Introduction: Prognosis of AML pts is still poor mainly because of refractoriness to or relapse after intensive chemotherapy. High rates of relapse are also attributed to LSCs, which are a small subset of cells with acquired abnormal self-renewal capacity and increased resistance to chemotherapy. A better understanding of LSCs is critical to improve outcomes of pts with AML. Ng et al. (Nature 2016;540:433) defined a 17 stemness-associated gene score that was highly prognostic.

Aims: The aim of this study was to validate the prognostic relevance of the 17-gene LSC score and explore its utility in the context of the ELN classification. We also examined gene mutations associated with the 17-gene LSC score.

Methods: We analyzed a total of 934 pts [729 aged <60 years (y) and 205 aged ≥60 y] with de novo AML. We used whole transcriptome expression data (RNAseq) to calculate the aforementioned 17-gene LSC score for each pt in our cohort. Similar to Ng et al., we used the median of the whole cohort to discriminate between pts with LSChigh and LSClow scores. The mutational status of 80 cancer- and leukemia-associated genes (Eisfeld et al. Leukemia 2017;31:2211) were determined using a targeted next-generation sequencing panel, CEBPA mutations using Sanger sequencing, and an internal tandem duplication (ITD) of the FLT3 gene using fragment analysis in pretreatment bone marrow or blood samples. All pts were treated on frontline Cancer and Leukemia Group B/Alliance protocols.

Results: A comparison of pretreatment clinical and genetic features revealed that LSChigh pts were older (P<.001; median age, 53 vs 46 y) and had higher platelet counts (P<.001; median, 63 vs 50x109/L) than LSClow pts. Pts with a LSChigh score more frequently had FLT3-ITD (P<.001) and mutations in the ASXL1 (P=.001), DNMT3A (P<.001), RUNX1 (P=.002), SRSF2 (P=.02), STAG2 (P=.009), TET2 (P=.008) and TP53 (P<.001) genes. Conversely, these pts had a lower frequency of biallelic CEBPA (P<.001), GATA2 (P=.008) and KIT (P<.001) mutations.

Because of differences in treatment intensity, we analyzed outcomes of younger and older pts separately. Younger pts with a LSChigh score had a lower complete remission (CR) rate (P<.001; 63% vs 87%), shorter disease-free survival (DFS; P<.001; 3-y rates, 26% vs 48%; Figure 1A) and overall survival (OS; P<.001; 3-y rates, 27% vs 59%; Figure 1B) compared to those of LSClow pts. In multivariable analyses including clinical and genetic factors that impact on outcome, a LSChigh score associated with lower remission rates (P<.001; HR: 0.36), shorter DFS (P<.001; HR: 1.67) and OS (P<.001; HR: 1.88) after adjusting for other co-variates. We also analyzed the prognostic impact of the LSC score with respect to the 2017 ELN classification. We found that LSC score associated with different ELN groups (P<.001), with LSChigh pts being more often classified in the Adverse or Intermediate group and less often in the Favorable group. Within the ELN Favorable and Adverse groups, LSChigh score retained its prognostic impact and identified pts with a lower CR rate and shorter DFS and OS (Table1).

In older pts, a LSChigh score also associated with lower CR rate (P=.004; 50% vs 72%), shorter DFS (P=.04; 3-y rates, 6% vs 17%; Figure 1C) and OS (P<.001; 3-y rates, 9% vs 27%; Figure 1D). In multivariable analyses, LSC score remained significant only for OS (P<.003; HR: 1.70) after adjusting for other co-variates. Regarding the ELN classification, pts with LSChigh score in the Favorable group had shorter OS (P=.05; 3-y rates, 17% vs 50%) and, by trend, shorter DFS (P=.09; 3-y rates, 17% vs 39%); no significant differences were found in Intermediate or Adverse groups.

Conclusions: We used RNAseq expression data and applied the previously established 17-gene LSC signature to score 934 de novo AML pts. We detected distinct mutational differences between LSChigh and LSClow pts, with LSChigh pts more often carrying gene mutations associated with age-related clonal hematopoiesis (i.e., ASXL1, DNMT3A, TET2, SRSF2 and TP53 mutations). Moreover, this score, derived from the expression of stemness-associated genes, has not only a prognostic impact on its own but also in the context of the current 2017 ELN classification.

Disclosures

Kolitz:Magellan Health: Consultancy, Honoraria. Powell:Rafael Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution