STRENGTHS AND LIMITATIONS OF THIS STUDY
-
We performed a meta-analysis of 11 studies covering 2704 patients to discuss the prognostic value of clonal plasma cells in newly diagnosed symptomatic multiple myeloma.
-
Literature search, study inclusion, data collection, quality assessment, statistical analysis and bias analysis were conducted in strict accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.
-
Our eligible data were only from published articles and there may be unpublished data we did not obtain.
Introduction
Multiple myeloma (MM) is a malignant plasma cell disorder with wide variation in clinical progression and prognosis. Despite the dramatic improvement of patients’ survival in the past decades with the incorporation of novel agents and autologous stem cell transplantation (ASCT) into therapeutic strategies, MM remains incurable,1 with the survival time from only a few months to more than 10 years.2 With the availability of different classes of approved agents, the choice of the optimal strategy at diagnosis represents a challenge. Early recognition of the high-risk cohort is important for individualised therapy.
The definition of a high-risk group is continuously evolving and remains controversial. The most widely accepted staging system is based on cytogenetic and clinical biomarkers according to the revised International Staging System (R-ISS)3 and Mayo Stratification of Myeloma and Risk-Adapted Therapy. In addition to classical prognostic factors, the molecular abnormalities and the clonal plasma cells (CPCs) are being identified as possible prognostic factors. Unfortunately, next-generation sequencing is expensive and time-consuming. The CPCs can be detected through various methods, such as routine peripheral blood smear by conventional microscopy,4 slide-based immunofluorescence assay,5 multiparametric flow cytometry (FCM), next-generation FCM (NG-FCM) and next-generation sequence6 and so on. FCM is now a widely used tool that can detect and quantify the absolute number of CPCs in a more sensitive, minimally invasive and reproducible manner than others.
The presence of CPCs is a marker for a highly proliferative and fulminant process. The prognostic value of quantitating CPCs using FCM in symptomatic MM patients remains unclear. The optimum cut-off value of CPC positive (CPCs+) to predict survival is also kept controversial. It has been reported that CPCs could predict high risk of progression to symptomatic MM in smouldering MM (SMM)7 8 and predict poor prognosis in newly diagnosed symptomatic MM (NDMM)9 and relapsed MM.10 Once the CPC count exceeding 20% or an absolute plasma cell count of >2000/mm evaluated by morphology, it is defined as plasma cell leukaemia (PCL). Although there is considerable overlap between MM and PCL with respect to clinical, immunophenotypic and cytogenetic features, PCL is more aggressive and has poorer survival.11 In 2021, the International Myeloma Working Group (IMWG) recommended that PCL should be redefined as the presence of ≥5% CPCs evaluated by morphology,12 for NDMM with 5%–19% CPCs showed similar poor survival with PCL.13 14 There was even a study proposed to define 2% of CPCs by using sensitive multiparameter FCM as a new cut-off for ultra-high-risk MM resembling behaviour of primary PCL.15 The recognition of the prognostic significance of CPCs in MM patients is increasingly acknowledged, alongside the growing appreciation for the sensitivity and minimal invasiveness of FCM.
Although a prior meta-analysis explored the prognostic value of CPCs in MM and primarily concentrated on detection methods and sample times,16 our meta-analysis is specifically conducted to evaluate the prognostic value of quantifying CPCs by FCM in symptomatic NDMM patients and to explore whether the ASCT could eliminate the adverse impact on survival.
Materials and methods
Literature search and search strategy
A comprehensive literature search was conducted across three main databases—PubMed, EMBASE (OVID) and Web of Science—without any restrictions, up to 27 August 2022. Meanwhile, we screened the references of all articles included in our studies. The complete search strategies can be found in the online supplemental data. The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (online supplemental data: PRISMA_2020_checklist).17 The review protocol has been registered in the PROSPERO International Prospective Register of Systematic Reviews (registration number: CRD42021272381).
Supplemental material
Supplemental material
Selection criteria
Studies were included in our meta-analysis if they met the following criteria: (1) the article was published up to 27 August 2022 focused on evaluating the prognostic value of CPCs in symptomatic NDMM patients; (2) the study offered the information on overall survival (OS) and/or progression-free survival (PFS) of patients; (3) the study was published as an original article in English; (4) FCM was used to detect and quantify the absolute number of CPCs. Exclusion criteria comprised case reports, meeting abstracts, review articles, letters, laboratory studies and articles with small sample size (<10 patients). For the overlapping data across multiple studies, only the latest or the highest quality study was incorporated into our meta-analysis.
Patient and public involvement
No patient and/or the public were involved.
Data extraction
Titles and/or abstracts of studies were independently screened by two investigators (XL and FW) to identify studies potentially meeting the inclusion criteria mentioned above. Subsequently, the two reviewers independently assessed and extracted relevant data from the full texts of these potentially eligible studies.
The extracted data from articles included the first author’s name, study timeline, publication year, country, total numbers and CPCs+numbers of patients, age and gender distribution of patients, disease stage during sample collection, detection method, cut-off value of CPCs+, staging systems, cytogenetics analysis, median follow-up time, initial treatment strategy, whether ASCT was performed or not, OS and PFS.
The primary endpoint was OS, measured from the day of diagnosis to death from any cause or the last follow-up. The secondary endpoint was PFS, measured from the day of diagnosis to disease progression, death or the last follow-up. HR and 95% CI for OS and PFS from COX multivariable models were used to calculate the pooled estimated effect, evaluating the prognostic value of CPCs in patients with MM. In instances where necessary data for analysis were not provided in published articles, we obtained the information from survival curves using the method published.18
Divergences in data extraction were resolved through discussion. If necessary, a third investigator (QY) was involved to address disagreements until a consensus was reached.
Quality assessment
Two investigators (XL and FW) independently assessed the methodological quality of each eligible study with the Newcastle-Ottawa Quality Assessment Scale (NOS) for cohort studies, which covers the following nine items: selection including four items, comparability including two items and outcome including three items.19 Cohort studies with six or more stars were considered as high quality. Discrepancies between two investigators were resolved by discussion.
Sensitivity analysis and publication biases
Sensitivity analysis was used to assess the effect of each included study on the stability of the pooled results by sequentially omitting one study each time. We identified publication biases of included studies through Begg’s test20 and Egger’s test.21 All the relevant calculations were performed using Stata V.12.0 software, and a significance level of p<0.05 was deemed statistically significant.
Statistical analysis
Stata V.12 software (Stata Corp, College Station, Texas, USA) was used to perform all the statistical analyses. The software script/code was offered in the (online supplemental appendix B). CPCs+ indicated a poor prognosis when the pooled HRs for OS or PFS were >1 and their 95% CIs did not include 1. Conversely, CPCs+ indicated a good prognosis when the pooled HRs for OS or PFS were <1, and their 95% CIs did not include 1. CPCs+ had no prognostic value if there 95% CIs include 1. The pooled HRs and their 95% CIs for OS and PFS were calculated using the random effects model with the inverse variance method. The heterogeneity of the included studies was evaluated by χ2 test with a significance level when p<0.05. The I2 was used for the quantitative assessment of heterogeneity with four classifications: I2=0%–25%: no observed heterogeneity; I2=25%–50%: moderate heterogeneity; I2=50%–75%: large heterogeneity; I2=75%–100%: extreme heterogeneity.22 Meta-regression was used to explore the source of heterogeneity among studies if the included study number is more than 10, otherwise the subgroup analysis was applied. Publication biases of included studies were identified through funnel plot (the included study number is more than 10), Begg’s test and Egger’s test.
Results
Study identification and selection
As shown in figure 1, there were 1295 studies through the initial database searching. In addition, there were seven studies by screening the references. A total of 928 studies were further screened in titles and/or abstracts after excluding the 374 duplications. Then 55 eligible studies were needed to full text assessment while the 873 irrelevant and improper types of articles were excluded. We further excluded 44 studies: 1 written in Chinese, 24 lacking survival data or having insufficient data, 10 studies in which the CPCs were not detected by FCM, 1 letter, 3 with overlapping data and 5 for non-NDMM patients. Ultimately, 11 studies were included in our meta-analysis covering 2704 cohort patients.
Flow chart of study selection. FCM, flow cytometry; NDMM, newly diagnosed symptomatic multiple myeloma.
Characteristic of selected studies
As presented in the online supplemental table 1, the 11 included articles were all retrospective cohort studies spanning the publication years 2005 to 2022, with a study timeline ranging from 1998 to 2021. Notably, all these articles were single-centre studies, with three studies conducted in the USA, four in Asia (China, Japan and Korea) and four in Europe (Italy and Spain). The identification and quantification of CPCs were achieved through FCM or NG-FCM, employing diverse antibody panels featuring a spectrum of colours, ranging from two (CD38 and CD45) to eight (CD38, CD56, CD19, CD138, CD45, CD27, Kappa and Lambda). There did exist variability in the cut-off value for CPC positivity across the 11 cohort studies, with a cumulative inclusion of 1509 CPCs+ patients and 1195 CPCs− patients. Noteworthy is the observation that the majority of patients in all studies, except for two, underwent initial treatment with novel agents. Furthermore, all patients in two studies underwent ASCT.
Quality assessment of included studies
The NOS was used to assess the quality of the 11 included studies in our meta-analysis. The average overall score of the 11 eligible studies was 8 (range: 6–9), indicating a high level of quality in the included research (see online supplemental table 2).
Clinical characteristics of MM patients with CPCs+
As shown in online supplemental table 3, there were no significant differences in age and sex between the CPCs+ group and the CPCs− group.23–27 The CPCs+ group exhibited a positive association with higher levels of β2-microglobulin (β2-MG) and lactic dehydrogenase (LDH) level,23 24 28 29 as well as the presence of high-risk cytogenetic abnormalities.23 28 29
The studies showed a correlation between an elevated level of CPCs and an advanced ISS stage,10 24 25 28 29 equally to the R-ISS stage as well.23–25 28 Furthermore, the CPCs+ group exhibited a lower treatment response rate than CPCs− group.24 30 31
Analysis of outcomes
The prognostic data of OS with CPCs was available in 10 cohort studies (see figure 2A and online supplemental table 4). The pooled HR for OS was used to evaluate the prognosis value of CPCs+ in patients with NDMM. The analysis indicated that CPCs+ served as an adverse prognostic factor for OS, with an HR of 1.95 (95% CI 1.24 to 3.07) and exhibited extreme heterogeneity (I2=83.8%, p=0.000) in the random effects model.
(A) Forest plots of the HRs for the overall survival in the total population in a random effects model. (B) Forest plots of the HRs for the progression-free survival in the total population in a random effects model. CPCs−, clonal plasma cells negative; CPCs+, clonal plasma cells positive.
The prognostic data of PFS with CPCs+ was available in 10 cohort studies (see figure 2B and online supplemental table 4). Analysis revealed that CPCs+ was a poor prognostic factor for PFS, with an HR of 2.07 (95% CI 1.79 to 2.39). Interestingly, no significant heterogeneity was detected within the total population (I2=0, p=0.459) in the random effects model.
Among the 11 studies included in the analysis, 2 studies were focused on transplant-eligible NDMM patients. The pooled HR for OS from these two studies was 1.89 (95% CI 1.01 to 3.53, p=0.109), and for PFS, it was 2.07 (95% CI 1.58 to 2.72, p=0.878). Both these results revealed an unfavourable prognosis influence of CPCs+ in transplant-eligible NDMM patients. Notably, there was substantial heterogeneity in the OS analysis (I2=61%), while no observed heterogeneity was noted in the PFS analysis (I2=0), as depicted in figure 3. The results indicated ASCT may not eliminate the adverse impact of CPCs+ on OS and PFS in this specific patient population.
(A) Forest plots of the HRs for the overall survival in transplant-eligible population in a random effects model. (B) Forest plots of the HRs for the progression-free survival in transplant-eligible population in a random effects model. CPCs−, clonal plasma cells negative; CPCs+, clonal plasma cells positive.
Owing to the extreme heterogeneity in HR for OS in the total population, we carried out a sensitivity analysis to assess the influence of individual studies on the pooled HRs for OS. Our results revealed that the study by Vasco-Mogorrón in 202132 had no evident effect on the combined HR for the OS of the total population in the random effects model. However, it seemed to have an impact on the fixed effects model, as illustrated in online supplemental figure 1. The result above indicated that the study might be a potential source of the heterogeneity. On excluding this study, the pooled HR increased from 1.95 (95% CI 1.14 to 3.53) to 2.2 (95% CI 1.24 to 3.07). Simultaneously, the I2 decreased from 83.8% (p=0.000) to 46.4% (p=0.059) (see figure 2 and figure 4).
Forest plots of the HRs for the overall survival after one article (Vasco-Mogorrón 2021)32 excluded in a random effects model. CPCs−, clonal plasma cells negative; CPCs+, clonal plasma cells positive.
We performed subgroup analysis rather than meta-regression to explore the possible source of heterogeneity in OS, given the limited availability of OS data from only 10 studies. When the population were divided into novel agents initial treatment group and novel agents combined with traditional chemotherapy initial treatment group, we observed moderate heterogeneity (I2 :46.6%, p=0.059, figure 5A). Our findings indicated the initial treatment maybe a source of the heterogeneity.
Forest plots of the HRs for the OS of subgroup analysis. (A) Subgroup analysis by initial treatment. (B) Subgroup analysis by FCM colour. CPCs−, clonal plasma cells negative; CPCs+, clonal plasma cells positive; FCM, flow cytometry; NG-FCM, next-generation flow cytometry.
We also found articles employing 5–8 colour FCM had lower heterogeneity compared with those using 2–3 colour FCM and NG FCM (I2: 33.7% vs 60.7% vs 61.0%, figure 5B). This suggested that the choice of FCM panel maybe a source that influence the heterogeneity.
Publication biases
Significant publication bias was detected in Begg’s test (p=0.049), but not in Egger’s test (p=0.830) for OS in the entire population of the 10 studies (online supplemental figure 2A, B). Via the trim and fill method, three missing studies were filled, and the logHR and its 95% CI were 0.669 (0.216 to 1.122) and 1.539 (1.010 to 2.347) before and after in the random effects model, respectively. The funnel plot showed that the points were essentially symmetric and nearly around the curve (online supplemental figure 2C). There was no publication bias in this meta-analysis, indicating the stable and credible results.
Discussion and conclusion
In 1993, it was first reported that detection of CPCs was a marker of disease activity in patients with plasma cell disorders and established an appropriate cut-off value of 3×106 /L or higher CPCs by slide-based immunofluorescence.5 Subsequently, an increasing number of studies have reported the prognostic significance of CPCs level in MM with the detection methods varied,4–6 either in NDMM or pretransplantation MM. Notably, FCM is a relatively more sensitive and economical tool than microscopy in detecting and quantifying the absolute number of CPCs.33
Our meta-analysis demonstrated that CPCs+ was a poor prognostic factor for OS (HR=1.95, 95% CI 1.24 to 3.07, I2=83.8%, p=0.000) and PFS (HR=2.07, 95% CI 1.79 to 2.39, I2=0, p=0.459) in NDMM patients compared with those with CPCs−. The ASCT could not eliminate the adverse impact of CPCs+ on OS (HR=1.89, 95% CI 1.01 to 3.53, p=0.109) and PFS (HR=2.07, 95% CI 1.58 to 2.72, p=0.878). Despite here was no significant difference between CPCs+ and CPCs− NDMM patients regarding sex and age, notable distinctions were observed in other clinical parameters. The CPCs+ patients always showed higher β2-MG and LDH level. Moreover, the CPCs+ NDMM patients were usually accompanied by high-risk cytogenetic abnormalities.
There was extreme heterogeneity in entire NDMM patients for OS (I2=83.8%), even when we excluded one article for its survival data was not provided directly (I2=46.6%). In order to find the possible source of the heterogeneity, we further performed the subgroup analysis. Given the limited number of articles included, a meta-regression analysis was not performed. The results indicated that the initial treatment might be a significant contributor to the observed heterogeneity. The heterogeneity decreased significantly in studies using 5–8 colour FCM, indicating that the FCM panel selection had an important influence on the heterogeneity. We carried out a sensitivity analysis to assess the influence of individual studies on the pooled HRs for OS. Our results revealed that excluding a possible heterogeneity study (Vasco-Mogorrón 2021) resulted in the pooled HR increasing from 1.95 (95% CI 1.14 to 3.53) to 2.2 (95% CI 1.24 to 3.07), meanwhile, the degree of heterogeneity decreasing from extremely high (I2=83.8%, p=0.000) to large (I2=46.4%, p=0.059). Importantly, no significant publication bias was detected for OS in our meta-analysis. Additionally, we analysed the potential heterogeneity study (Nowakowski et al9) and found that patients in this study underwent traditional chemotherapy as initial treatment.
Although our paper is the first meta-analysis regarding the prognosis value of quantitating CPCs using FCM in NDMM, several limitations existing in our meta-analysis should not be neglected. First, our meta-analysis was predominantly based on retrospective studies rather than randomised controlled trials. Second, except for the potential heterogeneity from two studies (Vasco-Mogorrón 2021,32 Nowakowski G. S 20059) there were still substantial clinical heterogeneity among studies, such as age and gender distribution of patients, cytogenetic abnormalities, staging system, cut-off values of CPCs+, follow-up durations and treatment regiments, which may influence the clinical outcomes. Third, the studies included in our analysis employed diverse antibody panels for CPC detection, ranging from 2-colour to 8-colour panels, additionally, two studies used NG-FCM for CPCs detection. Fourth, it is important to acknowledge other potential publication biases could exist in our meta-analysis, since our eligible data were exclusively sourced from published articles, and there may be unpublished data we did not obtain. Finally, we did not explore the relationship between CPCs levels and the cytogenetic and molecular abnormalities in NDMM patients owning to a limited number of studies exploring this relationship, which needs to be further studied. Vagnoni et al reported that CPCs levels had a prognosis impact on patients with standard-risk cytogenetics rather than those with high-risk cytogenetics.27 Recent reports speculated that the primary reason for the existence of a high number of CPCs is cytogenetic abnormalities, particularly t(4;14) abnormality.34
The prognostic factors and the staging systems are evolving with the development of detection methods and treatment strategies. The inclusion of CPCs into the R-ISS holds potential for identifying a subgroup at much higher risk of adverse outcomes, indicative of more aggressive disease in the near future. Han et al reported that the prognosis of patients with a high level of CPCs was significantly inferior to those patients with low level of CPCs in the bortezomib group.23 NDMM with 5%–19% CPCs displayed similar survival compared with PCL, and the IMWG had recommended that PCL should be redefined as the presence of ≥5% CPCs in patients meeting diagnostic criteria for MM.12–14 Although MM patients had achieved very good partial response or better at the time of peripheral blood stem cell collection due to the use of novel agents, the CPCs were still detectable in 18.8% (27/144) patients.30 Those studies implied that specific treatment for patients with high level of CPCs were needed. In summary, our meta-analysis including 11 studies demonstrated that CPCs+ was an independent poor prognostic factor for OS and PFS in NDMM. Notably, the ASCT could not eliminate the adverse impact of survival. Further studies are needed to explore whether the level of CPCs can be used as a predictor of treatment response.
Data availability statement
All data relevant to the study are included in the article or uploaded as online supplemental information.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
This post was originally published on https://bmjopen.bmj.com