Mutation Landscape and Prognostic Correlates of ASXL1 Variants in Primary and Secondary Myelofibrosis

Introduction: Myelofibrosis (MF), whether primary (PMF) or secondary (SMF) to polycythemia vera or essential thrombocytemia, is characterized by a complex and partially undeciphered molecular architecture. Besides mutations in driver genes (JAK2, CALR, MPL), somatic mutations in selected myeloid-associated genes have been shown to impact prognosis of MF patients (pts). Among these, ASXL1 mutations (ASXL1^MTs) are associated with poor outcomes in myeloid malignancies including PMF, where they are included in the category of "high molecular risk" (HMR) mutations along with EZH2^MTs, IDH1/2^MTs, and SRSF2^MTs (Vannucchi AM, Leukemia 2013). However, a recent study (Luque Paz D, Blood Adv 2021) questioned the value of ASXL1^MTs in MF. The current study aimed at further characterizing the prognostic role of ASXL1^MTs in MF.

Methods: After IRB approval, pts with WHO-defined MF were included in the study. Mutational analysis by targeted NGS was performed as previously described (Guglielmelli P, JCO 2017). All deposited variants were manually curated to assess pathogenicity. In this study, we also used the molecular model proposed by Luque Paz et al. that identifies 4 genetic groups: TP53^MT; High-risk (≥1 mutation in EZH2, CBL, U2AF1, SRSF2, IDH1/2); ASXL1^MT-only; and "Others".

Results: A total of 525 pts were included in the study, including 331 (63%) PMF and 194 (37%) SMF. Median age at diagnosis was 89 (18-90) years, 314 (60%) were male. The median follow-up time was 80 (98% CI, 68-90) months. Overall, 324 (62%) pts were JAK2^MT, 126 (24%) CALR^MT, 24 (5%) MPL^MT, 40 (8%) triple negative (TN), and 11 (2%) double mutated. Among non-driver genes, ASXL1^MTs were found in 158 (30%) pts, EZH2^MTs in 45 (9%), SRSF2^MTs in 37 (7%), NRAS^MTs in 30 (6%) U2AF1^MTs in 27 (5%), TP53^MTs and CBL^MTs in 25 (5%) each, IDH1/2 ^MTs in 18 (3%), and KRAS ^MTs in 15 (3%). Pts in the HMR category were 125 (38%) in PMF and 63 (32%) in SMF.

According to the above model, distribution of pts was as follows: TP53^MT n=25 (5%), High-risk n=137 (26%), ASXL1^MT-only n=64 (12%), and Others n=299 (57%). Pts in the TP53^MT and ASXL1^MT-only groups were more likely to be diagnosed with SMF compared to pts in the High-risk and Others groups (44% and 48% vs 28% and 38%, respectively). In addition, the High-risk group was enriched in TN pts (16%), while CALR^MTs were more common in the ASXL1^MT-only and Others compared to the TP53^MT and High-risk groups (25% and 27% vs 12% and 18%, respectively).

In univariate analysis, the TP53^MT and High-risk groups were associated with the worst overall survival (OS), with median values of 38 (14-110) and 55 (45-85) months (P=.0039), respectively (Fig 1A). Albeit remarkably better, the OS of pts in the ASXL1^MT-only group was inferior compared to pts in the Others group (median 124 [91-156] vs 193 [142-NR] months; P=.0118) (Fig 1A).

We then analyzed separately PMF and SMF cohorts. In the former, the TP53^MT and High-risk groups remained associated with the worst OS (median 58 [20-126] vs 55 [36-85] months), although with no significant difference, likely due to the low frequency (4%) of TP53^MTs mutations in PMF (Fig 1B). Concurrently, the negative prognostic impact of the ASXL1^MT-only group was confirmed in comparison to the Others group (median 103 [78-NR] vs 320 [178-NR] months; P=.0170). In pts with SMF, while the TP53^MT group (6%) had by far the worst OS (median 13 [6-NR] months), the OS of the ASXL1^MT-only group (median 141 [56-171] months) was comparable to that of the Others group (median 131 [106-NR] months; P=.5188) and not different from the High-risk group (median 58 [45-174] months; P=.3606) (Fig 1C). In a further analysis including only pts in the High-risk group, ASXL1^MTs were found in 62% and 63% of patients with PMF and SMF, respectively. In survival analysis, the presence of ASXL1^MTs was associated with an increased risk of death only in PMF (median OS 47 [31-73] vs 102 [34-317] months; P=.0240), unlike in SMF (median OS 90 [47-174] vs 25 [16-338] months; P=.3296) (Fig 1D-E).

Conclusion: In the current study, we critically re-addressed the prognostic impact of ASXL1^MTs by applying a genetic model recently developed by Luque Paz et al. to our cohort of molecularly annotated, WHO-defined MF pts. Overall, our results confirm that ASXL1^MTs -even in the absence of other co-occurring high-risk mutations- harbor a negative prognostic impact mainly in PMF. These findings also reinforce the idea that PMF and SMF represent two different biological entities.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal