R.S.C., and G.R.M provided critical input, extensively reviewed the first draft, and approved the final draft of the manuscript.
Conflict-of-interest disclosure: R.C. declares consulting for and serving on the advisory board of Janssen, Sanofi, and Adaptive Biotech. The remaining authors declare no competing financial interests.
Correspondence: Rajshekhar Chakraborty, Multiple Myeloma and Amyloidosis Program, Herbert Irving Comprehensive Cancer Center, Columbia University Irving Medical Center, New York, NY; e-mail: rc3360@cumc.columbia.edu.
