Classification of Philadelphia-Negative MPN As Low Risk and High Risk MPN Based on Peripheral Blood Values and Molecular Genetics Only

Background: Philadelphia-negative myeloproliferative neoplasms (Ph^- MPN) are characterized by overproduction of differentiated hematopoietic cells mediated in the majority of cases by mutations in JAK2, MPL or CALR (MPN driver genes). Ph^- MPN include polycythemia vera (PV), essential thrombocythemia (ET), primary myelofibrosis (PMF) and MPN, unclassifiable (MPN-U) which differ in clinical presentation, prognosis, fibrosis and leukemic transformation rate as well as therapeutic options. However, WHO based classification requires experienced pathologists, in particular to diagnose prefibrotic PMF as high risk MPN, and thus remains observer dependent.

Aim: Observer independent classification of Ph^- MPN solely based on molecular genetics and peripheral blood (PB) parameters.

Patients and Methods: We analyzed 267 Ph^- MPN cases diagnosed as ET (n=76), PV (n=74), PMF (n=68) or MPN-U (n=49) following the WHO4R classification including only patients with an MPN driver mutation (median age: 68 years [22-91]; female: 41%; median follow-up: 5 years). All samples were analyzed by cytomorphology and whole genome sequencing (median coverage 100x). Mutation status of 20 genes associated with myeloid malignancies were analyzed in detail.

Results: Genomic landscape of the MPN cohort showed mutations in JAK2 in 222 (83%) (n=132: single mutation; n=90: either mutation and copy neutral loss of heterozygosity or mutation and 9p gain), in CALR in 36 (14%), and in MPL in 9 (3%) patients. The median number of mutations per patient was two (range: 1-6). Within the cohort, 141 patients (53%) harbored MPN driver mutations only, while in 126 (47%) one to 5 additional mutations were found. The most frequent additional mutations were ASXL1 (n=50), TET2 (n=49; thereof 15 biallelic), SRSF2 (n=26), DNMT3A (n=18), U2AF1 (n=11) and TP53 (n=9; thereof 3 biallelic). Regarding PB parameters, high WBC (≥11×10⁹/ L) were detected in 127 patients (48%), high PLT (≥450×10⁹/ L) in 116 (43%) and anemia (HB <♂12 or ♀11 g/dL) in 50 patients (19%).

We then performed Cox regression analyses for overall survival (OS) including only genetic abnormalities and PB values. In univariate analyses, biallelic TET2 inactivation (TET2bi: hazard ratio/HR: 3.789), RUNX1 (n=4; HR: 13.479), ASXL1 (HR: 1.886), SRSF2 (HR: 3.218) mutations as well as high PLT (HR: 0.440), high WBC (HR: 1.998) and PB blasts >1% (n=57; HR: 1.904) showed significant associations with OS (for all p<0.02), while no effect was observed for MPN driver mutations, DNMT3A and monoallelic TET2 mutations. In addition, anemia showed a trend towards poor prognosis (HR: 1.595; p=0.079). Following this, we determined high and low risk MPN. In the absence of thrombocytosis, high risk MPN was defined if any of the four following risk factors were present: i) mutation in ASXL1 or SRSF2 or RUNX1 or TET2bi, ii) PB blasts >1%, iii) high WBC or iv) anemia. In the presence of thrombocytosis, at least two of these risk factors were required (Figure 1).

Survival analysis revealed a significantly worse median OS of high risk (n=137; 51%) compared to low risk MPN (n=130; 49%) (8 years vs. not reached; p<0.001). This proposed novel stratification reflected prognosis better than WHO entities (median OS: MPN-U: 7 years; MF: 8 years; PV/ ET not reached; overall p=0.001) (Figure 1A). All 76 ET cases (100%) and 44/74 PV cases (59%) were classified as low risk MPN, while 63/68 PMF cases (93%) and 44/49 MPN-U cases (90%) were classified as high risk MPN (Figure 1B). Within low risk and high risk MPN, no significant OS differences were observed between the WHO entities (p=0.114 and p=0.355, respectively).

Conclusions: JAK2-pathway mutated Ph^- MPN can be categorized into low and high risk MPN based on molecular genetics and PB values only. The presence of these objectively assessable risk factors in a patient with ET or PV warrants for a profound reevaluation of bone marrow histology (eventually by a reference pathologist for MPN) and clinical course before a high risk MPN (including prefibrotic PMF) can be ruled out. Whether the predictive value of low risk MPN based on molecular genetics and PB values only could be sufficient to omit BM biopsy in selected patients, needs to be evaluated in further prospective studies.