Degree of BM involvement by different methods.
The study included 77 MM, two solitary plasmacytoma (SP), seven MGUS and four AL amyloidosis (ALA). The mean proportion of infiltrating BM plasma cells was lowest in seven patients with MGUS: 0.7% (0.04–2.6%) by MFC and 5–7% by BMC (3–10%) and BMH (1–10%); intermediate in four patients with AL-amyloidosis : 0.8% (0.1–1.9%) by MFC and ~ 10% by CM (10–20%) and HM (5–15%); and highest in 77 patients with MM: 10% by MFC (0.002-83%) and 45–50% by BMC (3-100%) and BMH (< 1-100%) (Table 1).
Plasma cell assessment revealed discrepant results in the aspirates and trephine biopsies in nine cases (10%), with PC percentage being higher according to BMH as compared to MFC results, which may be due to peripheral blood dilution of the marrow samples used for MFC or sampling errors [26, 27].
NGS results: Mutation frequency and type
In total, 102 mutations were detected by NGS in 64/90 (71%) cases analyzed. We detected one mutation in 41/90 cases (46%), and more than one mutation per sample in 23/90 (26%) cases with a maximum number of five mutations/sample. In 26 cases (29%), no mutation was detected using our NGS panel.
The proportion of cases affected by mutations was highest in patients with MM (60/77 patients, 78%), and lowest in MGUS with 2/7 (29%) patients showing evidence of mutations. One out of two SP cases was positive for BRAFV600E mutation and one out of four ALA cases showed an isolated IDH1 mutation (Table 2).
The most frequent mutations across all PCD types included members of the MAPK signaling pathway – such as NRAS and KRAS mutations [6, 7, 20]. Mutations in those two genes were detected only in samples from patients with MM and were mutually exclusive, which is in line with previous reports [6, 28]. KRAS was mutated in 22/90 samples (24%), occurring as isolated mutation in 12 of the cases (55%). In two cases, two different mutations of KRAS were identified. Gly12, Gln61, and Gly13 were the most frequently mutated residues, representing the known major and two minor hotspots in KRAS [6, 20, 29]. NRAS and DIS3 were identified as the second most recurrently mutated genes, each of them affected in 14/90 (16%) and 11/90 (12%) of cases, respectively. In 8/14 cases (57%), NRAS was the only mutation detected per sample. In contrast, mutations in DIS3 were almost always (10/11 or 91%) accompanied by other mutations, representing a unique mutation/sample in one case only (MGUS).
Mutations of FAM46C (TENT5C) were the third most frequent in the cohort: in 11 patients (12%), 14 mutations were observed. FAM46C (TENT5C) mutations were found as isolated molecular alterations in six out of 11 cases (55%).
TP53 and BRAF mutations were identified in nine samples (10%) each. TP53 was an isolated mutational event in 2 out of 9 cases only, and BRAF in 7 out of 9 (78%) of cases. TP53 mutations were seen exclusively in MM cases and showed a high frequency (6 out of 7 cases) in those patients investigated due to relapse or progression of the disease, in accordance with previously published results [16, 30–32]. Regarding BRAF, in most cases (6/9, 67%) the activating “classic” V600E type I mutation was detected, in one case the type II G469A mutation or kinase-dead (D594A) type of alteration was identified, respectively [8, 29, 33].
TRAF3 was affected in six MM cases (7%) and was found to be an isolated mutation in 2 out of 6 cases. Four cases showed FGFR3 mutations (4.5%), all with a high-risk t(4;14) translocation. This association was already previously reported [34]. IDH1 was mutated in three cases (3%) at the hot spot position R132. IRF4 and EGR1 mutations were detected in two (2%) samples each, in both cases correlated with at least one other mutated gene, respectively. For the remaining genes on the panel, namely CCND1, IDH2, MYD88 and PRDM1, no mutations were observed in this cohort.
Correlation between NGS results and degree of bone marrow infiltration by clonal plasma cells
Our analysis suggested that in these three groups, the degree of BM infiltration was the highest in samples with more than one mutation/sample with a mean of 15% aberrant PC as defined by MFC (0.1–83%), and ~ 50% (range 1-100%) PC as defined by both BM cytology and histology (Fig. 1). In samples with one mutation, identified by NGS, the pathologic PC percentage by flow cytometry was lower (mean 11%; range 0.002–62%), but similar (~ 50%, range 3-100%) as defined by both BMC and BMH.
We observed the lowest plasma cell infiltration in BM samples with no mutation detected. The mean PC infiltration in this group was only 1.6% by MFC (0.003–11%), 20% by BMC (1–90%), and 24% by BMH (1–80%) (Fig. 1).
Those findings were further confirmed by positive Jonckheere-Terpstra test (p = 0.003). The pairwise comparison has also showed a significant difference between samples without mutations and both 1) samples with one mutation per sample (p = 0.006), and 2) samples with more than one mutation per sample (p = 0.003) (Fig. 2).
The influence on outcome of the number of mutations was further confirmed by the multivariate analysis with the odds ratio (OR) for PC infiltration being 1.023 (95% CI 1.008, 1.038; p = 0.002). For comparison of various prediction’s models, see Supplementary materials, Fig S1-3.
Correlation of NGS data and clinical/biologic parameters
We performed correlation analyses between the presence and number of mutations per sample with the clinical and biological parameters of the 90 patients in the cohort, as presented in Table 3.
The majority of samples without mutation (26 cases, 29%) were from patients undergoing initial staging (19/26 cases or 73% of this subgroup), whereas BM samples with > 1 mutation (23 cases, 26%) were mostly acquired from patients with relapsed/progressive disease (15/23 cases, or 65% of this subgroup).
The level of biological markers (mean), reflecting the tumor burden, was progressively increasing from MGUS to MM cases as well as from cases with no mutation to those with one and those with > 1 mutation/sample.