Abstract

Background

Although the majority of acute myeloid leukemia (AML) patients may achieve complete remission (CR) after induction chemotherapy, about half of these patients relapse with a grave prognosis. Serial studies to assess the dynamic changes of molecular aberrations during follow-ups can help clarify the underlying mechanisms of leukemia progression. The advance and progress in the high throughput next generation sequencing provide us a good chance to comprehensively analyze gene mutations longitudinally during clinical follow-ups.

Method

We used TruSight myeloid panel (Illumina, USA), focusing on 54 genes related to myeloid neoplasms, to investigate the gene mutations of the bone marrow mononuclear cells from 154 adult de novo AML patients who obtained CR but relapsed and had adequate samples for paired samples analyses. The HiSeq platform (Illumina) was applied which had a median reading depth of 12000X. The variant calling was accomplished by CallSomaticVariants 3.6.2 (Illumina) or Genome Analysis Tool Kit (GATK) HaplotypeCaller. For detection of the large insertion of FLT3 and KMT2A, Pindel v0.2.5b8 was used specifically for chromosome 13 and 11, respectively, with default setting. Complete serial cytogenetic data were available in 92 (59.7%) patients.

Results

At diagnosis, 98% patients had at least one genetic alteration including 44.7% patients had both cytogenetic abnormalities and molecular gene mutations, 48.0% patients had only gene mutations and 5.3% patients had only cytogenetic abnormalities. The most prevalent gene mutations at diagnosis were NRAS mutations (19.4%), followed by FLT3 -ITD (16.9%), NPM1 (16.9%), and DNMT3A (16.2%) mutations. The first relapse of these patients occurred at a median of 7.5 months (range, 1.2-64.4 months) after achieving CR and 75.3% of them had early relapses within one year from CR. At relapse, 77.3% patients had genetic evolution, including 12.0% with cytogenetic evolution, 37.0% with molecular gene evolution, and 28.3% with both cytogenetic and molecular gene evolution. Comparing the gene mutations between diagnosis and relapse, IDH1 (stability 100%), SRSF2 (100%), STAG2 (100%), SMC3 (100%), NPM1 (96%), ASXL1 (90%), and TP53 (85.7%) mutations remained stable during disease evolution; in contrast, some gene mutations were unstable, such as NRAS (28%), SF3B1 (50%), FLT3 (60%), and KRAS (60%) mutations. Sixty-seven (43.5%) of 154 patients had CR samples at 30±15 days after the first standard induction chemotherapy. Among them, 23 (34.3%) patients with persistent leukemia associated gene mutations (variant allele frequency at median 3.3%, range, 1.3-18.3%) at CR after first induction chemotherapy; these patients had a significant poorer overall survival (OS) than those without detectable gene mutations at CR (19.5±3.0 vs. 25.1±8.4 months, P=0.019). The 154 patients could be separated into two groups according to the patterns of clonal genetic evolution at disease evolution: Group 1 of patients had stable mutations (n=57, 37.0%) or gain of novel mutations (n=32, 20.8%) at relapse (type 1 of evolution); Group 2 had clonal sweeping, such as loss of previously harbored mutations (n=43, 27.9%) or mixed gain and loss of mutations (n=22, 14.3%), at disease progression (type 2). Type 2 evolution was positively associated with inv(16) (P=0.010), NRAS (P<0.001), and CEBPA mutations (P<0.001) but negatively associated with FLT3 -ITD (P=0.032) at diagnosis. In univariate analysis, type 1 evolution pattern predicted a shorter OS than type 2 (19.5 ±2.1 vs. 25.1±3.9 months, P= 0.034). In multivariate Cox proportional hazards regression analysis for OS, which included those prognostic factors with P < 0.1 in univariate analysis, type 1 was still an independent poor prognostic factor (relative risk 1.58, 95% confidence interval 1.03-2.44, P = 0.038).

Conclusion

The majority of de novo AML patients had genetic evolutions at relapse, including cytogenetic changes, molecular genetic evolutions, or both. The persistence of leukemia associated gene mutations at CR was associated with a significant poorer OS. Genetic evolution patterns were heterogeneous in AML and type 2 clonal evolution was associated with distinct genetic alterations and better outcomes.

Disclosures

Chen: Celgene International Sàrl: Research Funding. Tsai: Celgene International Sàrl: Research Funding. Tang: Celgene International Sàrl: Research Funding. Kuo: Celgene International Sàrl: Research Funding. Lin: Celgene International Sàrl: Research Funding. Yao: Celgene International Sàrl: Research Funding. Li: Celgene International Sàrl: Research Funding. Huang: Celgene International Sàrl: Research Funding. Ko: Celgene International Sàrl: Research Funding. Hsu: Celgene International Sàrl: Research Funding. Lin: Celgene International Sàrl: Research Funding. Wu: Celgene International Sàrl: Research Funding. Chen: Celgene International Sàrl: Research Funding. Tsai: Celgene International Sàrl: Research Funding. Hou: Celgene International Sàrl: Research Funding. Tien: Celgene International Sàrl: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.