Myelofibrosis (MF), which includes both primary MF (PMF) and secondary MF (SMF) that develops from pre-existing polycythemia vera (PV-MF) or essential thrombocythemia (ET-MF), represents a heterogeneous and complex hematologic malignancy.¹ While  specific non-driver myeloid neoplasms-associated cancer gene variants (CGVs) are well - es tablished in PMF, their significance in SMF, and their impac t on patient outcomes has remained less clearly defined.¹ At the EHA Congress 2025, Dr. Mora Barbara from Fondazione IRCCS Ospedale Maggiore Policlinico, Italy, presented results from the updated MYSEC (MYelofibrosis SECondary to PV and ET) project, which shed light on the pattern of distribution, correlations, and prognostic implications of CGVSs in a large cohort of SMF patients, providing valuable insights for clinical decision-making.¹

The MYSEC project enrolled 1,284 patients with SMF from 19 centers across Europe, the United States, and Australia.¹ This analysis was conducted on 644 SMF patients with a targeted myeloid-panel next-generation sequencing (NGS) assay, within 12 months prior to SMF diagnosis, or within three months before evolution into blast phase.¹ The analysis focused exclusively on pathogenic (or likely pathogenic) CGVs with a variant allele frequency (VAF) of ≥1%.¹

Strikingly, two-thirds of patients (66.6%) harbored at least one CGV, with 36.3% presenting a single variant, 18.6% carrying two variants, and a smaller subset bearing three or more.¹ Among the variants detected in at least 2% of cases, ASXL1 was most frequent (29.7%), followed by TET2 (19.9%), DNMT3A (6.7%), EZH2 (5.1%), SF3B1 (4.7%), U2AF1 (4.3%) and TP53 (4.2%).¹ Many of the identified variants were correlated with more severe hematologic disruptions.¹ Compared to unmutated cases, patients harboring ASXL1, EZH2, U2AF1, TP53, ZRSR2, and CBL mutations exhibited lower median hemoglobin levels (p<0.05), while ASXL1, U2AF1, TP53, and SRSF2 mutations were linked to reduced platelet counts (p<0.05), and ASXL1 mutations were associated with a higher proportion of circulating blasts (p=0.03).¹

At a median follow-up of 4.2 years, the median overall survival (OS) for the cohort was 9.2 years; more detailed survival estimates underscored the prognostic weight of specific mutations.¹ ASXL1 mutations corresponded to a median OS of 7.5 years.¹ In contrast, mutations in U2AF1 and TP53 were associated with poor outcomes—median OS of only 2.5 years—while SRSF2 mutations were associated with a median OS of merely two years.¹ Recognizing this, researchers grouped U2AF1, TP53, and SRSF2 into a unified UTS high-risk category, which then defined three distinct molecular risk classes: standard risk (no ASXL1 or UTS mutations), high risk (ASXL1 mutated but UTS wild-type), and very high risk (UTS-mutated).¹ Median OS for these groups were 11.8 years, 8.2 years, and 3.9 years, respectively.¹

To translate these insights into clinical practice, the molecular data were incorporated into a refined prognostic tool—the molecularly enhanced MYSEC-Prognostic Model (MYSEC-PM).¹ This scoring system integrates molecular features—1 point for high molecular risk (ASXL1 mutation, UTS wild-type) and 4 points for very high molecular risk (UTS mutation)—alongside traditional clinical covariates such as age, presence of constitutional symptoms, hemoglobin level, platelet count, and circulating blast percentage.¹ By doing so, patients are stratified into four risk categories: low, intermediate-1, intermediate-2, and -high.¹ These groups demonstrate clearly separate survival trajectories—with median OS of 14.5, 7.5, 4.2, and 1.8 years, respectively.¹ Patients in the intermediate-2 and high-risk categories may benefit from a more aggressive management strategy to improve outcomes.¹ Compared with the original MYSEC-PM, the enhanced model demonstrated superior prognostic accuracy, as evidenced by a higher Harrell’s C-index (0.7564 vs. 0.7373) and a lower Akaike Information Criterion (AIC) score (2,289.96 vs. 2,328.94).¹

In conclusion, the study revealed that most frequently identified CGVs were located in the ASXL1, TET2, and DNMT3A genes.¹ Significantly, the presence of ASXL1 and UTS (U2AF1, TP53, and SRSF2) variants was identified as an indicator of high and very high molecular risk categories, respectively.¹ The new molecularly enhanced MYSEC-PM was shown to stratify SMF survival into four distinct risk categories, demonstrating superior performance compared to the original MYSEC-PM.¹