Smoldering multiple myeloma: Reviewing the rationale for intervention

Abstract

Smoldering multiple myeloma has been recognized for over 40 years and represents a pre symptomatic phase of the 2nd most common hematologic malignancy. 1/3 of patients will remain asymptomatic at 10 years. There is an identifiable subset of patients that will develop CRAB within 2 years of recognition and these patients are considered for therapeutic intervention before the development of potentially irreversible complications. Obstacles to widespread implementation of therapeutic guidelines is limited by the variable definitions associated with this high-risk group as well as the poor concordance between classification schemes. Analysis of clinical trial outcomes as well as uniform eligibility helps determine whether a given patient should be considered for therapeutic intervention outside of a clinical trial.

Introduction:

In 1980 Kyle and Greipp described 6 patients that were monitored without therapeutic intervention for 5 years or greater. These patients were selected based on having a bone marrow greater than 10% plasma cells or serum monoclonal protein greater than 30 gram/liter. This represented a retrospective selection among 334 patients. The authors recognized that there was a subset of patients who fulfilled criteria for multiple myeloma but could do well without treatment.¹ They described idiopathic Bence-Jones proteinuria 2 years later, reporting 7 patients with over 1 gram of light chain in the urine who had been monitored for a minimum of 7.7 years. This represents the light chain equivalent of MGUS in the era before the development of the immunoglobulin free light chain assay.².

Subsequently a prospective cohort of patients fulfilling these criteria (light chain smoldering myeloma had not been defined at this time) were monitored over a period of 26 years and a triphasic pattern was recognized. Over the first 5 years the development of symptomatic multiple myeloma occurred with an annual rate of 10%. The slope leveled off after 5 years and the evolution rate fell to 3% annually for the next 5 years. At 10 years therefore 1/3 of these patients remained on observation not requiring therapy. After 10 years the rate of progression was 1% annually identical to that of monoclonal gammopathy of undetermined significance. Even at 20 years of monitoring these patients, who fulfilled criteria for smoldering multiple myeloma, 20% remained free of therapy.³.

It was recognized, there were instances where potentially irreversible complications would develop as these patients evolved into active malignancy even with optimal monitoring.⁴. The goal is to identify patients with smoldering multiple myeloma with a high risk of developing active multiple myeloma in the near term. For these patients the risks of withholding therapy, not wishing to expose asymptomatic patients to toxic treatments, are higher than the benefits. This would include the prevention of myeloma cast nephropathy or spinal compression fractures that could result in lifelong pain. The questions therefore are as follows: Can a subset be identified that is considered high risk for whom the benefits of immediate therapy outweigh the risks of long-term therapy related complications. A proportion of patients might never need treatment but could be exposed to multi drug chemotherapy including stem cell transplantation? Have we identified the correct subset for intervention? What constitutes optimal treatment of the subgroup of high-risk smoldering multiple myeloma?

Classification of the high-risk group of multiple myeloma

High-risk smoldering multiple myeloma is not a disease entity. It is a statistical concept to predict those patients most likely to develop active disease in the near term for whom a chemo prevention strategy is justified. In the original classification scheme from Mayo Clinic high-risk smoldering multiple myeloma was considered those patients who had both a monoclonal protein greater than 30 grams/liter and plasma cells> 10%. This group had a 5 and 10-year risk of progression to symptomatic multiple myeloma of 69 and 77% respectively. I would not consider this a sufficiently high risk to justify therapeutic intervention outside of a clinical trial and I suspect that patients who were told that there was a 30% chance of not needing therapy at 5 years, many would decline participation in a trial based on these definitions of high risk smoldering myeloma.

The 1st model to be reported is from the Spanish multiple myeloma group.⁵ this model was used for randomization in the 1st phase 3 trial on chemo prevention for smoldering multiple myeloma.⁵ This model however is not applied in practice because the application of flow to detect the aberrant plasma cells in the bone marrow is not in widespread use⁶. The 2nd model published by Mayo Clinic in 2008 modeled the risk of progression based on the size of the M protein, the percentage of plasma cells in the bone marrow, and the free light chain ratio. (table 1) This produced 3 groups of almost equal size with different risks of progression to multiple myeloma. However even in the highest risk group only half actually developed overt multiple myeloma in the first 2 years. Both studies included patients with smoldering multiple myeloma based on data available at the time, but this preceded the concept of the myeloma defining event. Therefore, patients to be evaluated for these studies did not require whole-body MRI or PET scanning^7,8 which if positive would remove the patient from the smoldering myeloma group and put them into the active multiple myeloma group. Even though light chains were measured in the 2008 Mayo group the cohort may have included patients whose free light chain ratio was greater than 100 and therefore would not be considered smoldering multiple myeloma. Caution needs to be exercised because populations previously reported would contain patients today considered to have active multiple myeloma and not candidates for observation. This would result in a reduced risk of transformation of high-risk smoldering patients if those who already have multiple myeloma by todays criteria are removed.

Table 1.

Risk Stratification Schemes for Smoldering Myeloma

Stratification model	Parameters	Projected outcomes	Comments
Spanish Risk Model	1 point for: Suppression of uninvolved immunoglobulins, aPC/BMPC≥95% or DNA index abnormal	5-year risk of progression; 0–4%; 1–46%; 2–72%	Flow for aPC/BMPC not widely available
Mayo 2008	1 point for: BMPC≥10%, M protein ≥ 30g/L, iFLC/uFLC>8	Median time to MM yrs.; 1–10; 2–5.1; 3–1.9	High risk group only 50% risk of MM@2years. Result not confirmed by British. Poor concordance with Spanish. Myeloma defining events not excluded
Mayo 2018	1 point for: BMPC>20%, M protein >20g/L, iFLC/uFLC>20 in	Median time to myeloma mos.: 0–110; 1–68; 2–3 29	Ultra-high-risk group not defined. Light chain ratio may only be applicable with light chain proteinuria.
IMWG 2020	Point scale from 0–17; BMPC 4 strata; M protein size 3 strata; FLC 4 strata; FISH	Risk of progression at 2 yr, %; 0–4 3.8; 5–8 26.2; 9–12 51.2; ≥12 72.5	Likely the best model but very complex with a 17-point scale. 12 points defines a 72.5% risk of overt disease in 2 years so approaches true ultra-high risk. Confirmation by other groups pending

Modern systems for risk assessment include the Mayo 2018 model easily remembered as the 20/20/20 model.⁹ This results in a point system from 0 to 3 where the highest risk category had a median time to development of active myeloma of 29 months. (table 1) One would not consider this an ultra-high-risk category if only half of the patients develop multiple myeloma in 2–1/2 years. The most recently reported model was that by the International myeloma Working group. While this system may be useful in defining clinical trial eligibility it is unlikely to be used in clinical practice because of its complexity. These criteria are based on a point system that can range from 0 to 17. The percentage of plasma cells in the bone marrow provides from 0 to 6 points, the size of the serum M protein can provide 3 or 4 points, the free light chain level can provide from 0 to 5 points and the presence of genetic abnormalities can provide further stratification. This data set of 1996 patients also validated the Mayo 2018 staging system and added the importance of FISHmag04 abnormalities which included t(4;14), t(14;16), +1q, and/or del13q. ¹⁰ One advantage of the system is that it provides a graded risk assignment rather than a simple binary assignment characteristic of the Mayo staging system. In addition, the Mayo staging systems have equal weighting for each parameter (0 points or 1 point), the IWMG does not assign equal weights to each parameter making it a sounder method for an overall risk assignment.

When patients are treated for multiple myeloma meeting CRAB criteria the percentage of plasma cells in the bone marrow, or the free light chain ratio is prognostic but not relevant in terms of initiating therapy. However in smoldering multiple myeloma when the risk assessment criteria is greater than or less than 20% plasma cells the method of estimation becomes important since there is significant discordance in the percentage of plasma cells when estimated by histologic review, smear review or by flow.¹¹ Concordance rates were only at 60.7% between bone marrow aspiration and bone marrow biopsy and 35.7% between bone marrow flow and bone marrow biopsy.¹² In a total of 389 newly diagnosed MM patients, 81.2% (316/389) patients had a PC >10%, and the remaining 18.8% (73/389) of patients had either a PC <10% or were unavailable for bone marrow aspirate PC estimation. Of these 73 patients, 55 patients showed PC >10% after estimation by image analysis in bone marrow biopsy.¹³ Awareness of how the laboratory is providing the plasma cell estimate is important in assessing smoldering myeloma risk of progression.

A similar problem exists when assessing the immunoglobulin free light chain ratio. Interlaboratory variability as well as differences based on the degree of sample dilution can result in ratio changes. This is a particular problem when the uninvolved light chain level is low where very small changes can result in dramatic alterations in the ratio.¹⁴ Significant measurement variability is common in the measurement of sFLC. ¹⁴

Despite the importance of these risk stratification models in smoldering multiple myeloma, there remain major drawbacks. Ideally the various classification schemes should identify the same group of patients that would benefit from early intervention. It is somewhat disconcerting if the models identified different patients that would recommend treatment in 1 model but not in another model. Using an independent validation cohort of 77 patients with smoldering myeloma. Patients were assigned low intermediate and high risk based on both the Spanish model and the Mayo 2008 model. The Mayo Clinic model identified 38, 35 and 4 patients as low, medium, and high risk, respectively. The Spanish PETHEMA model classified 17, 22 and 38 patients as low, medium, and high risk, respectively. Note that among the same 77 patients four were identified as high risk by the Mayo model and 38 by the Spanish model. A completely different patient population will therefore be offered chemo prevention. When the Spanish model of smoldering myeloma was compared to the Mayo model, which was the basis of the Eastern Cooperative Oncology group smoldering myeloma trial,¹⁵ the concordance rate was only 28.6% suggesting that cross trial comparisons of high-risk smoldering multiple myeloma using different classification schemes is inappropriate.

A trial comparing Mayo 2008, Mayo 2018, and the Spanish model in two clinical trials of 145 patients with smoldering myeloma was performed. The overall global rate of agreement across all 3 models for all 3 categories was a disappointing 16.6%. Even the Mayo Clinic comparison of the 2008 and 2018 models which used the same 3 parameters with different cutoff values identified high risk in 8 versus 46 patients respectively among the 145 patients.¹⁶

Before subjecting asymptomatic patients to therapy they should be assured of a significant risk of developing multiple myeloma over the subsequent 24 months. A study of 157 smoldering myeloma patients from the United Kingdom used the Mayo 2018 criteria to assess the risk of progression to active myeloma. In the original Mayo publication high risk patients had a median time to progression of 29 months and in the British smoldering cohort trial the median time to progression for patients with high risk was 67 months (46% at 5 years). Mayo 2018 staging may be population dependent. When staging is applied to a different population results may change. In the United Kingdom study patients that had free light chain ratios of greater than 100 were included as they were managed as smoldering myeloma. Presumably if they were excluded the progression rate would be even lower. This heterogeneity in time to progression in patients with smoldering myeloma with the same criteria applied is disturbing since it makes cross trial comparisons of outcomes next to impossible and renders interpretation of single arm phase 2 trials, other than safety assessment, of very little utility because of its population dependency.¹⁷ The most striking example of this variability is a direct comparison of the Spanish and US cooperative group trial. Table 2 demonstrates that the 3-year progression-free survival in the Spanish cohort was 30%, in the U.S. cohort was 66%. Even in the treated population 80% were progression-free in the Spanish group, 91% in the American group. A survival advantage was seen at 3 years in the Spanish trial but not in the United States trial. This illustrates the variability of smoldering myeloma populations and the inadvisability of attempting a cross trial comparison. A review of 17 trials for smoldering myeloma had 6 different definitions of high-risk smoldering myeloma for enrollment. Only 4 adopted the definition of high-risk smoldering multiple myeloma by the International myeloma Working group. Results may not be applicable to studies that have used other entry criteria.¹⁸ This variability also applies to the interpretation of outcomes of any phase 2 trial where the outcome may be as much dependent on the population selected as to the therapy itself.

Table 2.

Reported results of Phase 3 trials of intervention high risk smoldering myeloma

	Intervention	PFS @3 years; %	OS@3 yr.;%
Spanish Trial	Len/dex vs 0	80 vs 30 p<0.001	93 vs 75 p=.03
ECOG US trial	Len vs 0	91 vs 66 p=0.002	98 vs 96 p NS

The endpoints for smoldering myeloma trials are not limited to response duration and response depth. There are vaccine trials underway designed to harness the patient’s immune system and maintain long term stability of smoldering disease. One approach is to use the patient’s plasma cells to create a personalized vaccine. (NCT03631043). A 2nd approach is to use PD-L1 blockade to prevent progression of the myeloma.¹⁹

Phase 3 trials attempting to improve outcomes in high-risk smoldering myeloma

With the concerns raised above phase 3 trials designed to improve outcomes are desirable to assess the relative merits of available therapeutic interventions. In newly diagnosed multiple myeloma it is generally agreed that relevant end points include the MRD negativity rate and a stringent complete response rate. It is also agreed that for relapsed multiple myeloma, progression-free survival is appropriate when compared to a standard of care regimen. What however constitutes an ²¹appropriate endpoint for chemo prevention trials in asymptomatic patients?²² It is questionable whether an endpoint such as response depth or duration is appropriate for smoldering myeloma trials because the risk of lead time bias from the initiation of therapy in patients who may be asymptomatic for years is very real. Lead time bias may distort interpretation of outcomes. Lead-time bias will occur when the definition of high-risk smoldering myeloma reclassifies asymptomatic patients as active disease. The earlier diagnosis (upstaging) may do nothing to change the course of the disease.²³ Overall survival and Health related quality of life are important endpoints for these types of trials and control arms are essential to answer whether early treatment will result in improved survival or QOL compared to careful observation. Trials of smoldering multiple myeloma intervention are fundamentally chemo prevention trials. Quality of life should be used as another biomarker to ensure intervention on asymptomatic patients is valuable. The impact on QOL should be modest in order to justify their use.²⁴

The literature is currently replete with references to cure as a goal in smoldering myeloma trials.²⁵ Unfortunately, the tools currently available have not realized any so-called cure. Even patients who achieve stringent complete response with MRD negativity at 10 ⁻⁶ are not showing plateaus in the survival curves that could suggest an appreciable proportion of patients are being cured. ²⁶ It will be important to report in future trials whether clinically significant irreversible end-organ damage is prevented with early intervention, as opposed to M protein measures of progression. As early as 1985 it was suggested that the optimal method of identification of stable smoldering myeloma was serialized measurements of the M protein.27 The use of dynamic monitoring of parameters over time has been validated by demonstrating that evolving changes in the level of the myeloma protein and decline in hemoglobin accurately predicts for progression of smoldering myeloma. A decline in hemoglobin of greater than 0.5 gram/deciliter within 12 months and rise in the M protein greater than 10% within 6 months or a 25% increase in M protein within 12 months all were sensitive measures of the likelihood of symptomatic myeloma. Following the diagnosis of monoclonal gammopathy of undetermined significance and smoldering myeloma the velocity at which the M protein increases over time has been found to be highly prognostic and should be included in the assessment of risk of evolution to smoldering myeloma.^28,29 These dynamic risk assessments may be better parameters for monitoring rather than attempting to make a decision at a single time point.³⁰Deciding on intervention may also be impacted by a patients frailty as these patients are less likely to tolerate interventions well when compared to fit counterparts. These patients tend to have lower response rates and greater levels of therapy related toxicity.^31,32

The 1st trial and the only one that has demonstrated a survival advantage is the Spanish trial randomizing patients to lenalidomide dexamethasone versus observation ³³ When this trial activated demonstration of the absence of skeletal disease with PET or MRI was not part of the eligibility criteria. The presence of a free light chain ratio> 100 now recognized to be a myeloma defining event was not an exclusionary criterion for this protocol so some of the patients may have fulfilled criteria for active myeloma. Patients with a bone marrow plasma cell percentage greater than 60% were also admitted to the trial. Although a survival advantage was seen, the trial was not powered for this type of analysis. For patients on the placebo arm to begin active therapy CRAB criteria had to be fulfilled and no allowance was made for rapid change in the M protein over time. Per protocol specifications delaying intervention in the face of a rising M protein awaiting a skeletal event can impact overall survival. Quality of life analysis was not performed. This trial was done in Spain where lenalidomide was not approved for the initial therapy of multiple myeloma patients that progressed and went off study. These patients did not have a defined therapy after they went off trial but may easily have lacked access to an immunomodulatory agent. The specifics of the salvage therapy in a placebo-controlled trial such as this ideally should be pre specified (preferably crossover to the therapy arm not simply off study) to ensure inferior therapy at the time of progression is not contributing to the inferior survival. Patients were eligible for enrollment if the diagnosis of smoldering multiple myeloma was made within 5 years of trial entry. One would suspect that the risk of transformation in a patient who at study entry had been monitored for 5 years would not be the same as a newly recognized patient fulfilling criteria for high-risk smoldering myeloma. Second primary malignancies were seen in 6% of the treatment group and 2% of the non treatment group. The risk of 2nd primary malignancy must be considered when asymptomatic patients are offered therapy.

The most recent phase 3 trial was a cooperative group effort, randomizing patients recognized within 5 years of registration to lenalidomide versus observation.³⁴ No survival benefit was seen, and the patients were not uniformly high risk as 49 of the patients that were enrolled were low risk smoldering by Mayo Clinic criteria. Forty-seven of the patients had an abnormal magnetic resonance imaging making it unclear whether they fulfilled criteria for a myeloma defining event. The study randomized only 29 high risk patients into groups of 15 and 14 respectively. In this trial at 3 years 2/3 of the patients in the control arm were stable and had not progressed to multiple myeloma reflecting the low-risk status of the enrolled population. In my practice I would have some difficulty convincing a patient to embark on therapy when their chance of developing active disease over the next 3 years was only 1 in 3. Despite the finding of no survival benefit the successor study has no control arm and will randomize patients to 2 arms; daratumumab, lenalidomide, dexamethasone versus lenalidomide, dexamethasone, the DETER-SMM (NCT03937635). These patients fortunately will have to have been recognized to have smoldering multiple myeloma within 12 months of registration. Another phase 3 trial currently actively registering patients is comparing carfilzomib lenalidomide dexamethasone to lenalidomide dexamethasone (NCT03673826). Other agents that are being explored in the phase 2 setting in combination for high-risk smoldering myeloma include isatuximab, elotuzumab, ixazomib, iberdomide, leflunomide and dendritic cell vaccination.

What are characteristics of appropriate interventions for smoldering myeloma?

If we assume that high-risk smoldering myeloma represents a heterogeneous group of patients with MGUS like illness, overt multiple myeloma in evolution and slowly progressive plasma cell disorders that ultimately will evolve into multiple myeloma, it is unclear how 1 regimen would be appropriate for this heterogeneous group. Standard of care for the treatment of newly diagnosed multiple myeloma has been established and includes triplet therapies including bortezomib lenalidomide dexamethasone, bortezomib thalidomide dexamethasone, daratumumab lenalidomide dexamethasone as well as carfilzomib containing triplet induction therapy. ³⁵ Recently several trials have incorporated quadruplet therapy with anti CD38 antibodies as part of induction in transplant eligible populations.³⁶ If the consensus for induction of myeloma is triplet or quadruplet therapy by what rationale is lenalidomide or lenalidomide dexamethasone appropriate induction for high-risk smoldering disease outside of a clinical trial setting? One could easily argue that for patients at high risk of developing overt multiple myeloma doublet therapy could potentially constitute undertreatment of disease not meeting todays standard of care for newly diagnosed disease. This was appropriate during clinical trial investigations at the start of trials attempting to manage smoldering myeloma. However, making the jump into clinical practice simply does not translate appropriately. Most current phase 2 trials of newly diagnosed myeloma have made the transition to 3 and 4 drug combinations. This represents the best approach to patients with myeloma but also increases the risk of multi drug toxicity for that subset of patients that are not destined to develop multiple myeloma in the subsequent 3 to 5 years. Depending on the regimen those patients, not destined to develop multiple myeloma, may yet still develop intractable painful neuropathy seen in as many as 30% of patients receiving a bortezomib containing regimen, recurrent upper respiratory infections related to suppression of immunoglobulin production, venous thromboembolism in spite of prophylaxis, cardiopulmonary complications most commonly associated with carfilzomib and 2nd primary malignancies which is now being reported in upwards of 6% of lenalidomide exposed patients. In a recent report, 12 transplant ineligible myeloma patients among 280 developed a 2nd primary malignancy.³⁷ A recently published trial showed a 10 year 2nd primary malignancy rate of 16.6% in non transplanted patients.³⁸ When quality of life is a relevant endpoint additional years of dexamethasone becomes an important consideration as this appears in virtually every new phase 2 and phase 3 trial on clinical trials.gov for the management of smoldering myeloma. This agent is generally despised by patients because of the insomnia, hyperirritability, hypomania, and significant mood swings that impact on the patient and family. Steroid’s ability to increase infection risk, accelerate bone loss and precipitate diabetes mellitus, often insulin-dependent, is well known. These toxicities must always be balanced against the positive preventative impact these agents have in preventing overt multiple myeloma.

Conclusion

Clinical trials investigating interventions for the prevention of the development of overt multiple myeloma are important to improve outcomes for this disease. Ongoing enrollment of patients into multi agent clinical trials is strongly encouraged. However, outside the context of clinical trials the community oncologist should be warned that a patient that fulfills statistical criteria for high-risk smoldering myeloma should not be prescribing lenalidomide or lenalidomide and dexamethasone as part of routine daily practice. Given the rapid introduction of new therapies for multiple myeloma, it is possible that delayed intervention could potentially allow patients access to newer and more effective regimens. The field requires further definition of appropriate regimens and further refinement of current systems to identify patients truly at high risk for the imminent development of symptomatic multiple myeloma. The inherent heterogeneity of a smoldering myeloma population is a challenge in finding the correct population for intervention. Future research aimed at myeloma defining genomic events may improve our ability to differentiate asymptomatic patients with myeloma destined to remain stable or progress to symptomatic disease.³⁹ This is required before any chemoprophylactic regimen can be considered standard of practice in the community.