SLiM CRAB criteria revisited: temporal trends in prognosis of patients with smoldering multiple myeloma who meet the definition of ‘biomarker-defined early multiple myeloma’—a systematic review with meta-analysis

Summary

Background

Biomarker-defined patients with smoldering multiple myeloma (SMM) were included in the diagnostic category of multiple myeloma (MM) by the International Myeloma Working Group (IMWG) in 2014. This includes ≥60% bone marrow plasma cells (BMPCs), free light chain ratio (FLCratio) ≥100, and >1 MRI-defined ≥5 mm focal lesion, also called SLiM CRAB MM. We examined whether the risk of progression of SLiM CRAB MM patients to CRAB positive MM described in recent studies differs from that reported in earlier studies published before the introduction of the new diagnostic criteria.

Methods

We conducted a systematic review with meta-analysis, and included studies on Embase and PubMed (01/01/2010–01/11/2022), selecting studies with digitizable progression curves. Inconsistent studies were excluded. We created forest plots using random effects models from digitized and published data and Kaplan–Meier curves. Main outcomes were median time to progression (TTP), 2-year progression risk, and odds ratios (ORs) comparing 2-year progression risks.

Findings

Our meta-analysis including 11 studies with 3482 patients found an approximately 3-fold longer TTP and 50% lower 2-year progression risk of SliM CRAB MM patients in recent (published after 2014) compared with earlier studies. Median TTP in patients with ≥60% BMPCs was 30.31 months [18.71–62.93] in recent compared with 9.20 months [6.02–15.56] in earlier studies; the 2-year progression risk was 45.45% [20.12–62.75] compared with 86.21% [65.74–94.45] in the respective time periods. In patients with a FLCratio ≥ 100, the median TTP was 48.06 months [40.51–64.91] vs. 15.33 months [9.38–19.10], and the 2-year progression risk was 31.61% [25.30–37.39] vs. 73.00% [62.39–80.62] in recent and earlier studies, respectively. Tests for heterogeneity showed that the two time periods differed significantly in their ORs when comparing patients who met the high-and low risk criteria. No appropriate recent studies on focal lesions have been published.

Interpretation

Recent studies show significantly improved prognosis of biomarker-defined MM with ≥60% BMPCs and FLCratio ≥ 100. This warrants careful evaluation for signs of progression before treatment initiation.

Funding

Funding was provided by the Austrian Forum against Cancer.

Research in context.

Evidence before this study

The updated diagnostic criteria by the IMWG published in 2014 included patients with SMM fitting the criteria of ≥60% BMPC or FLC ratio ≥100, or ≥1 MRI defined focal lesion in the category of MM. This decision was implemented because of the short time to progression indicated by the evidence available at that time. As a consequence, those patients were deemed to be candidates for treatment initiation.

We searched PubMed and Embase (restricted to 01/01/2010–01/11/2022) using the search terms ‘myeloma’, ‘smoldering multiple myeloma’, ‘gammopathy’, ‘MGUS’, ‘SLiM CRAB’, ‘bone marrow plasma cells’, ‘BMPC’, ‘Free light chain’, ‘FLC’, ‘lesion∗’, and ‘progress∗’, excluding Case Reports and Reviews. We included publications with suitable graphs and excluded publications which were inconsistent. Our search resulted in 11 studies (3482 patients) suitable for meta-analysis, which resulted in estimates for the median TTP (≥60% BMPCs: recent study: 30.31 months; early studies: 9.20 months; FLCratio ≥ 100: recent studies: 48.06 months; early studies: 15.33 months) and the 2-year progression risk (≥60% BMPCs: recent study: 45.45%; early studies: 86.21%; FLCratio ≥ 100: recent studies: 31.61%; early studies: 73.00%). Tests for heterogeneity showed the two time periods differed significantly in their ORs when comparing high- and low-risk patients (BMPCs: late OR = 3.27 vs. early OR = 27.01, I² = 78.6%, p = 0.009; FLCratio: late OR = 2.69 vs. early OR = 7.03, I² = 67.8%, p = 0.005). No suitable recent studies were published on focal lesions. Risk of bias was deemed low by use of a funnel plot.

Our analysis revealed a significantly improved prognosis of patients with biomarker-defined MM (≥60% BMPC or a FLC ratio of ≥100) in studies published after the IMWG consensus statement in 2014 compared to earlier reports. Based on these new findings, those patients should be carefully evaluated for signs of progression before treatment initiation, as the majority of them will not progress within 2 years. Future research is needed to refine the prognostic tools for risk assessment to optimize decisions between immediate treatment initiation or a wait-and-see strategy.

Added value of this study

The time to progression is significantly longer in patients with biomarker-defined MM and ≥60% BMPC or a FLC ratio of ≥100 than previously found.

Implications of all the available evidence

Myeloma therapy should not automatically be initiated in patients with biomarker-defined MM and 60% BMPC or a FLC ratio of ≥100. Instead, patients should but be carefully evaluated for signs of progression, which would merit treatment initiation.

Introduction

Since the first description of smoldering multiple myeloma (SMM) by Kyle and Greipp in 1980,¹ numerous studies have aimed to better identify patients with a high risk of early progression, as they may benefit from early treatment intervention. Among the risk stratification models described2, 3, 4, 5, 6, 7 only few identify patients with a greater than 80% risk of progression within 2 years.^5,8 Applying the expression patterns of four genes Khan et al., identified patients with an 85.7% risk of progression at 2 years.⁵ The IMWG group validated the 20/2/20 model of the Mayo clinic⁹ that used >20% BMPCs, >2 g/dl M-protein, FLCratio >20 as cutoffs, and extended the algorithm by incorporating FISH cytogenetics. With this, they could identify a small group with a 2-year progression risk of 88.9%.⁸

Early reports on the impact of single risk factors indicated a high risk of progression in patients with specific markers. The study by Rajkumar et al.,¹⁰ published in 2011, included 655 patients with smoldering MM (SMM); 21 (3.2%) of those patients presented with ≥60% clonal BMPCs, and their median TTP was only 7 months, with a 2-year risk for progression of 95%. Kastritis et al.,¹¹ reported similar results with a median TTP of 15 months. A high risk for progression was also noted in patients with an FLCratio ≥100 by Larsen et al.,¹² in a Mayo patient cohort in 2013. The 2-year risk of progression was 72% and the median TTP was 15 months. Kastritis et al., reported even more alarming results in the same year,¹¹ with a median TTP of 8 months. Another high risk factor for rapid progression was reported by Hillengass et al., in 2010.¹³ These authors noted a median TTP of 13 months in patients with >1 MRI-defined focal lesions (FL) of 5 mm or greater diameter. Two additional reports confirmed the impact of >1 MRI-defined focal lesion on rapid progression to MM with very similar findings: Kastritis et al.,¹⁴ reported a median TTP of 15 months. In the publication by Merz et al.,¹⁵ the exact median TTP was not reported but can be inferred from the graph shown in their publication to a median of approximately 22 months. These studies were published between 2010 and 2014 and prompted the International Myeloma Working Group to update the diagnostic criteria for MM and to include patients meeting the above-cited criteria (plus an involved FLC level of ≥100 mg/dl for patients with a FLCratio ≥100) in the definition of MM.¹⁶ The consequences for those ‘biomarker-defined MM’ or ‘SLiM CRAB positive’ patients are clinically highly relevant, as this classification implies that initiation of therapy is warranted. However, since 2014, a number of relevant studies re-examining this issue were published, which indicate a longer time to progression than initially observed. These findings are supported by clinical experience obtained in patients with high-risk smoldering multiple myeloma, many of which may be followed for years without signs of progression. Therefore, we conducted an extensive literature review and meta-analysis to compare the evidence on the interdependence of these risk factors with progression to MM available at the time of the consensus report (October 2014) with the data published thereafter (until November 1, 2022). We aimed to evaluate whether the median time to progression (TTP) and 2-year progression risk in the high-risk groups would be similar or not, both in absolute and relative (compared to low-risk groups) terms, in more recent studies than in the studies originally available which prompted the inclusion of those patients in the diagnostic criteria of MM.

Methods

For this systematic review with meta-analysis, we conducted a comprehensive literature review following the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines and searched PubMed (https://pubmed.ncbi.nlm.nih.gov/) and EMBASE (https://www.embase.com/) databases from January 1 2010 to November 1 2022. The search terms used are shown in Supplementary File, Table S1. Two authors (HL and SK) conducted independently from each other hand-searching of the citation lists of the selected papers and of abstracts recently presented at EHA. Articles not including Kaplan–Meier-curves on TTP and those using thresholds other than ≥60% BMPCs, FLCratio ≥100, or >1 focal lesion were excluded (Supplementary File, Table S1). In three cases of discrepant assessments the authors discussed the arguments of each other and finally agreed upon inclusion or not.

The quality of the selected papers was assessed using the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies by the National Heart, Lung, and Blood Institute (NIH; https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools),¹⁷ but we excluded question 3, 5, 10, and 13, due to inapplicability (Supplementary File, Table S2). The Quality Assessment Tool was applied by two independent reviewers, and discrepancies in judgement were resolved by discussion. Although the quality of the publications varied (Supplementary File, Table S3), we chose to use all available studies due to the good quality of the graphs and underlying data. The relevance of the research question was discussed during a scientific meeting in Würzburg organized by Prof. Einsele, with Prof. Gareth Morgan and Prof. Martin Kaiser commenting. In addition, the authors had a few informal meetings. This review was not registered beforehand and no review protocol is available.

To obtain the best possible proxy for individual patient data, we digitized (using WebPlotDigitizer™) the published TTP curves of patients enrolled in the respective studies, and then applied the algorithm described by Guyot et al.,¹⁸ in R¹⁹ to obtain individual patient outcomes (Supplementary File, Table S3 for a comparison of the digitized data and published estimates, where available). From these digitized data, we then constructed TTP curves according to Kaplan Meier for the combined data of the SLiM CRAB MM groups available before and after the consensus was published using SPSS Statistics 27. Hazard ratios and log-rank test p-values were calculated to compare curves from earlier publications with those from later publications. We calculated the median TTP as well as the 2-year progression risk for each group, as these are common endpoints reported in clinical studies, and in particular a high risk for progression within two years may be an indicator for considering early treatment. We further conducted a meta-analysis of the odds ratios comparing the 2-year progression risks, using published instead of digitized data wherever available. The meta-analytic procedures were performed using the software ReviewManager (RevMan V. 5.3, The Cochrane Collaboration, 2014): Forest plots (random effects models) were constructed to show the pooled odds ratios for risk of progression at 2 years of patients with high-risk criteria compared to those with criteria not meeting the threshold. Between-study and between-subgroup heterogeneity (I²) was tested.²⁰ Of note, our methodology does not compare individual patient data from different studies directly. Rather, this methodology pools the data stratified by study, i.e. adjusts for study effects. We also did a sensitivity analysis excluding five studies which ranked lowest in the quality assessment. Additionally, we constructed a funnel plot (by plotting the odds ratios against their standard errors for the FLCratio) and did a trim-and-fill analysis (by estimating the true center). These analyses were deemed inappropriate for BMPCs and focal lesions due to the small number of studies.

Role of the funding source

Funding was provided by the Austrian Forum against Cancer, which had neither involvement in study design, collection, analysis and interpretation of data, nor in writing of the report and in the decision to submit the paper for publication. The raw data were accessed by HL, SK and AH.

Results

Our literature search revealed 3159 records by searching databases and 27 abstracts by hand searching (Fig. 1). After eliminating duplicates, publications not addressing biomarker-defined MM or SMM, and publications without graphs suitable for data digitization, we ended up with 11 studies (non-randomized observational cohort studies) that were included in the meta-analysis. In addition to the six studies published up until the end of 2014,10, 11, 12, 13, 14, 15 which formed the basis of the current treatment guidelines, we identified five recent studies21, 22, 23, 24, 25 containing data and diagrams that were viable for the meta-analysis, of which one contained novel data both on patients with increased BMPCs and on high FLCratio,²⁵ four on high FLCratio only,21, 22, 23, 24, 25 and none on focal lesions. The data by Rago et al.,²⁶ on the impact of BMPCs, albeit containing a suitable graph, were not used for the 2014 consensus. We also did not include them in our analysis because of inconsistencies in the publication. Data from individual studies are shown in Fig. 2. In total, the studies included 3482 patients, none of whom was treated for myeloma during the duration that was taken into consideration.

Fig. 1.

Flow-Chart demonstrating study selection for meta-analysis (early = until 2014, later = after 2014). A closer description of the 9 publications without suitable graphs can be found in Appendix 1 pp 1–3.

From: http://www.prisma-statement.org/.

Fig. 2.

Median time to progression and risk for progression at 2 years (+95% CI) for ≥60% vs. <60% BMPCs (A), ≥100 vs. <100 FLCratio (B), and >1 vs. ≤1 MRI defined focal lesions (>5 mm; C), separated by risk group and individual study. (p: published data, only used where estimate plus 95% CI were available. n.r.: not reached).

Fig. 3 shows the Kaplan–Meier curves of TTP to MM in the different patient cohorts. The TTP of patients with ≥60% BMPCs reported in the early studies was significantly shorter than the one for data published in later years (HR: 0.27, 95% CI 0.14–0.51, logrank p < 0.0001); in contrast, a borderline significant weak trend in the opposite direction was recorded in patients with <60% BMPCs (HR: 1.22, 95% CI 1.01–1.47, logrank p = 0.0421). Similar results were found for patients with an FLCratio ≥100 (Fig. 3B): their TTP curve was significantly longer in the later compared to the earlier publications (HR: 0.28, 95% CI 0.21–0.37, logrank p < 0.0001). Importantly, a similar difference was also noted for the two time periods in patients with an FLCratio <100 (HR: 0.56, 95% CI 0.49–0.64, logrank p < 0.0001).

Fig. 3.

Time to progression of data obtained from early vs. recent publications. A: Patients with ≥60% BMPCs vs. <60%, B: Patients with FLCratio ≥100 or <100, and C: Patients with >1 focal lesion vs. ≤1 (no new data published after consensus).

Median TTP and 2-year risk of progression of biomarker-defined patient cohorts positive or negative for the respective criteria are shown in Table 1, comparing data available for the 2014 consensus and data published afterwards, as well as combined results. In patients with ≥60% BMPCs, median TTP increased from 9.2 months in the earlier compared to 30.3 months in later studies; this was associated with a marked decrease in the 2-year risk of progression from 86.2% to 45.5%. Similar trends were observed in patients with an FLCratio ≥100. Median TTP increased from 15.3 to 48.1 months between both time periods, while the 2-year risk of progression decreased from 73.0% to 31.6%. All studies on the impact of >1 focal lesion on time-to-progression were published between 2010 and 2014, and no reports on this topic were published thereafter. Median TTP was 15.1 months and 2-year risk of progression 67.3%. These findings are supported by a sensitivity analysis that excluded five trials that ranked lowest in the quality assessment (Supplementary File, Table S5). Like in the analysis of all studies, a significant prolongation of the TTP and a reduction in the 2-year risk of progression was noted. A funnel plot did not reveal any indication for a publication bias (see Supplementary File, Fig. S1), a finding which was supported by a trim-and-fill analysis (estimated number of missing studies: 0 (SE = 1.51)). The only discenable asymmetry is due to the outlying, small-sample study of Kastritis et al.,¹¹ arguably indicating some publication bias during the early time period, but not supported by the trim-and-fill analysis.

Table 1.

Median TTP and risk for progression @2 years for combined data of the two time periods, and for overall combined data.

Published	Median time to progression (months) [CI]		Risk for progression at 2 years (%) [CI]
	BMPCs ≥ 60%	BMPCs < 60%	BMPCs ≥ 60%	BMPCs < 60%
Early publications	9.20 [6.02–15.56]	101.47 [89.90-NA]	86.21 [65.74–94.45]	21.19 [18.15–24.12]
Recent publications	30.31 [18.71–62.93]	80.46 [70.97–115.48]	45.45 [20.12–62.75]	20.32 [15.18–25.14]
Combined data	15.48 [10.93–21.93]	96.73 [87.01-NA]	68.63 [52.92–79.09]	20.97 [18.37–23.48]
	FLCratio ≥ 100	FLCratio < 100	FLCratio ≥ 100	FLCratio < 100
Early publications	15.33 [9.38–19.10]	58.59 [52.78–65.80]	73.00 [62.39–80.62]	26.58 [22.89–30.09]
Recent publications	48.06 [40.51–64.91]	115.15 [105.96–118.81]	31.61 [25.30–37.39]	16.79 [14.91–18.64]
Combined data	30.40 [25.43–38.69]	93.19 [81.37–105.96]	43.82 [38.14–48.97]	19.45 [17.75–21.12]
	>1 focal lesion	≤1 focal lesion	>1 focal lesion	≤1 focal lesion
Early publications	15.07 [10.49–32.98]	102.42 [69.67–102.42]	67.30 [48.97–79.05]	16.14 [11.10–20.90]

The analysis of the outcome of patients who had lower than 60% BMPCs between both time periods showed a change in the median TTP (101.5 vs. 80.5 months) but no difference for the 2-year risk of progression (21.2 vs. 20.3%). In contrast, in patients with a lower FLCratio (<100), a statistically significantly longer TTP (115.2 vs. 58.6 months) and lower 2-year risk of progression (16.8% vs. 26.6%) was noted in the more recent studies.

Fig. 4 shows the meta-analysis with the odds ratios (OR) for the 2-year progression risk in patients meeting the respective biomarker-defined criteria vs. those not fitting the threshold. The OR for 2-year risk of progression for patients fitting the biomarker-defined criteria for BMPCs vs. those not meeting the threshold was significantly higher in the two earlier (OR: 27.01, 95% CI 4.49–162.34, p = 0.0003) compared to one later report (OR: 3.27, 95% CI 1.37–7.99, p = 0.009). This was substantiated by a test for heterogeneity that showed the two time periods differ significantly in their ORs (I² = 78.6%, p = 0.009). Similar findings were obtained for patients with FLCratio ≥100 vs. those with FLCratio <100 in early (OR: 7.03, 95% CI 4.34–11.37, p < 0.0001) and recent publications (OR: 2.69, 95% CI 1.77–4.09, p < 0.0001), with a heterogeneity test revealing a statistically significant interaction between these time periods and prognostic impact (I² = 67.8%, p = 0.005). The OR for the 2-year risk of progression between patients with >1 and those with ≤1 MRI-defined focal lesions was also highly significant (OR: 10.53, 95% CI 4.96–22.37, p < 0.0001) for the early data. As no recent studies have been published since then, testing against newer data was not possible.

Fig. 4.

Forest plots. Odds ratios for progression risk @ 2 years of A: Patients with ≥60% BMPCs vs. <60%, B: Patients with FLCratio ≥100 or <100 in earlier and more recent studies and C: Patients with >1 focal lesion vs. ≤1 (no new data published after 2014).

Discussion

This analysis revealed a significant heterogeneity between earlier and later studies and an approximately three times longer TTP and 50% lower 2-year progression risk of patients with biomarker-defined MM with ≥60% BMPCs or a FLCratio ≥100 reported in studies published after 2014 as compared to earlier findings. For patients with >1 MRI-defined focal lesion, no comparison between time periods was possible, as no additional data were published after the initial reports dating from 2010 to 2014.

A similar trend as in patients with an FLCratio ≥100 was noted for patients with an FLCratio below the threshold of 100. The TTP and 2-year risk of progression showed a more insidious course in patients enrolled in studies reported after the consensus, indicating a shift in patient characteristics towards more favorable prognostic factors in recent publications. This notion is supported by the meta-analysis, which showed a significant reduction in the odds ratios between patients meeting the thresholds for high-risk and those below the threshold for the respective markers in the later time period (p = 0.009 and p = 0.005 for BMPCs and FLCratio, respectively). The observed improvement of the prognosis of biomarker-defined MM patients in recent compared to early studies is probably due to a phenomenon called stage migration.²⁷ Patients classified today as ‘biomarker-defined’ myeloma have undergone skeletal imaging studies that were not available in the early period. Larsen et al.,¹² for instance, included smoldering myeloma patients who were diagnosed between 1970 and 2010 in their report on the prognostic value of an FLCratio ≥100 (Supplemental File, Table S4). This suggests that a proportion of the 586 patients included in this study might have met the criteria for CRAB-positive myeloma, skewing the progression risk above that of truly positive biomarker-defined myeloma patients with an FLCratio ≥100. A similar scenario can be anticipated for early studies on the impact of high BMPCs on prognosis and possibly for one study¹⁴ investigating the role of MRI-defined focal lesions. The odds ratio for the 2-year progression risk of patients with MRI-defined focal lesions vs. those negative for this criterion was markedly higher in this report (OR: 14.00, CI 2.84–69.08), compared to two other studies^13,15 (OR: 9.71, CI 3.60–26.23, and OR: 9.69, CI 1.81–51.83), which is likely due to underreporting of lesions outside the spine, as only spine- and not whole-body images were available for the analysis.¹⁴

The increasing awareness of monoclonal gammopathies and referal to hematologists likely has led to more frequent work-up, to exclusion of CRAB positive MM, and thus to inclusion of more patients with an insidious course of the disease in recent studies. The uptake of newer diagnostic measures seems to also explain the shift in the control groups with a lower FLCratio (<100) toward a more favorable prognosis in recent reports.

The small number of eligible studies available for the comparison of patients with BMPCs ≥60/<60% with only one recent publication, as well as the small number of studies on focal lesions, is a limitation of this analysis. Another important limitation of this meta-analysis was the unavailability of accurate censored data for digitized individual patient data. Even though estimates for median TTP and 2-year-progression risks obtained from the digitized data are very similar to the published values (see Supplementary File, Table S4), the exact number of progressions or censorings, as well as their timing, could not be accurately obtained from the published figures, inter alia due to a lack of censoring ticks. Nevertheless, our estimates of median TTP and 2-year-progression risk should be quite accurate (assuming a low proportion of early censorings), while confidence intervals, significance tests and hazard ratios may deviate slightly from the original data. We must also mention that some studies used in this meta-analysis received a low quality ranking according to the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (Supplementary File, Table S3), although we have confidence in the quality of the graphs and underlying data, which is supported by a sensitivity analysis excluding these low-scoring studies (Supplementary File, Table S5). A further limitation is the lack of novel data on >1 MRI-defined focal lesion. New studies could reveal whether a similar shift toward improved prognosis could be observed. This is not completely unlikely with the use of new imaging tools such as diffusion-weighted MRI or MRI in combination with PET/CT. Furthermore, it should be mentioned that the exact definition of biomarker-defined MM patients with an FLCratio ≥100 includes a minimal involved free light chain concentration of ≥100 mg/dl. This criterion was taken into account in five of the seven published studies. Two studies did not employ the latter criterion,^11,25 and a third study did not use it for their diagram,¹² a fact which might have led to underestimation of the risk of their ‘marker positive’ patients. Lastly, this study has not been registered in PROSPERO, an international prospective register of systematic review protocols in health and social care. This registry informs investigators about planned and ongoing systematic reviews thereby avoiding possible doublication of efforts. Searching PROSPERO did not reveal any projects with competing design and aims.

Our findings highlight the importance of using modern skeletal imaging techniques in all patients with ‘biomarker-defined MM’ to exclude patients with myeloma bone involvement. Furthermore, including additional risk factors for biomarker-defined MM patients would improve the identification of patients with a high probability of rapid progression to active myeloma and could facilitate treatment decisions in these patients, an opinion which is shared by Akhlaghi et al.²¹ Combining the existing criteria with risk factors such as cytogenetic factors,⁴ light chain proteinuria,²⁴ gene expression data,^5,28 circulating plasma cells,²⁹ whole exome³⁰ or genome,³¹ and immune³² profiling, among others, should increase the prognostic accuracy of biomarker-defined MM. Very recent studies suggest to adapt the definition of patients with a high FLCratio and to include the detection of concomittant proteinuria (>200 mg/24 h) as an essential criterion for high-risk in those patients.^21,24,33 Taking up this suggestion would imply that the IMWG consensus needs to be adapted also in this respect. Furthermore, an MGUS phenotype has been shown to signal a very low risk for progression³⁴ and identification of these patients should spare them the burden of anti-myeloma therapy. A reasonable approach to better characterize the aggressiveness of the underlying disease is dynamic follow-up. Evolving patterns of certain markers of disease, such as an increasing M-component and BMPC infiltration or decreasing hemoglobin levels,⁷ may be used together with static data.

In summary, we found a roughly threefold longer TTP and an about 50% lower 2-year progression risk in patients with biomarker-defined MM and ≥60% BMPCs or a FLCratio ≥100 in recent studies compared to those observed in earlier publications. Our findings have several implications. First, patients with an FLCratio ≥100 and an involved FLC level ≥100 mg/dl should not automatically be considered as early myelomas, as about half of properly staged patients will remain progression-free for roughly four years. Second, more recent data on patients with high BMPCs (≥60%) showed a much lower 2-year progression risk (45.5%) than the early studies (86.21%). Therefore, a more conservative approach should be applied when considering treatment initiation. Third, the natural history of truly smoldering and biomarker-defined MM should be prospectively studied using modern diagnostic criteria as envisaged in the CARRISMM study,³⁵ and therapeutic studies in these patients should include a carefully followed untreated control group with one of the endpoints being overall survival.^32,36

Patients with biomarker defined MM should be carefully followed up by skeletal imaging (whole body MRI, or diffusion-weighted MRI, or PET/CT, or skeletal CT) and laboratory tests at reasonable intervals. In case of an evolving pattern, including the M-component, hemoglobin, GFR, and skeletal findings, initiation of therapy should be considered. Finally, it is not the disease itself which has changed its behavior; it is rather the advancement in myeloma diagnostics, better access to care, and care by myeloma specialists that allows us to better define the prognosis of patients with smoldering myeloma and of those with biomarker-defined MM. Clearly, further advances in prognostication are needed to finetune management of individual patients.

Contributors

HL designed the study and prepared the first draft of the manuscript. SK and AH conducted the statistical analyses. MS and NZ contributed to writing the manuscript and interpreting the results. All authors participated in the discussion of the available evidence and main findings of the study. All authors commented on the first and subsequent drafts of the manuscript.

Data sharing statement

The data presented in this article will not be made available to others.

Declaration of interests

The authors declare no relevant conflicts of interest.