Triple-negative primary myelofibrosis: a comparative analysis of phenotype, genotype, and outcome

Primary myelofibrosis (PMF), along with polycythemia vera (PV) and essential thrombocythemia (ET) constitute the three major myeloproliferative neoplasms (MPNs) that are characterized by somatic JAK2 , CALR , or MPL mutations [ 1 ]. JAK2 V617F or other JAK2 mutations occur in virtually all patients with PV [ 2 ] but their incidence is lower in ET and PMF with mutational frequencies of 50–60% [ 3 , 4 ]. CALR and MPL mutations are rarely seen with PV but are detected in 20–30% and 3–7% of patients with ET [ 4 ], respectively, and in 25–30% and 5–7% of those with PMF [ 3 ]. It is now well established that these “MPN-driver mutations” might not be routinely detected in ~10–15% of patients with ET or PMF, operationally designated as “triple-negative MPN” [ 5 ]. Several studies have looked into triple-negative MPN in terms of underlying genomic alterations; whole-exome and Sanger sequencing studies, including paired tumor and control samples, have so far failed to reveal novel recurrent mutations but disclosed, instead, rare somatic or germline MPL or JAK2 variants [ 6 ]. Subsequent targeted deep sequencing studies have confirmed the presence of non-canonical MPL (i.e., other than MPL 515L/K) or JAK2 (i.e., other than JAK2 V617F) variants in some patients with triple-negative MPN [ 7 , 8 , 9 , 10 ]. The current study is focused on triple-negative PMF, assigned by institutional mutation screening methods with analytic sensitivity of 0.06% for JAK2 V617F, 5% for CALR , and 10–20% for MPL .