Clinical Validation of a NGS Capture Panel to Identify Mutations, Copy Number Variations and Translocations in Patients with Multiple Myeloma

Introduction

Multiple Myeloma (MM) is a heterogeneous disease with a complex clonal and subclonal architecture with few recurrent mutations. The arrival of next-generation sequencing (NGS) has allowed us to have a deeper understanding of the disease. Due to that complexity and the low recurrence of “driver” mutations, the study of general mutational profile, copy number variation (CNV) and translocations is crucial to make an accurate diagnosis and prognosis. For that reason, a clinically validated NGS capture panel has been designed to analyze in a single assay all interesting genetic aberrations simultaneously including SNVs, indels, CNVs and chromosomal translocations.

Methods

In addition to genomic DNA (gDNA) from 33 healthy donors to create a robust baseline forCNV detection,we studied 161 DNA samples from 149 newly diagnosed MM patients enrolled in the GEM2012MENOS65clinical trial and treated homogeneously: gDNA from 149 BM CD138+ plasma cells and 12 paired cfDNA from peripheral blood samples obtained at diagnosis.

First, starting with 100 ng of gDNA and 200ng for cfDNA samples, a custom targeted NGS panel using SureSelect capture technology (Agilent) followed by NextSeq500 (Illumina) sequencing identified SNVs, indels, CNVs and the most relevant IGH translocations within 26genes involved with MM. The average sequencing depth was 609x across samples; and 98% of the targeted regions were sequenced with >200x. Second, NGS custom panel results were compared with FISH (n=88) and whole exome sequencing (SureSelect XT V6) (n=48) results. Both NGS panel and exome raw data were analyzed by DREAMgenics applying a custom bioinformatic pipeline. Finally, 5 discordant cases were followed-up by SNP-arrays.

Results

We have identified 408 exonic and non-synonymous variants. At least 1 oncogenic mutation was detected in 86% (128/149) of patients. NRAS was mutated in25% of patients, followed by KRAS(23%), BRAF(12%) DIS3(11%) and TP53(9%). Other interesting pathogenic mutations were identified in FGFR3 and HIST1H1E genes in 9% and 5% of patients, respectively.

In 92% (11/12) of cfDNA samples at least 1 oncogenic mutation was detected. For this 12 cases, paired samples (BM CD138+vscfDNA) were available. A total of 39 somatic mutations were identified in those cases. In cfDNA, 10 mutations were detected, and 5 were present in both samples. Furthermore, a mean decrease of 0.19VAF was observed in cfDNA (0.12; 0.01-0.48) vs plasma cells (0.31; 0.01-0.51).

Regarding to CNV, 1q gain was detected in 32% of patients (28/88), and 1p and 17p deletions in 17% (15/88) and 13% (11/88), respectively. Additionally, ATR and CRBN gene amplifications were detected in 22% and 16%, respectively. When these data were compared to FISH, a 75% of sensitivity and 91% of specificity was achieved by our method, with a PPV of 68% and a NPG of 93%.

Translocations were identified in 28% (25/88) of patients, including 7% (6/88)) t(11;14), 14% (12/88), t(4;14), and 1 patient t(14;16). We also detected t(6;14)(p21;q32) IGH/CCND3 in 3 patients that had also been described in MM.Translocations were detected with a 94% of sensitivity, 99% of specificity, with a predictive positive value of 94% and a predictive negative value of 99%. Importantly, NGS-based method revealed a t(10;14) in 3 patients that had not been identified by FISH, a new translocation implying the miRNA hsa-mir-4537.

Finally, the impact on PFS from FISH and NGS results was analyzed separately. PFS was similar for translocations and 17p deletions. However, 1p^-detected by NGS showed a higher negative prognostic impact (p=0,006 vs p=0,127 by FISH)(Fig.1). Furthermore, our panel showed that 7% of patients (6/88) had a bi-allelic TP53 inactivation. Survival analysis showed that these patients relapsed significantly earlier than the others (p=0.028). Amplifications (≥4 copies) in 1q⁺could not identified with this panel.

Conclusions

Our custom NGS-based test allows in a single assay a more comprehensive study of the genomic landscape of MM patients by (a) detecting with a high sensitivity the most important and recurrent mutations and cytogenetic alterations, (b) identifying translocations and CNVs not previously detected by FISH and (c) identifying a double-hit MM patient. Additionally, cfDNA could be analyzed with this NGS strategy identifying molecular alterations in most of the patients.