Driver Mutations May Predict Patient Outcomes in Myeloproliferative Neoplasms

Published on: 

Myeloproliferative neoplasms have varied progression rates, and a complex genetic landscape may contribute to heterogeneity in the outcomes of patients with these diseases.

Myeloproliferative neoplasms have varied progression rates, and a complex genetic landscape may contribute to heterogeneity in the patient outcomes. However, patients with these diseases are currently diagnosed with essential thrombocythemia (ET), polycythemia vera (PV), or myelofibrosis (MF) on the basis of clinical and laboratory criteria rather than on underlying biological factors. New research, supported by the Wellcome Trust and published in The New England Journal of Medicine, suggests that genomic characterization can allow for personalized predictions of patients’ outcomes.

The researchers undertook targeted sequencing of the full coding sequence of 69 genes and genome-wide copy-number information in a total of 1887 patients, and an additional 148 patients underwent previous whole-exome sequencing.

Of the full cohort, 1321 patients had ET, 356 had PV, 309 had MF, and 49 had other myeloproliferative neoplasms. In total, 33 genes had a driver mutation in at least 5 patients; in 45% of patients, mutations in JAK2, MPL, and CALR were the sole abnormalities observed, and accounted for 1831 driver mutations. A total of 1075 driver mutations were identified across the remaining genes.


PPM1D was the eighth most commonly mutated gene in myeloproliferative neoplasms, with 10 patients having PPM1D mutations detectable while undergoing treatment with hydroxyurea, and in 20 patients within 1 month after diagnosis. Mutations in MLL3 were detected in 20 patients. Noncanonical variants in JAK2 and MPL were detected in 16 patients with triple-negative essential thrombocythemia and in 1 patient with triple-negative MF.

The researchers also found that CALR and MPL mutations more commonly occurred early in the course of disease, while mutations in NRAS, TP53, PPM1D, and NFE2 generally occured later. Early-occurring geentic mutations, including in SF3B1 and DNMT3A, were associated with age-related clonal hematopoiesis. In patients with multiple mutations, JAK2 V617F was more commonly a later event in patients with ET and an earlier event in those with PV or MF.

On the basis of their findings, the authors used a Bayesian model to identify genomic subgroups in myeloproliferative neoplasms:

  1. TP53 mutation.These mutations often occur later in the disease, and are associated with poor prognosis and a high risk of transformation to acute myeloid leukemia (AML).
  2. Chromatin or spliceosome mutation. Patients with these mutations have an increased risk for transformation to MF and shorter event-free survival.
  3. CALR mutation.These patients have a clinical course that is similar to that found in the 2 JAK2 subgroups identified below.
  4. MPL mutation.These patients are at an elevated risk of AML transformation.
  5. Homozygous JAK2 or NFE2 mutation. MF transformations occur more frequently in this subgroup.
  6. Heterozygous JAK2 mutation.These patients typically have favorable outcomes.
  7. Myeloproliferation with no known driver mutation.This subgroup has generally benign outcomes.
  8. Myeloproliferation with other driver mutations.This group includes patients with mutations in genes such as TET2 and DNMT3A (which are not disease-specific) or with mutations in genes that have been associated with other myeloid cancers.

“A key determinant of the treatment of patients with myeloproliferative neoplasms is the predicted prognosis,” write the study’s authors. “For example, patients who are expected to have a benign future clinical course would probably benefit from treatments that are aimed at minimizing thrombotic risk, and those who are expected to have progression to leukemia or myelofibrotic bone marrow failure could be candidates for intensive therapy or clinical trials of new agents.”

To this end, the researchers have launched a free online calculator of individualized patient outcomes that can allow for extrapolation of data from the cohort, and that can generate new patient predictions according to patients’ clinical, laboratory, and genomic features.


Grinfeld J, Nangalia J, Baxter J, et al. Classification and personalized prognosis in myeloproliferative neoplasms. N Engl J Med. 2018;379:1416-1430. doi: 10.1056/NEJMoa1716614.