Abstract
Primary myelofibrosis (MF) is a genetically diverse myeloproliferative neoplasm (MPN) that is often driven by somatic mutations in JAK2, CALR and MPL. Subsequent dysregulation in cytokine signaling results in collagen deposition and impairment to functional hematopoiesis. Alternative signaling pathways can also be dysregulated and alterations within the RAS/MAP Kinase (MAPK) pathway have been identified in a subset of patients. Specifically, mutations in KRAS, NRAS and CBL have been associated with poor prognostic features, shortened overall survival, increased risk for progression to acute myeloid leukemia (AML) and decreased response to JAK inhibition. This investigation centers on a comprehensive assessment of the impact of these mutations on MF molecular risk stratification, responsiveness to therapy and overall survival.
Patients with both primary and secondary MF were evaluated through the Moffitt Cancer Center comprehensive MF database and separately examined according to the presence or absence of RAS/MAPK pathway mutations. Data was collected pertaining to baseline patient and disease characteristics, including constitutional symptom presence, splenomegaly, MIPSS70 scoring and peripheral blood monocyte count. We then examined differential responses to ruxolitinib and allogenic hematopoietic stem cell transplantation (AHSCT) as well as comparative overall survival (OS) utilizing the Kaplan-Meier method.
We identified 74 patients with RAS/MAPK pathway mutations, which included those BRAF (5), CBL (24), KRAS (16), NRAS (22) and PTPN11 (7)variants and 589 without such alterations, termed RAS/MAPK wildtype (WT). Of patients with assessable data, baseline disease characteristics were comparable between these groups with several exceptions, including a higher median age (72.5 vs. 67.1 years, p<0.001) and more frequent splenomegaly (56/64 or 88.9% vs. 405/589 or 70.4%, p=0.008) in those with RAS/MAPK mutations. Notably, patients with RAS/MAPK mutations had higher-risk disease by various risk assessment tools, including MIPSS70+, with a mean scoring of 1.65 versus 1.31 in RAS/MAPK mutated and wildtype patients, respectively (p<0.001). The baseline peripheral blood monocytes were also substantially higher in patients with RAS/MAPK mutations (2.06 vs. 0.75 x 103/µL, p<0.001).
In patients who received treatment with ruxolitinib, 22/40 with RAS/MAPK mutations (55%) and 192/347 without (55.3%) responded to treatment (p= 1.0). Median duration of ruxolitinib therapy did not differ significantly between the two groups, although there was a trend toward longer exposure at 31.4 months in patients with RAS/MAPK mutations as compared to 23.9 months in wildtype (p=0.071). Interestingly, amongst patients whose mutations were identified within 1 year of initiating ruxolitinib, 10/21 (47.6%) responded to therapy. In those whose mutations either developed or were identified >1 year after therapy was instituted, 10/14 (71.4%) exhibited a response. Serial data pertaining to baseline and post-treatment mutational variant allele frequencies (VAFs) was not uniformly available at the time of this analysis.
Median OS was inferior for patients with RAS/MAPK mutations as compared to those without at 3.56 and 7.23 years, respectively, with a hazard ratio (HR) for death of 1.748 (CI 1.25-2.442, p<0.01). When adjusting for the MIPSS70+ scoring across risk groups, RAS/MAPK mutations independently predicted for adverse outcomes (HR 1.744; CI 1.546-1.967, p<0.001). AHSCT significantly impacted outcomes amongst patients with RAS/MAPK mutations as a median overall survival (OS) not reached (NR) in patients who underwent transplant as compared to 2.502 years in those who did not (log-rank p=0.0026).
The presence of baseline RAS/MAPK mutations being associated with a comparatively suboptimal response than in those whose mutations emerge later in the course of therapy is a noteworthy finding. This may suggest that RAS/MAPK mutations may not only predict for MF progression but, in subsets of patients, also represent evolution of disease rather than its true baseline character. Patients with RAS/MAPK mutations who undergo AHSCT have significantly improved overall survival compared to those who are not transplanted. Collectively, we advocate for early consideration of transplant in this population as well as for the exploration of integration of these mutations into molecular risk stratification models.
