JAK2 or CALR mutation status defines subtypes of essential thrombocythemia with substantially different clinical course and outcomes

Abstract

Patients with essential thrombocythemia may carry JAK2 (V617F), an MPL substitution, or a calreticulin gene (CALR) mutation. We studied biologic and clinical features of essential thrombocythemia according to JAK2 or CALR mutation status and in relation to those of polycythemia vera. The mutant allele burden was lower in JAK2-mutated than in CALR-mutated essential thrombocythemia. Patients with JAK2 (V617F) were older, had a higher hemoglobin level and white blood cell count, and lower platelet count and serum erythropoietin than those with CALR mutation. Hematologic parameters of patients with JAK2-mutated essential thrombocythemia or polycythemia vera were related to the mutant allele burden. While no polycythemic transformation was observed in CALR-mutated patients, the cumulative risk was 29% at 15 years in those with JAK2-mutated essential thrombocythemia. There was no significant difference in myelofibrotic transformation between the 2 subtypes of essential thrombocythemia. Patients with JAK2-mutated essential thrombocythemia and those with polycythemia vera had a similar risk of thrombosis, which was twice that of patients with the CALR mutation. These observations are consistent with the notion that JAK2-mutated essential thrombocythemia and polycythemia vera represent different phenotypes of a single myeloproliferative neoplasm, whereas CALR-mutated essential thrombocythemia is a distinct disease entity.

Figure 1

Hematologic parameters in patients with CALR-mutated ET, JAK2 (V617F)-mutated ET, and JAK2 (V617F)-mutated PV. Data are shown in a box plot depicting the upper and lower adjacent values (highest and lowest horizontal line, respectively), upper and lower quartile with median value (box), and outside values (dots). Of note, patients with JAK2 (V617F)-mutated ET had markedly lower serum Epo values than those of patients with CALR exon 9–mutated ET, despite the fact that the difference in median Hb levels was <1 g/dL.

Figure 2

Granulocyte mutant allele burden in JAK2 (V617F)-mutated and in CALR-mutated myeloid neoplasms. Data are shown in a box plot depicting the upper and lower adjacent values (highest and lowest horizontal line, respectively), upper and lower quartile with median value (box), and outside values (dots). (A) This analysis includes 250 patients with ET and 212 patients with PV at presentation, and 18 patients with post-ET myelofibrosis and 55 with post -PV myelofibrosis. In these JAK2 (V617F)-mutated myeloid neoplasms, progression from the primary disease to secondary myelofibrosis appears to be related to the mutant allele burden. In particular, the proportion of patients with values >50% increases progressively, indicating an increasingly higher proportion of cells that are homozygous for the mutation as a result of copy neutral loss of heterozygosity of chromosome 9p. Most patients with post-ET or post-PV myelofibrosis have values for granulocyte JAK2 (V617F)-mutant allele burden greater than 75%, consistent with a dominant population of homozygous cells. (B) This analysis includes 38 patients with ET at presentation and 10 patients with post-ET myelofibrosis. Also within these CALR-mutated myeloid neoplasms, progression to secondary myelofibrosis appears to be associated with a significant increase in the mutant allele burden. However, only 1 patient with post-ET myelofibrosis had a value consistent with a dominant population of homozygous cells. This might suggest that the higher mutant allele burden in patients with post-ET myelofibrosis most often reflects the progressive expansion of a heterozygous clone that eventually achieves full dominance in the bone marrow.

Figure 3

Relationship between granulocyte JAK2 (V617F)-mutant allele burden and hematologic parameters in patients with ET or PV. The mutant allele burden was directly correlated with Hb level (ρ = 0.53, P < .001) and hematocrit (ρ = 0.61, P < .001), and inversely correlated with PLT count (ρ = −0.18, P < .001) and serum Epo level (ρ = −0.23, P < .001). These correlations suggest that the mutant allele burden is a determinant of the phenotypic features of JAK2 (V617F)-mutated MPN.

Figure 4

Cumulative incidence of polycythemic transformation, thrombotic events, myelofibrotic transformation, and leukemic transformation in patients with CALR- or JAK2-mutated ET and in those with PV. The cumulative incidences were estimated with a competing risk approach, considering death for all causes as a competing event. (A) In patients with JAK2-mutated ET, the cumulative incidence of polycythemic transformation was 28.6% (95% CI 20.7-37.0) at 15 years. No progression to PV was observed in patients with CALR-mutated ET. (B) Patients with CALR-mutated ET showed a lower incidence of thrombosis than those with JAK2-mutated ET (10.5% vs 25.1% at 15 years, P = .001), or those with PV (10.5% vs 34.7% at 15 years, P < .001). By contrast, patients with JAK2-mutated ET and those with PV did not differ in terms of cumulative incidence of thrombosis (P = .314). These differences in risk of thrombosis remained statistically significant even after adjusting for age, as detailed in the text. (C) The 15-year cumulative incidence of myelofibrotic transformation was 13.4% (CI 95% 5.4-25.2) in CALR-mutated ET, 8.4% (CI 95% 3.9-15.3) in JAK2-mutated ET, and 13.6% (CI 95% 7.3-21.9) in PV, without any significant difference among these 3 subgroups even after adjusting for age. (D) The 15-year cumulative incidence of leukemic transformation was 2.5% (CI 95% 0.2-11.3) in CALR-mutated ET, 4.3% (CI 95% 1.9-8.2) in JAK2-mutated ET, and 14.6% (CI 95% 8.4-22.3) in PV. Although CALR-mutated patients showed a lower risk of leukemic transformation in comparison with both those with JAK2-mutated ET (P = .026) and those with PV (P < .001), no significant difference was observed after adjusting for age.