Plasma Cell Myeloma Residual Disease Quantitation Using a Next-Generation Sequencing-Based IGH Clonal Rearrangement Assay with the Aid of a "Spike-in" Clonal Sequence

Introduction: Next-generation sequencing (NGS)-based IGH clonal rearrangement assays can characterize and subsequently track disease-associated clonal sequences for lymphoid and plasma cell neoplasms, even at very low levels. As IGH PCR primers are used, the detected clonal sequences are usually reported as % of sequencing reads, roughly corresponding to % of B and plasma cells (PC) in samples, rather than % of total cellularity, hampering accurate disease burden assessment. In this study, we evaluated a method for calculating residual disease burden as % of total cellularity, with the aid of adding a known quantity of "spike-in" clonal sequence to the samples, and compared to concurrent 10-color flow cytometry (FC) quantitation of abnormal PC.

Methods: DNA was extracted from 40 plasma cell myeloma patient marrow biopsies sent for disease monitoring purposes at Memorial Sloan Kettering Cancer Center (MSKCC), with previously-characterized clonal sequences specific to the patients' myelomas. All samples had concurrent FC analyses and aspirate differential counts performed. 100 cell equivalent of DNA with a known clonal sequence (LymphoQuant®, LQ) was added to 700ng of patient DNA (~100,000 cell equivalent), and testing was performed using Lymphotrack^TM, a NGS-based assay. Following PCR amplification using IGH FR1 primers, sequencing was performed on the Illumina MiSeq^TM instruments at the molecular laboratory of MSKCC. Reproducibility studies were conducted on a subset of samples at the laboratory of Invivoscribe, Inc. using identical methodology. LymphoTrack MRD data analysis tool (MRDDAT) v.1.0.3 was used to search for both the myeloma-specific and LQ clonal sequences. Disease as # of cell equivalent was calculated as: (% reads for myeloma clonal sequence/% reads for LQ) X 100 cells. Disease as % of total cellularity was calculated as: (# of cell equivalent/100,000 cells) X 100%.

Results: Disease as % of total cellularity calculated by LQ showed a median of 0.7576% cells (range: 0.000614% to 39.89%), compared to abnormal PC as % of total WBC by FC with a median of 0.355% cells (range: 0.00061% to 44.70%). Overall, a good correlation between disease quantitation by LQ and FC could be observed for cases with ≤10% total PC by aspirate count (r=0.79), while the correlation is lower for cases with >10% total PC (r=0.51). 12/40 samples were tested in two different laboratories, and showed excellent correlation in disease quantitation by LQ (r=0.94). As expected, detectable clonal sequences as % of sequencing reads (rather than as % of total cellularity) showed poor correlation with FC quantitation (r=0.32), due to variability of total B and plasma cell content in different samples.

Conclusions: Disease as % of total cellularity calculated with the aid of a known "spike-in" sequence in the NGS-based assay showed good correlation with the quantitation of abnormal PC by FC, when total PC was ≤10% by aspirate count. The correlation between the two declines when total PC was >10%. When patient samples contain a high number of B and/or plasma cells, the PCR amplification efficiency of the very small amount of the admixed "spike-in" clonal sequence may be hampered, affecting accurate quantitation. Furthermore, FR1 primers may not anneal optimally to some patients' clonal sequences due to somatic hypermutations in binding sites, underestimating the % of disease clone. Utilization of a second "spike-in" sequence and other primer sets (FR2, FR3) may improve disease % calculations in some cases.