Unexpected Heterogeneity of Newly Diagnosed Multiple Myeloma Patients with Plasmacytomas

In multiple myeloma (MM), malignant plasma cells infiltrate the bone marrow. In some cases, plasma cells migrate out of the bone marrow creating either para-skeletal plasmacytomas (PS) or infiltrating soft tissues as extramedullary plasmacytomas (EMD). The aim of this study was to define risk groups in newly diagnosed MM (NDMM) patients with PS and EMD plasmacytomas. In total, 523 NDMM patients with PS plasmacytomas and 196 NDMM patients with EMD plasmacytomas were diagnosed in the Czech Republic between 2004 and 2021 using modern imaging methods. Patients’ data were analyzed from the Registry of Monoclonal Gammopathies of the Czech Myeloma Group. In NDMM patients with PS plasmacytomas, we found a subgroup with <5% of bone-marrow plasma cells to have the best prognosis (mPFS: 58.3 months (95% CI: 33.0–NA); mOS: not reached). The subgroup with >5% of bone-marrow plasma cells and ≥3 plasmacytomas had the worst prognosis (mPFS: 19.3 months (95% CI: 13.4–28.8), p < 0.001; mOS: 27.9 months (95% CI: 19.3–67.8), p < 0.001). Our results show association between tumor burden and prognosis of NDMM patients with plasmacytomas. In the case of PS plasmacytomas, NDMM patients with low BM PC infiltration have an excellent prognosis.

MM is characterized by infiltration of the bone marrow (BM) by malignant plasma cells (PCs). In some cases, these PCs migrate out of the BM creating two types of plasmacytomas: paraskeletal lesions (tumors arising directly from the bone lesion; PS plasmacytomas) or

Patient' Selection
This multicentric real-life retrospective study was carried out in major hematologic centers in the Czech Republic between 2004 and 2021. For the data search of NDMM with plasmacytomas, the Registry of Monoclonal Gammopathies (RMG) of the Czech Myeloma Group was used. In total, 7123 NDMM patients fulfilling International Myeloma Working Group (IMWG) diagnostic criteria for MM were evaluated. We excluded all patients treated only with conventional chemotherapy or diagnosed only by skeletal survey; 523 NDMM patients with PS plasmacytomas, 196 NDMM patients with EMD plasmacytomas, and 2440 reference NDMM patients with clear absence of plasmacytoma (proven by high-sensitivity imaging methods) were identified. Of the enrolled patients with BM PCs <10% fulfilled MM diagnostic criteria with plasmacytoma/bone lesion tissue histology together with CRAB (i.e., osteolytic lesions, hypercalcemia etc.), or myeloma-defining events according to IMWG criteria. Patients with other plasma-cell dyscrasias (i.e., solitary plasmacytoma or solitary plasmacytoma with minimal marrow involvement) were not enrolled into the study. All participants provided written informed consent approved by institutional ethics boards in accordance with the latest Helsinki declaration.

Imaging Methods
The entire study cohort was evaluated for the presence of plasmacytomas by modern imaging methods-computed tomography (CT), focused/whole body magnetic resonance imaging (MRI/WB-MRI), or positron-emission tomography/computed tomography (PET/CT) [21]. Multiple diagnostic methods were performed on patients when clinically needed. If there was a clinical need and when safe for the patient, plasmacytomas were confirmed by surgical sampling.

Bone-Marrow Assessment
Bone-marrow samples were evaluated at the time of NDMM diagnosis. The number of BM PCs was evaluated by cytology, the clonality of BM PCs was evaluated by flowcytometry, and interphase fluorescent in-situ hybridization (I-FISH) analyses of commonly found aberrations was performed on separated PCs as previously described [22].

Response Assessment and Survival Intervals
Treatment response was assessed according to the current International Myeloma Working Group (IMWG) criteria [23]. Survival intervals (progression-free survival, PFS and overall survival, OS) were assessed from the NDMM diagnosis.

Statistics
Data were described by absolute and relative frequencies of categorical variables and median with 5th-95th percentile range for quantitative variables. Fisher's exact test was used to evaluate the association of selected features. The differences in survival (OS and PFS) among individual patient groups were assessed by the Kaplan-Meier method, and the statistical significance of differences in survival was evaluated using the log-rank test. The univariable Cox proportional-hazards model was used to quantify the effect of individual clinical features on the survival measures. The independence of selected features as prognostic survival factors was tested in the multivariable Cox proportional-hazards model in the context of R-ISS (Revised International Staging System). Statistical significance of hazard ratios (HR) was assessed by means of the Wald test. The cut-off for BM PCs was defined as the value where multivariable Cox regression adjusted to ISS showed highest HR and significance for OS and PFS, and the numbers of patients in the resulting groups was still sufficient. All statistical tests were performed at a significance level of α = 0.05 (all tests two-sided We also analyzed i-FISH data of all three subgroups of patients. Statistically significant differences between these groups of patients were not found.

Heterogeneity in PS and EMD Subgroups of Patients
In addition to generally accepted prognostic markers, such as the ISS and R-ISS stage, we found intra-and/or extramedullary tumor burden to be an important prognostic indicator of both subgroups of plasmacytomas.
In the PS subgroup, we found <5% of BM PCs (regardless of plasmacytoma count) to have the best prognosis (mPFS: 58.3 months (95% CI: 33.0-NA), p < 0.001; mOS: not reached). These patients had more frequent low ISS stage (ISS I), low monoclonal immunoglobulin secretory activity, lower ratio of clonal BM PCs (<95% from all BM PCs), intact IgH gene, and relatively low plasmacytoma count.

Discussion
In the last decade, MM has gradually become a more manageable disease. Modern induction-treatment protocols lead to deeper responses and longer remissions in most NDMM patients [25][26][27][28]. Moreover, even treatment results of NDMM patients with plasmacytomas have improved when compared to historical data [13,14,19,20,29].
In this work, we analyzed NDMM patients with both paraskeletal and extramedullary plasmacytomas. As treatment regimens based on conventional chemotherapy are more than two decades obsolete in NDMM treatment [30], we censored historical patients treated with this approach from our analysis, with the aim of understanding the prognostic impact of plasmacytomas found in NDMM patients in real-life treatment scenarios. Similarly, as discussed above, we censored NDMM patients evaluated without high-sensitive imaging methods, such as CT, MRI, WB-MRI, or PET/CT.
In light of modern real-life treatment, we found NDMM patients with PS and EMD plasmacytomas to have comparable initial treatment results to those of NDMM patients without plasmacytomas (median PFS: 25.8 and 17.9 vs. 23.3 months; p = NS). Similarly to our analysis, in a study of NDMM with mostly PS plasmacytomas, predominantly IMID-based regimens lead to treatment results comparable to NDMM patients without plasmacytomas (median PFS: 25.3 months vs. 25.2 months; p = 0.46) [14]. In another analysis of NDMM patients treated with bortezomib-based induction, patients with single PS plasmacytoma had comparable treatment results to NDMM patients without any plasmacytomas (median PFS: 34.6 months vs. 38.1 months; p = 0.662). These results were improved with ASCT (median PFS: 46.0 months vs. 15.3 months; p = 0.073), but only 16% of patients with plasmacytomas in this study underwent ASCT [20]. Tandem ASCT did not show significant benefit in NDMM patients with plasmacytomas when compared to single ASCT [19] but may somehow have improved the inferior outcome of NDMM patients with plasmacytomas with high-risk cytogenetic aberrations [31]. Taken together, real-life induction protocols containing PIs, IMIDs and, in eligible patients, high-dose chemotherapy followed by ASCT, could change the outcome of NDMM patients with PS plasmacytomas.
Unfortunately, even our real-life dataset did not have many patients treated with anti-CD38 antibodies. While there is a lack of data about activity of anti-CD38 antibodies in the NDMM patients with plasmacytomas, data from RRMM patients with plasmacytomas are not promising [19,32].
In recently published papers, heterogeneity of MM patients with plasmacytomas is evident [7,8]. Moreover, this heterogeneity is strongly reflected in patients' prognosis. It is obvious that the clinical course of NDMM patients with plasmacytomas and RRMM patients with plasmacytomas is dramatically different [12,14,18,19].
Further, there is a difference in clinical characteristics and prognosis between patients with EMD and PS plasmacytomas [7,8]. In accord with recent papers [7,8,33], our clinical and research groups have long considered PS and EMD plasmacytomas as very different entities in both NDMM and RRMM patients. Thus, we have studied them separately [9,12,18,34]. However, other groups describe all plasmacytomas outside of BM environment as one unit often making comparisons impossible [13,28,35].
The robust size of our dataset provides deeper insight, surprisingly revealing a significant heterogeneity even within subgroups of patients with PS or EMD plasmacytomas. This heterogeneity was reflected in far different prognoses.
In our study, the subgroup with PS plasmacytomas with very low BM PC (<5%) infiltration had surprisingly good prognosis; we found significantly more patients with clinical features associated with better prognosis, such as low ISS stage and lower proportion of clonal PCs in BM, called MGUS-like phenotype. MGUS-like phenotype is associated with an indolent clinical course and long survival [36]. in addition to low intramedullary tumor burden, plasmacytoma count was also low.
On the other hand, patients with higher BM PC infiltration together with increasing plasmacytoma count had worse prognosis. More patients with unfavorable characteristics were present, together with a higher frequency of numerous osteolytic lesions and higher paraprotein secretory activity. Our findings are supported by recent analyses, showing the negative prognostic impact of BM infiltration (cut-off >30% BM PCs) together with multiple plasmacytomas [20]. Similarly to other analyses, multiple PS plasmacytomas were also connected with inferior prognosis [11,14].
The most important result of our analysis was the prognostic impact of EMD plasmacytomas in NDMM patients. Generally, the presence of EMD plasmacytoma is a strong negative prognostic factor in MM patients, more pronounced in RRMM patients [11][12][13]16,18,19,35]. Interestingly, beside long-known prognostic indicators, such as ISS and R-ISS stage, we found EMD plasmacytoma count to be a strong biomarker of worse prognosis. The subgroup of our patients with numerous EMD plasmacytomas had dismal prognosis, with median overall survival near 17 months. On the other hand, we showed that in NDMM patients with a low number (1-2) of EMD plasmacytomas, treatment results were somehow comparable with NDMM patient without plasmacytoma presence (median PFS: 20.9 months vs. 23.3 months, median OS: 50.9 months vs. 55.0 months). Our results highlight the importance of high-sensitive imaging methods especially in patients with EMD plasmacytomas. According to the high prognostic impact of EMD plasmacytoma count, we recommend whole-body methods, such as PET/CT or WB-MRI in the standard diagnostic workup of NDMM patients with plasmacytomas [21].
These findings are in strict contrast with the dismal effect of EMD presence in RRMM patients [12,18,19] but, to the best of our knowledge, we have described the largest sofar published cohort of NDMM patients with EMD plasmacytomas detected by modern diagnostic methods and treated in real-life scenarios. Our findings are supported by another analysis of transplant-eligible NDMM patients where EMD patients with localized plasmacytoma involvement had a comparable 3-year PFS to that of NDMM patients without plasmacytomas (HR 1.03 (95% CI: 0.66-1.62; p = 0.88)), but patients with disseminated EMD plasmacytomas had worse prognosis (3-year PSF: HR 3.40 (95% CI: 1.74-6.61; p < 0.001) [11].
We found well-known high-risk cytogenetic aberrations del(17p) a t(4;14) retaining their negative prognostic impact predominantly in NDMM patients with PS plasmacytomas, as described in other studies [37,38]. Moreover, we found a negative prognostic impact of gain(1q21) in NDMM patients with EMD plasmacytomas. This finding is interesting, as we have previously published the higher risk of future plasmacytoma development in NDMM patients with gain(1q21) [18]. Similarly, analyses of a small groups of patients described higher incidence of gain(1q21) in NDMM patients with numerous EMD plasmacytomas [39,40]. Unfortunately, our results are based on a low number of evaluated samples. Another limitation was the absence of plasmacytoma tissue cytogenetics analysis in our work. Routine evaluation of plasmacytoma tissue can be complicated, mostly due to possible risk for patients from surgical sampling (i.e., paraspinal plasmacytomas and EMD in parenchymatous organs or the CNS).

Conclusions
Taken together, there is emerging evidence of the importance of the distinction between EMD and PS plasmacytomas in NDMM patients. Moreover, within these two entities, we found significant clinical heterogeneity, based on intra-and extramedullary tumor burden. These easy-to-assess biomarkers might reflect far different disease biology. Patients with PS plasmacytomas and low BM PC infiltration predominantly harbor low-risk features and have surprisingly good prognosis. In contrast, patients with higher BM PC infiltration and numerous PS plasmacytomas have poor prognosis. Prognosis of NDMM patients with EMD plasmacytomas is highly dependent on extramedullary burden. Patients with 1-2 EMD plasmacytomas had surprisingly comparable outcomes to NDMM patients without plasmacytoma. On the other hand, those with numerous EMD plasmacytomas had a dismal prognosis, resembling aggressivity of EMD plasmacytomas in RRMM patients.