Non-driver mutations and JAK2V617F allele burden have been involved in progression to myelofibrosis (MF) or acute myeloid leukemia (AML) in patients with polycythemia vera (PV) and essential thrombocythemia (ET). It is unknown if both mechanisms play a different role in disease transformation and if they are useful in routine clinical practice.


JAK2V617F allele burden was monitored in 208 patients (PV n=106, ET n=102) for a median of 6.5 years (range: 1-13). Quantification of JAK2V617F allele burden was assessed on the first available sample and every year thereafter. The evolutionary pattern of JAK2V617F allele burden was categorized as persistently low (<50%), persistently high (>50%), progressive increase (> 25% from baseline) or unexplained decrease (not therapy related).

Next generation sequencing (NGS) analysis of 51 myeloid-related genes was performed in 100 patients with a median molecular follow-up of 10 years including all cases with transformation to AML or MF. Detected mutations by NGS in the last sample were studied in first paired sample obtained in the chronic phase (median time from diagnosis: 1.6 years).

Time to myelofibrosis and time to AML were calculated according to the presence of non-driver mutations or the JAK2V617F evolutionary pattern. Multivariate analysis was performed by Cox regression.


With a median follow-up of 13 years (range: 1-30) 32 patients died whereas 24 and 12 patients progressed to MF and AML, respectively. Median age at diagnosis was 63 years (range: 20-94), 115 were women (55%) and 173 (83%) received cytoreduction. A persistently low JAK2V617F allele burden was observed in 62% of patients whereas the remainder presented a persistently high (25%), a progressive increase (11%) or a non-therapy-related decrease of JAK2V617F allele burden (2%).

Non-driver mutations were detected in last sample in 48% of patients. Median number of mutations was 1 (range: 1-5). Mutational frequencies were: TET2 12%, DNMT3A 12%, TP53 9%, ASXL1 7%, RUNX1 4%, SF3B1 4%, SRSF2 4%, IDH1/2 4%, SH2B3 3 % and <2% for EZH2, ZRSR2, SETBP1, FLT3, NPM, BCOR, CBL, PTPN11, KAT6A, NF1, U2AF1, GNAS, PHF6, KTM2A, JAK2, BCORL1, NOTCH1, MPL and PRPF40B.

Mutations were detected in first sample in 28% of patients (58% of those with mutations in last sample). Frequencies of mutations in first sample were: TET2 10%, DNMT3A 5%, TP53 5%, ASXL1 4%, SRSF2 4%, IDH1/2 3%, and < 2% for SF3B1, SH2B3, KMT2A and ZRSR2. The evolutionary pattern of JAK2V617F allele burden was not associated with the presence of mutations in first or last sample.

Twelve patients progressed to AML (post-PV n=7, post-ET n=5), nine of them presented mutations in first sample. AML transformation at 15 years was 27% and 6.8% for patients with and without additional mutations in first sample, respectively (p=0.001). Mutated genes associated with a higher probability of AML transformation were DNMT3A (p<0.0001), SRSF2 (p<0.0001) and IDH (p<0.0001). Other variables associated with a higher probability of AML were age > 65 years (p=0.012) and exposure to busulfan (p=0.003). Evolutionary JAK2V617F pattern was not associated with the probability of AML (p=0.667). In multivariate analysis, an increased risk of AML transformation was observed for patients with additional mutations in the chronic phase (HR: 6.3; 95%CI 1.6-24.7, p=0.008) after adjusting for initial diagnosis, age and exposure to busulfan.

Twenty-four MF transformations were documented (post-PV n= 18, post-ET n=6). Presence of additional mutations was not associated with the probability of MF (p=0.189). Patients with persistently high or a progressive increase of the JAK2V617Fallele burden showed a higher probability of MF transformation (24% versus 1.5% at 10 years, p<0.0001) that persisted in multivariate analysis after correction for age and initial diagnosis (HR 13.9, 95%CI: 2.9-65.6, p=0.001).


Non-driver mutations are involved in the progression of PV and ET to AML but not to MF. NGS could be useful for identifying patients with PV or ET at risk of AML transformation.

Acknowledgment:This work was supported by grants from the Instituto de Salud Carlos III, Spanish Health Ministry, FISPI13/00557, FISPI1300393, RD012/0036/0004, 2014 SGR567. Alicia Senín received a grant from Sociedad Española de Hematología y Hemoterapia.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution