Personalized Genetic Monitoring of myeloma using analytic multivariate analysis of total Fully-Sequenced Circulating DNA

There is no single sequence or breakpoint region linked to all patients with myeloma. It was thus elected to extract, amplify and fully sequence all (total) DNA present in circulating blood (CNA) using pyrosequencing with the Roche/454 sequencer (GSFLX). The origin of circulating DNA was investigated by local alignment analyses using BLAST and compared with the circulating DNA of 47 healthy individuals, aged 18 to 64 years. Thirty-one samples were subjected to total CNA sequencing during the course of the disease for a patient with lambda Bence Jones myeloma sequentially monitored from relapse, through complete response for 38 months and then with development of subsequent relapse. Conventional monitoring included both serum free light chain analyses and whole body CT-PET imaging. For comparative purposes the CNA was categorized by origin, functionality, chromosomal localization and genes, based on public databases (human reference genome build 36.2 and corresponding annotation file seq_gene.md as downloaded from NCBI (ftp.ncbi.nih.gov). The average length of sequenced nucleotides was 185.7 nucleotides and the total number of nucleotides sequenced per sample was 1.2 – 2.6 million. CNA sequences occurring with a frequency significantly greater or less than those observed for healthy controls were evaluated in detail. The next steps were to examine correlations with initial active myeloma, response to treatment, the complete remission state off all therapy, and subsequent relapse. The presence of 10 DNA sequences was highly correlated with the disease course: ZMYM2; TSPAN5; TLL1; PRKD1; ANTXR1; SYT14; PRKCH; MBNL1; EGFR and EDG7. Of these ZYMYM2 was highly correlated with disease reactivation and relapse off therapy. TSPAN5 showed a clear pattern with reduction to background levels during remission, but increase at the very earliest sign of relapse evident on CT-PET. Conversely, other sequences (GRID1; KIF16B) increased substantially during the peak impact of successful therapy. Multivariate analyses were used to identify the best combinations of gene sequences predictive of disease course and outcome. The combination of four genes (GRID1; PRKD1; ANTXR1; GAB1) was sufficient to define the early active disease data points vs the normal controls (odds-ratio OR: 135 [10.6 to 172]; p<.0001). When the resulting model was kept and used as a search engine for the whole 2 years course, the time course followed the treatment success (OR: 54; p=.0017) and failure as well as the final relapse(OR: 13.3; p=.004)

In conclusion: Total sequencing of DNA has proved promising in providing personalized molecular monitoring for myeloma. The details of associated DNA sequences provide insights to the underlying molecular pathophysiology linked to myeloma disease progression, response to successful therapy as well as subsequent relapse. Further testing is ongoing.