The prognostic impact of bone marrow fibrosis in primary myelofibrosis.

Abstract

To the Editor: We would like to congratulate Guglielmelli et al. on their important study describing the prognostic impact of bone marrow fibrosis in 490 patients with primary myelofibrosis (PMF), evaluated at diagnosis with relevant clinical and molecular assessment and extensive follow-up information. 1 The authors conclude that fibrosis grade 2 and greater on a 0–3 scale was associated with clinical characteristics indicative of a more advanced disease including constitutional symptoms, cytopenias, larger splenomegaly, a higher IPSS risk category and more frequent prognostically adverse mutations (i.e., ASXL and EZH2) and a greater total number of same. Median survival was significantly reduced in patients with Grades 2 and 3 fibrosis as compared with Grade 1 and on multivariate analysis, the fibrosis grade independently predicted for survival. This study is important, as the authors point out it has the advantage of a “real-life” setting: fibrosis grading was performed as part of routine diagnostic approach by local pathologists, who were experienced in fibrosis assessment in the context of being established at reference centers with large accrual of patients with myeloproliferative neoplasms. Although we acknowledge and applaud the significance of their observations we had some thoughts on the authors deliberate “inclusion of only patients with a diagnosis of PMF fulfilling the 2008 WHO criteria and fibrosis grade ≥1 to avoid the inclusion of early/prefibrotic (prePMF) forms of PMF” even though fifty (9.3%) such patients were initially presented as part of a total of 540 patients originally included in this study 2. In their paper, they cite the European consensus classification of bone marrow fibrosis, and its grading: Grades 0–3 and note these criteria have been reinforced in the revised 2016 WHO criteria for diagnosis of PMF 3. As the 2016 WHO classification of PMF with fibrosis grades 0 or 1 now defines prePMF (Table 1), inclusion of patients with Grade 1 fibrosis in a study of the prognostic significance of fibrosis in PMF is problematic, now that the diagnosis of PMF requires the presence of Grade 2 or 3 fibrosis 3. The original report of this study in which patents with fibrosis grades 0–3 were included (n = 540), provided much needed information on both prefibrotic and overt PMF. The authors reported that compared with patients with fibrosis grades 0 and 1, those with fibrosis grades 2 and 3, presented more clinical features of advanced disease. Importantly there was no correlation between fibrosis grades (0–3) and phenotypic driver mutations and triple negative patients were equally distributed (10, 10.6, 14.3, and 8.8% from Grades 0–3, respectively). In addition, according to fibrosis grades 0–3, the frequency of ASXL1 (12, 15, 23.5, and 36%) and EZH2 (2, 3.9, 8.2, and 13.2%) mutations, showed similar frequencies in PMF with grade 0–1 versus 2–3 fibrosis. Similarly, the overall survival of patients with Grade 0 (which they had used as their reference group) and Grade 1 fibrosis was significantly longer than with fibrosis grades 2–3 (i.e., not reached and 14.7 vs. 6.7 and 7.2 years, respectively). As such, retaining the patients with fibrosis grade 0 and combining these with fibrosis grade 1 would have allowed for the demonstration of the clear differences (clinical, molecular, and prognostic) between prePMF (fibrosis grades 0–1) and frank PMF (Grades 2–3), as currently WHO-defined, and would have increased the current relevance of this excellent and important study 2, 3. This valuable work would have been further enriched with the inclusion of details regarding the risk of progression of lower to higher grades of fibrosis, although perhaps this will require longer follow up. Michelle A. Elliott,* Ayalew Tefferi Division of Hematology, Department of Internal Medicine, Mayo Clinic College of Medicine, 200 First St. SW, Rochester, MN 55905