Introduction

Multiple myeloma (MM) is a plasma cell neoplasm characterized by the presence of monoclonal immunoglobulin in serum and/or urine and clinical symptoms related to the end organ damage by MM comprising hypercalcemia, renal insufficiency, anemia, and bone complications (CRAB symptoms) [1]. Recently, the International Myeloma Working Group (IMWG) has updated the diagnostic criteria to address a change in the diagnosis of a subset of patients without CRAB symptoms but with biomarkers of malignancy (called SLiM, derived from an acronym of the biomarkers) from smoldering MM (SMM) to MM requiring treatment [2]. As such, MM is a heterogeneous disease and survival outcome varies considerably depending on the risk status of each patient, determined by patient-related factors such as age and frailty, disease-related factors such as clinical stage and cytogenetic abnormalities (CA), and treatment-related factors such as the use of novel agents and autologous stem cell transplantation (ASCT).

Survival of patients with MM has shown a continued improvement over the past decade by the emerging novel therapies comprising initially bortezomib and thalidomide, followed by next-generation proteasome inhibitors, carfilzomib and ixazomib, and next-generation immunomodulatory drugs (IMiDs), lenalidomide and pomalidomide, as well as other new class of agents such as elotuzumab and daratumumab classified as monoclonal antibodies [3]. However, MM patients are not always treated with the novel therapies because of comorbidities and/or socio-economical reasons [4].

To evaluate tumor burden of MM, the IMWG initially developed the International Staging System (ISS) based on the serum levels of albumin and β2-microglobulin in 2005 [5]. Subsequently, several investigators have reported the importance of LDH and CA in the prognostication of MM [6, 7]. Recently, the IMWG has proposed the Revised ISS (R-ISS) as a more potent prognostic system than the original ISS by incorporating both LDH as a function of tumor burden and the presence of high-risk CA as a function of disease biology into the ISS [8]. Although the R-ISS was constructed based on the data obtained solely from patients enrolled in clinical trials testing novel agents, its validity has also been evaluated in patients of routine clinical practice in many countries including North America, Europe, and Asia [9,10,11]. However, there have been no studies evaluating the relevance of the R-ISS in the context of different treatment modalities in clinical practice.

In the present study, we have evaluated and compared the clinical relevance of the R-ISS with the original ISS in Japanese MM patients treated with different modalities in routine clinical practice.

Patients and methods

Patients

The Japanese Society of Myeloma has conducted two pivotal retrospective studies. The first study was carried out based on the collected data from 1383 patients diagnosed and treated in clinical practice between January 1990 and December 2000, before the era of novel agents, to disclose the real figure of the management of MM at that time in Japan [12]. The second study was to analyze changes in outcome by comparing the data of the previous study with more recent data involving 2234 patients diagnosed and treated between January 2001 and December 2012 [13]. Subsequently, the second study was further extended to a total of 3270 patients from 38 affiliated hospitals. Among these, a sufficient data set including FISH (fluorescence in situ hybridization) analysis and details of initial treatment was obtained in 718 patients and served as the basis of the present study.

The diagnosis of MM was made according to the IMWG criteria [1] and the clinical stage of MM was determined based on the Durie and Salmon staging system [14]. Patients with asymptomatic (smoldering) MM and primary amyloidosis were excluded. Baseline demographic, clinical, and laboratory data, and details regarding treatment and response to treatment were collected retrospectively during the period between January 2014 and May 2014. This study has involved unselected patients treated consecutively in routine clinical practice in the participating hospitals of the Japanese Society of Myeloma. This study was conducted in accordance with the institutional guidelines with approval of the Ethics Committee/Institutional Review Board of Tokushima Prefectural Central Hospital.

Stages according to the ISS and R-ISS

The ISS and the R-ISS stages in each patient were determined according to the respective IMWG criteria previously published [5, 8]. Abnormal LDH was defined as a serum level greater than the upper limit of normal. CA was analyzed on the CD138 selected bone marrow plasma cells by FISH method. The cut-off level of ≥ 10% was used to determine the presence of CA in each FISH analysis. High-risk CA was defined as the presence of any of t(4;14), t(14;16), or del(17p) according to the R-ISS criteria [8].

Treatment

Treatment for each patient was determined by the respective physician-in-charge. The short-term use of dexamethasone for the emergency purpose was not considered as a therapeutic line. Initial therapy was classified based on the use of novel agents incorporating thalidomide, lenalidomide, and bortezomib or the use of ASCT. ASCT was conducted with high- or intermediate-dose melphalan conditioning followed by peripheral blood stem cell transplantation according to the institutional protocol and was regarded as initial therapy only when it was performed upfront after induction therapy. Treatment response was assessed according to the uniform response criteria by the IMWG [15].

Statistical analysis

Fisher’s exact test was used to compare differences between categorical variables, whereas the Mann-Whitney U test was used for continuous or nominal values. Kaplan-Meier method was used to create the overall survival (OS) curves, and differences between the curves were analyzed by the log-rank test and the Holm test for multiple comparisons when applicable. Statistical analyses were performed with EZR version 1.30 (Saitama Medical Center, and Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [16].

Results

Patient characteristics

In the present study, we collected the clinical data of 3270 patients diagnosed and treated in the period of between 2001 and 2012 (whole group). A sufficient data set including FISH analysis and the information of initial therapy was evaluable in 718 patients of them and was subjected to the analysis (study group). The follow-up periods of these patients ranged from 0.4 to 152.8 months (median, 29.5 months).

Baseline characteristics of the patients are summarized in supplementary Table 1. There were 374 males and 344 females. The median age was 67 years old (range, 31–89). The isotype of M protein was IgG in 433 patients (60.2%), IgA in 138 (19.2%), IgD in 17 (2.4%), and Bence Jones type in 111 (15.5%). The Durie and Salmon stages I, II, and III were distributed in 9.0%, 26.2%, and 64.8% of the patients, respectively. The ISS stages I, II, and III were distributed in 31.8%, 35.7%, and 32.5%, respectively. Patient baseline features including age, gender, M protein isotype, Durie and Salmon stage, and ISS stage were not significantly different between the whole patient group and the study patient group. The percentage of patients with high β2-microglobulin levels was similar in both groups (51.1% vs 54.7%). The percentage of patients with abnormal serum LDH (>normal upper limit) was significantly lower in the study patient group (19.4% vs 26.5%, p < 0.0001). Among the 718 patients, those with adverse CA such as t(4;14), t(14;16), and del(17p) were observed in 142 (19.8%), 14 (1.9%), and 63 (8.8%) patients, respectively. The percentage of patients with abnormal karyotype was slightly higher in the study patient group (33.3% vs 28.2%, p = 0.0219). As for the initial therapy, the proportion of patients treated with conventional chemotherapy alone was lower (33.1% vs 48.7%), whereas that of the patients treated with novel therapy alone was higher in the study patient group than in the whole patient group (36.3% vs 18.7%). ASCT was performed in 30.6% of the study patient group and in 32.6% and the whole patient group, respectively.

Initial therapy

Treatment regimens of the initial therapy for the 718 study patients are shown in supplementary Table 2. Two hundred and thirty-eight patients (33.1%) received conventional chemotherapy alone including melphalan + prednisolone (MP) reaching to 21.1%, followed by vincristine + doxorubicin + dexamethasone (VAD) and high-dose dexamethasone. Two hundred and sixty patients (36.3%) received novel agent-based regimens alone including bortezomib + dexamethasone ± cyclophosphamide (BD/VCD), bortezomib + MP (VMP), and thalidomide + MP (MPT). A total of 220 patients underwent upfront ASCT either after the induction therapy with conventional chemotherapy alone represented by VAD and high-dose dexamethasone in 69 (9.6%) patients or after the induction with novel agents including BD, VCD, and bortezomib + doxorubicin + dexamethasone (PAD) in another 151 patients (21.0%). Maintenance therapy was given to 125 transplanted patients and to 240 non-transplanted patients in different doses and schedules.

Distribution of the ISS and R-ISS stages

The distribution of patients according to the ISS and the R-ISS stages in the study patients is shown in Table 1 and comprises 228 patients (31.8%) assigned to the ISS stage I, 257 (35.7%) to the stage II, and remaining 233 (32.5%) to the stage III, whereas 154 patients (21.4%) were assigned to the R-ISS stage I, 438 (61.1%) to the stage II, and remaining 126 (17.5%) to the stage III, respectively. It was noted that the distribution of the 718 patients according to the R-ISS stages showed a higher predilection for the stage II although the same 718 study patients were equally distributed by the ISS stages.

Table 1 Distribution and median OS of patients according to the ISS and the Revised ISS stages
Full size table

Best response according to the ISS and R-ISS stages

Best response according to the ISS and R-ISS is shown in Table 2. The overall response rates (≥partial response, PR) were as follows: 71.5% for the ISS stage I, 64.6% for the ISS stage II, and 61.2% for the ISS stage III (p = 0.14); and 70.4% for the R-ISS stage I, 64.6% for the R-ISS stage II, and 61.4% for the R-ISS stage III (p = 0.45), respectively. The rate of deep response such as stringent complete response (sCR) and CR differed by the stages; 18.7% for the ISS stage I, 11.8% for the ISS stage II, and 7.6% for the ISS stage III (p = 0.01); and 18.4% for the R-ISS stage I, 10.9% for the R-ISS stage II, and 10.5% for the R-ISS stage III (p = 0.12), respectively. Thus, the percentage of patients achieving PR or better was similar between the ISS stages and the R-ISS stages. Although the overall response rate and deep response rate were higher in stage I than in other stages by the ISS and R-ISS, it seemed difficult to predict therapeutic responses with these classifications.

Table 2 Best response according to the ISS and R-ISS stages
Full size table

Best response according to initial therapy

In contrast to the similar distribution of patients achieving best response according to the ISS and R-ISS stages, the distribution of patients achieving best response differed considerably according to the initial therapy categories as shown in Table 3. The overall response rate (≥PR) was 41.6% for the conventional chemotherapy group, 68.8% for the novel agent group, 76.1% for the conventional chemotherapy + ASCT group, and 91.7% for the novel agents + ASCT group, respectively (p = 1.07 × 10−17). The deep response rate (sCR and CR) according to initial therapies was 1.8%, 12.3%, 4.5%, and 34.0%, respectively, and was the highest in the novel agent + ASCT group (p = 5.1 × 10−15).

Table 3 Best response according to initial therapy
Full size table

OS according to the ISS and R-ISS stages by treatment modalities

We then analyzed OS of the study patients according to the ISS and R-ISS stages. Figure 1 and Fig. 2 show the survival curves according to respective ISS stages and R-ISS stages. In both figures, survival curves of total patient population are shown in A, those of patients treated without novel agents in B, with novel agents in C, those treated without ASCT in D, and with ASCT in E. The difference between stages was evaluated by the p value of three-stage comparison by the log-rank test. The p values between each stage were analyzed by the Holm test as shown in supplementary Table 3.

Fig. 1
figure 1

Overall survival according to the ISS. OS in total patients (a); patients treated with conventional chemotherapy ± ASCT (b) or with novel agents ± ASCT (c); patients treated without upfront ASCT (d) or with ASCT (e) as initial therapy. NR, not reached

Full size image
Fig. 2
figure 2

Overall survival according to the R-ISS. OS in total patients (a), patients treated with conventional chemotherapy ± ASCT (b) or with novel agents ± ASCT (c); patients treated without upfront ASCT (d) or with ASCT (e) as initial therapy. NR, not reached

Full size image

The median OS of the study patients according to the ISS was 100.7 months for stage I, 65.2 months for stage II, and 50.9 months for stage III, respectively (p = 4.0 × 10−10, Fig. 1a). In the analysis evaluating the effect of the use of novel agents as initial therapy, the difference in OS was more pronounced in the novel agent group (p = 4.1 × 10−6, Fig. 1c) compared with the conventional chemotherapy group (p = 4.8 × 10−4, Fig. 1b). In the comparison according to the use of transplantation, survival curves between the stages were clearly separated in the non-transplant group (p = 1.4 × 10−8, Fig. 1d), but not in the transplant group (p = 0.13, Fig. 1e).

According to the R-ISS, the corresponding median OS for each stage was 152.8, 65.2, and 45.3 months, respectively (p = 9.0 × 10−15, Fig. 2a). The difference in survival curves between the stages was significant whether the initial therapy was novel agents-based (p = 3.4 × 10−7, Fig. 2c) or conventional chemotherapy-based (p = 2.3 × 10−7, Fig. 2b). When comparing the survival curves of patients treated with ASCT, the R-ISS clearly distinguished the difference between the stages in the non-transplant group (p = 2.4 × 10−12, Fig. 2d). Although the R-ISS could show a superb survival of the stage I patients treated with upfront ASCT, the survival curve for stage II overlapped with that of stage III (p = 0.033, Fig. 2e).

The difference between survival curves according to each stage was more significant in the R-ISS than in the ISS (supplementary Table 3). However, there was a tendency that the difference between stages II and III was smaller than that between stages I and II, especially in patients treated with novel agents (Figs. 1c and 2c) and those with transplantation (Figs. 1e and 2e).

Comparison of the OS within the R-ISS stages

When comparing R-ISS with ISS, more patients were assigned to stage II of R-ISS than to stage II of ISS, and patients with the R-ISS stage II were heterogeneous in terms of prognostic factors such as elevated LDH and adverse CA (Table 1). Therefore, we further evaluated OS of each constituent of the R-ISS stage II according to the risk factors (Table 1). The R-ISS stage II is composed of six subgroups: (i) ISS stage I with 1 risk factor, (ii) ISS stage I with 2 risk factors, (iii) ISS stage II without risk factor, (iv) ISS stage II with 1 risk factor, (v) ISS stage II with 2 risk factors, and (vi) ISS stage III without risk factor. We have confirmed that the outcome of the six subgroups of the R-ISS stage II patients was extremely variable than we had anticipated; median OS for the subgroup (i); either 52.6 months or 100.7 month depending on which of the two risk factors was met, subgroup (ii): 34.0 months, subgroup (iii); 130.8 months, subgroup (iv): either 49.5 or 53.8 months depending on which of the two factors was met, subgroup (v): 46.9 months, and subgroup (vi): 61.8 months, respectively (p = 6.8 × 10−6, and p values between each subgroup are shown in supplementary Table 4). These findings would implicate that the presence of abnormal LDH and/or high-risk CA defined in the R-ISS might have a much stronger impact on OS than the two factors defined in the original ISS (i.e., serum albumin and β2-microglobulin).

Discussion

In the present study, we have evaluated the clinical relevance of the original ISS and R-ISS based on the long-term follow-up data of Japanese MM patients treated initially with conventional chemotherapy or with novel agents ± ASCT in routine clinical practice.

We first assessed the treatment response according to the ISS and R-ISS stages. The percentage of patients achieving PR or better was similar between the ISS stages and the R-ISS stages. The percentage of patients achieving deep response (sCR and CR) was significantly higher in stage I patients than in stage II and III patients in the ISS, but such a difference was not observed in the R-ISS stages. Next, we assessed the response according to the treatment modality. It was found that the use of novel agents and ASCT in initial therapy significantly yielded a better response with higher deep responses. Accordingly, the treatment modalities such as the use of novel agents and ASCT are more important in treatment response than the ISS and R-ISS stages. Our previous study in routine clinical practice has shown that initial treatment with novel agents and/or ASCT achieved a deep response such as sCR and CR, which was associated with a long-term survival [13]. In this context, deep response achieved by treatment is more important as a prognostic factor for patient outcome. Recently, meta-analyses of clinical trials have shown that minimal residual disease (MRD) negativity was one of the most powerful prognostic factors in OS regardless of treatment or risk status [17, 18]. Thus, we need to develop treatment regimens incorporating the next-generation novel agents that would achieve a higher MRD negativity status [19].

We evaluated the potential of the two staging systems in association of survival outcome of 718 patients treated upfront with either conventional chemotherapy or novel therapy followed by ASCT or not. It was found that separation between stages was more distinct in the R-ISS than the ISS in accordance with the previous reports [9,10,11], particularly in non-transplanted patients (Figs. 1d and 2d). However, the usefulness of the ISS and R-ISS was limited in the groups of patients treated with novel agents (Figs. 1c and 2c) or with ASCT (Figs. 1e and 2e). According to the original report, the median OS of R-ISS stage III was reported to be 42 months in the transplanted patients [8]; however, it was extended to 73.8 months in the corresponding patients in our study. We understand the comparison itself is not pertinent, but as a possible explanation for this outstanding result, we would presume that the prognosis of the stage III patients could well be improved by the use of upfront ASCT and novel salvage therapies in transplant-eligible patients. In fact, another Japanese study group has indicated that recent induction therapy with novel agents followed by ASCT would be able to overcome the poor prognosis of the ISS stage III disease and has implicated that the ISS would no longer be able to stratify the prognosis of the transplanted patients [20]. Importantly, bortezomib has been shown to improve renal insufficiency and survival of patients with renal failure having a high level of β2-microglobulin to be diagnosed with ISS stage III disease [21]. Recent large-scale clinical phase III studies on upfront ASCT after induction therapy including bortezomib have demonstrated a significant impact of ASCT on deep responses associated with MRD negativity and prolongation of progression-free survival [22, 23]. In addition, recent clinical studies have disclosed a favorable effect of proteasome inhibitors such as bortezomib and carfilzomib to improve the response and OS in high-risk patients harboring t(4;14) and del(17p) [24]. Taken together, frontline treatment with novel agents + ASCT and probably subsequent novel salvage therapy would likely to improve outcome even in patients with risk features such as renal failure and adverse CA. Of interest, the R-ISS could clarify a superb outcome of the stage I patients treated with upfront ASCT (Fig. 2e), which was not detected by the original ISS. Thus, the R-ISS could also be beneficial for identifying a good prognosis group without abnormal LDH and adverse CA, and these patients would mostly benefit from recent advances in treatment with novel agents and ASCT.

Notably, according to the definition of R-ISS, more patients would be obliged to be assigned to the R-ISS stage II. We have found that a higher number of patients were classified as R-ISS stage II, and more importantly, we have also found that the outcome of these patients was quite heterogeneous depending on the presence or absence of risk factors comprising elevated LDH and adverse CA. This heterogeneity of the R-ISS stage II has also been pointed out by other groups [25, 26], indicating that elevated LDH and adverse CA were more powerful prognostic factors than albumin and β2-microglobulin under the current treatment. Moreover, a recent genome-wide analysis of the largest set of patients has identified an extremely poor outcome group that is called as Double-Hit MM, either with bi-allelic TP53 inactivation or with amplification of CKS1B in ISS stage III patients [27]. Thus, the importance of genomic abnormality in prognostication is being further clarified in patients with MM, and the clinical relevance of the prognostic factors for a staging system needs to be reevaluated along with advances in research and treatment in the future.

In conclusion, this real-life, multicenter, retrospective study has demonstrated that the R-ISS is a valid and reliable tool for prognostication in both transplant-eligible and -ineligible MM patients regardless of initial treatment either with conventional chemotherapy or with novel agents. Notably, the R-ISS is useful for the discrimination of the stage I disease with very good prognosis when treated with upfront ASCT. However, the R-ISS still has a challenge because R-ISS stage II patients are composed of quite heterogeneous population in terms of OS due to variable fulfillment of prognostic factors and also because the prognosis of a proportion of the stage III has become improved due to advances in novel therapies and ASCT. Our results would provide an important basis for evaluating outcome achieved by the progress of treatment in the future clinical practice of MM.