The utility of non‐invasive liquid biopsy for mutational analysis and minimal residual disease assessment in extramedullary multiple myeloma

Abstract

Multiple myeloma (MM) is an incurable B cell malignancy characterized by the clonal proliferation of plasma cells (PCs). Extramedullary multiple myeloma (EMM) is an advanced subset of MM, defined by the presence of clonal PCs outside of the bone marrow, which is a multi-focal disease of great heterogeneity and linked to poor prognosis (Bolli et al., 2014; Touzeau & Moreau, 2015). Tissue biopsy is invasive and is not always feasible. In recent years, many studies have illustrated the potential of liquid biopsy in molecular profiling, tracking minimal residual disease (MRD), and assessing prognosis. Some research has indicated that circulating tumor DNA (ctDNA) profiling could be adopted as an alternative to MM bone marrow (BM) aspirates (Kis et al., 2017; Oberle et al., 2017; Levin et al., 2018). In addition, an EMM study focusing on longitudinal plasma ctDNA profiling suggested that ctDNA can be used as an adjunct to currently available modalities. However, the concordance between ctDNA and extramedullary plasmacytoma-derived DNA has not previously been determined. To explore the association between plasma cell-free DNA (cfDNA) quantity and MM disease advancement, we measured 22 plasma samples from 8 EMM patients and 23 plasma samples from 10 MM patients without extramedullary spread, whose median cfDNA concentrations were 43·58 ng/ml (range 3·53–313·5 ng/ml) and 16·4 ng/mL (range 4·32–214·7 ng/ml), respectively. Significantly higher quantities of cfDNA were obtained from more advanced cases who had extramedullary spread (P = 0·009) compared to patients without extramedullary spread (Figure S1). Two MM-focused sequencing panels targeting the coding sequence regions of the same 22 recurrently mutated genes were designed (Tables S1 and S2) (Egan et al., 2012; Joaquin et al., 2014; Martinez-Lopez et al., 2014; Walker et al., 2015). Their amplicon insert sizes were distributed from 59 to 104 bp (for ctDNA) and 85 to 335 bp (for genomic DNA) separately. Genes such as KRAS, NRAS, and BRAF in the MAPK pathway, TRAF3, CYLD, TLR4, RIPK4, and LTB in the NF-κB pathway as well as TP53, CCND1, and RB1 in the cell cycle pathway were included (Chapman et al., 2011; Lohr et al., 2014; Robiou et al., 2018). Overall patient characteristics are summarized in Table 1. BM CD138+ cells were isolated using antibody-coated magnetic microbeads for subsequent sequencing if they were flow cytometry-detectable (patients EMM-1, 5, 6). For the other five patients, who had no detectable monoclonal medullary PCs, total BM aspirate was used for sequencing. The mean target coverage of medullary DNA and extramedullary tumor DNA was over 5000× (5470–12 167). The depths of coverage for time-matched ctDNA samples were all over 20 000× (20 723–70 659). A detailed description of the bioinformatics analyses performed is provided in the methods. A total of 17 different somatic mutations were detected (extramedullary tumor, n = 16; plasma, n = 10; BM aspirate, n = 4), among which, MGB probes for 16 (16/17, 94·12%) mutations were successfully designed and were verified using droplet digital PCR (Tables S3 and S4). The LTB p.Gln67* mutation, whose probe was not available, was validated by Sanger sequencing. False negative tests were then performed on all paired samples. Four mutation sites that were negative by next-generation sequencing methods (KRAS p.Gly12Val in plasma from EMM-4, BRAF p.Val600Glu in plasma sample from EMM-7, and TP53 p.Arg249Ser in the BM aspirate sample from patient EMM-5) or excluded during analysis (KRAS p.Gly12Arg in the plasma of EMM-5) were newly identified by ddPCR (Table S5). These results suggest that ddPCR was more sensitive than NGS in our system. We then analyzed the molecular profiles of paired tissue biopsies, BM aspirates, and ctDNA samples from eight EMM patients. Out of the 16 mutations detected from the extramedullary samples, 12 (66·67%) were also detected in time-matched plasma ctDNA, and only 5 (31·25%) were detected in paired BM aspirates (Fig 1B). Thus, the analytical sensitivities associated with ctDNA samples and BM aspirates were 66·67% and 31·25%, respectively. Specificity was assessed based on mutants that were detected in time-matched ctDNA samples or BM aspirates but undetectable in paired tissue biopsies (Newman et al., 2016). Only one ctDNA mutation and one BM aspirate mutation were undetectable in tissue biopsies. Negative genes were defined as those for which no mutation was detected in all three paired samples. The specificities calculated for ctDNA samples and BM aspirates were both 99·37%. The concordances of plasma ctDNA samples and BM aspirates with extramedullary tumor biopsies were assessed based on receiver operating characteristic (ROC) curves (Fig 1C) (Kjetil, 2009). The areas under the curve (AUCs) for plasma ctDNA and BM aspirates were 0·873 (P = 8·66e−7) and 0·621 (P = 0·109), respectively. These results show that ctDNA exhibited strong concordance with time-matched extramedullary plasmacytoma biopsies, and may be an appropriate surrogate when tissue biopsies are unavailable, meanwhile BM aspirate is not an appropriate surrogate for tumor tissue biopsy in EMM, because it may miss genetic alterations within extramedullary tumor cells. Plasma mutant allele fractions of patients EMM-5, EMM-2, and EMM-8 were tracked by ddPCR using sequential plasma samples spanning 206 days (EMM-5), 65 days (EMM-2), or 242 days (EMM-8). The ‘start’ time for tracking was set as the time when relative sequencing samples were acquired. The dynamics of ctDNA over time were compared with M protein levels. In patient EMM-5, The MAF of KRAS p.Gly12Arg which should represent the aggressive clone initially decreased, and then rapidly increased after day 84, and lambda LC showed a similar but delayed pattern. In patient EMM-2, the kappa LC levels and NRAS as well as TP53 mutations in plasma followed a similar pattern: all initially decreased, and then increased. In patient EMM-8, the MAFs of KRAS p.Gly12_Gly13delinsAlaThr and MAX p.Ser49* followed the same pattern as that of lambda LC, indicating a correlation between ctDNA and M protein levels (Fig 1D). In conclusion, we conducted molecular profiling of time-matched extramedullary plasmacytoma biopsies, plasma ctDNA samples, and BM aspirates from eight EMM patients using targeted deep sequencing and ddPCR assays. MAPK pathway mutations were identified in 100% of tumor biopsies. The concordance of plasma ctDNA to extramedullary tumor concordance was 0·873, which is much higher than that of BM aspirates. Therefore, ctDNA is a promising surrogate material for EMM mutational characterization, particularly when extramedullary plasmacytomas are inaccessible, and may also be used to track disease progression. This work was supported in part by the National Science Foundation for Young Scientists of China (No. 81700160 to Liting Chen), National Natural Science Foundation of China (General Program, No. 81873452 to Chunrui Li), and National Natural Science Foundation of China (No. 81830008 to Jianfeng Zhou). L.C., J.Z. and Q.X. designed the research; X.L., Y.L., Q.X., J.G., H.C., D.W., J.X., T.L., X.Z., M.X., Y.W., and X.M. performed the experiments; L.C., C.L., X.L, and Y.L. analysed the data; L.C., and X.L. wrote the paper. The authors have no competing interests. Table S1. Standard DNA library summary. Table S2. cfDNA library summary. Table S3. ddPCR assays information. Table S4. False positive rates and thresholds for all ddPCR assays. Table S5. Verification of sequencing results. Figure S1. Performance of targeted next-generation sequencing. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts