Page 189 - 2021_06-Haematologica-web

MDS progression involves cohesin and RAS mutations
transformation, as previously explained. After a stringent analysis18 as described in the Online Supplementary Appendix, 61 variants were identified as likely somatic mutations: 40 were called driver and 21 were called passen- ger using the novel bioinformatics tool “Cancer Genome Interpreter”19 (Online Supplementary Table S4). A total of 47 variants in genes known to drive myeloid malignancies were further validated with a true positive rate of >89% using TDS and VAF correlation between two platforms was high (Pearson´s r=0.90) (Online Supplementary Figure S2). However, the application of TDS revealed that several driv- er mutations were not detected by WES as they were poor-
ly covered and displayed a low VAF. Then, we decided to more comprehensively study disease progression by apply- ing the TDS panel in a larger cohort of serially collected samples.
We performed TDS on these 16 patients of the initial dis- covery cohort and in additional 26 patients. A total of 159 mutations were identified at diagnosis of the 42 patients (Online Supplementary Table S5). The most recurrently mutated genes were TET2 (14 of 42, 33%), SF3B1 (13 of 42, 31%), SRSF2 (ten of 42, 24%), DNMT3A (10 of 42, 24%), STAG2 (8 of 42, 19%), TP53 (8 of 42, 19%) and ASXL1 (6 of 42, 14%). At the time of the second sampling, the sAML
AB
CD
Figure 1. Boxplots showing the differences in the number of mutations (A-B) and variant allele frequency (C-D) between the two times analyzed during the evolu- tion of the disease and between the different French-American-British/World Health Organization (FAB/WHO) subtypes at the time of diagnosis. (A) Differences in the number of mutations between diagnosis and follow-up/secondary acute myeloid leukemia (sAML) stages within and between the control and discovery cohorts. These graphs show a statistically significant increase in the number of mutations during disease evolution in patients who progressed to sAML (P<0.0001). (B) No significant differences in the number of mutations between the FAB/WHO subtypes at time of diagnosis (number of mutations low-risk myelodysplastic syn- dromes [LR-MDS] vs. high-risk MDS [HR-MDS], P=0.588). (C) Differences in variant allele frequency (VAF) of detected mutations between diagnosis and sAML stage in the discovery cohort. The VAF was higher at the time of sAML progression (P<0.0001). (D) No significant difference in VAF between the FAB/WHO subtypes at time of diagnosis (VAF: LR-MDS vs. HR MDS, P=0.528). NS: not significant; *P<0.05; ****P<0.0001.
haematologica | 2021; 106(8)
2217