Mutation landscape in patients with myelofibrosis receiving ruxolitinib or hydroxyurea

Abstract

Refractoriness to ruxolitinib in patients with myelofibrosis (MF) was associated with clonal evolution; however, whether genetic instability is promoted by ruxolitinib remains unsettled. We evaluated the mutation landscape in 71 MF patients receiving ruxolitinib (n = 46) and hydroxyurea (n = 25) and correlated with response. A spleen volume response (SVR) was obtained in 57% and 12%, respectively. Highly heterogenous patterns of mutation acquisition/loss and/or changes of variant allele frequency (VAF) were observed in the 2 patient groups without remarkable differences. In patients receiving ruxolitinib, driver mutation type and high-molecular risk profile (HMR) at baseline did not impact on response rate, while HMR and sole ASXL1 mutations predicted for SVR loss at 3 years. In patients with SVR, a decrease of ≥ 20% of JAK2V617F VAF predicted for SVR duration. VAF increase of non-driver mutations and clonal progression at follow-up correlated with SVR loss and treatment discontinuation, and clonal progression also predicted for shorter survival. These data indicate that (i) ruxolitinib does not appreciably promote clonal evolution compared with hydroxyurea, (ii) VAF increase of pre-existing and/or (ii) acquisition of new mutations while on treatment correlated with higher rate of discontinuation and/or death, and (iv) reduction of JAK2V617F VAF associated with SVR duration.

Fig. 1. Landscape plot of mutations in the study population.

Each column represents an individual patient. a: ruxolitinib treated patients; b: HU-treated patients. color code: gray indicates a mutation detected at baseline that remained unchanged at the latest follow-up sample; pink colour indicates a mutation whose VAF increased of at least 20% compared to baseline; red colour indicates a newly acquired mutation at follow-up; light green colour indicates a mutation whose VAF decreased of at least 20% compared to baseline; green colour indicates a mutation that, while detected at baseline, was no longer detected in the follow-up sample

Fig. 2. Changes of driver mutation variant allele frequency (VAF) over study period.

The VAF of the JAK2V617F, MPLW515x and CALR driver mutation was measured in samples collected at baseline (BL) and at the latest available follow-up (FU) in patients receiving ruxolitinib (a) and hydroxyurea (b)

Fig. 3. Correlation of mutation profile, at baseline and during follow-up, with maintenance of spleen volume response, treatment duration and outcome in patients receiving ruxolitinib.

Kaplan-Meyer estimates of the proportion of patients treated with ruxolitinib who presented loss of spleen response, according to IWG-MRT criteria, are shown in panels a–d in relation to: a HMR status at baseline (a); modifications of JAK2 V617F VAF at latest follow-up compared to baseline (b); clonal progression, considered as the appearance of a novel somatic variant in the follow-up sample (c); modifications of the VAF of any non-driver mutation al follow-up sample compared to baseline (d). Kaplan-Meyer estimates of the proportion of patients who discontinued ruxolitinib in relation to acquisition of clonal progression at follow-up sample are shown in e. f shows Kaplan-Meyer estimates of overall survival, measured from therapy initiation in patients treated with ruxolitinib depending on the acquisition of clonal progression in the follow-up sample