A novel online calculator that combines genetic and clinical data appeared to better predict individual disease outcomes than current classification systems among patients with myeloproliferative neoplasms, according to study results published in The New England Journal of Medicine.
“Our model accurately identified a minority of patients with chronic-phase myeloproliferative neoplasms who were at substantial risk for disease progression,” Jacob Grinfeld, MB, ChB, researcher in the department of hematology at Cambridge Institute for Medical Research and Wellcome Trust/MRC Stem Cell Institute of University of Cambridge, and colleagues wrote. “Such patients could be considered for clinical trials of new therapeutic agents, since they are the most likely to benefit and the trials would be more efficient if higher-risk patients are preferentially enrolled.”
Myeloproliferative neoplasms are chronic conditions that can progress to more advanced forms of disease, including acute myeloid leukemia, with poor survival outcomes. Determining which patients are most at risk for progression has been a challenge with the current classification system, which was created in the 1950s when myeloproliferative neoplasms were divided into three clinical types.
“Current prognostic models for myeloproliferative neoplasms, which are focused on myelofibrosis, use simple scoring systems and group patients into broad prognostic categories,” the researchers wrote. “Many factors influence clinical outcomes, with a wide range of effect sizes, which means that current schemes discard information that is relevant to prognosis.”
To better understand the biological factors associated with myeloproliferative neoplasms, Grinfeld and colleagues examined 69 myeloid cancer genes from 2,035 patients.
The cohort included 1,321 patients with essential thrombocythemia, 356 patients with polycythemia vera, 309 patients with myelofibrosis, and 49 patients with other myeloproliferative neoplasms.
Researchers identified 33 genes with driver mutations in at least five patients. Single abnormalities in JAK2, CALR or MPL occurred in 45% of patients and accounted for 1,831 driver mutations.
Although loss of heterozygosity was infrequent in CALR and MPL mutations, loss was frequent in JAK2 V617F, particularly in those with polycythemia vera.
Based on these findings, the researchers identified eight genomic subgroups that were clinically different from each other.
Using individual genetic and clinical data, the researchers then developed a calculator that allowed for individual predictions of disease outcomes.
Results of internal cross-validation showed concordances of 76% to 86% for OS, EFS and transformation to acute leukemia as well as good performance on absolute predictive accuracy, according to the researchers.
“This research proves the potential of personalized medicine, using genetics,” Peter Campbell, MD, PhD, head of cancer, aging and somatic mutation at Wellcome Sanger Institute and joint head of the Cancer Genome Project, said in a press release. “Modern genomics will empower clinicians and support their decisions regarding the best therapies and clinical trials for each patient. We hope our study will be a game changer for patients with these blood cancers by providing better predictions for how their disease may behave in the future, and inform treatment choice.” – by Jennifer Southall
For more information:
The MPN Personalized Risk Calculator is available online at cancer.sanger.ac.uk/mpn-multistage.
Disclosures: The study was funded by the Wellcome Trust. Grinfeld reports grants from Bloodwise and Kay Kendall Leukaemia Trusts during the conduct of the study. Please see the study for all authors’ relevant financial disclosures.