原发骨纤维化长期生存预测:基于历史队列的独立危险因素分析与预后模型构建
Predicting long-term survival in primary myelofibrosis
原发骨纤维化长期生存预测:基于意大利帕维亚大学单中心历史队列中 25 年以上长期生存者数据,分析预后模型。独立危险因素包括:年龄小于 50 岁、血小板计数大于 450×10⁹/L 及骨髓纤维化 0 级。
研究者基于历史队列数据,给出了可量化的原发骨纤维化长期生存预测模型。
Introduction
The World Health Organization (WHO) and the International Consensus Classification (ICC) delineate three canonical categories of BCR::ABL1-negative classical myeloproliferative neoplasms (MPNs): polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF) [1, 2]. Within PMF, bone marrow histopathology provides an additional level of biological resolution by distinguishing pre-fibrotic (pre-PMF) from overtly fibrotic PMF (overt-PMF) [3].
Although both variants represent chronic clonal myeloid neoplasms, their natural histories diverge markedly. Overt-PMF is associated with a substantially reduced overall survival (OS), with median estimates of approximately 6–7 years [4, 5], whereas pre-PMF follows a more indolent course, with median OS ranging from 15 to 17 years [6, 7]. These prognostic differences carry direct implications for clinical management. In overt-PMF, conventional clinical-hematologic variables at diagnosis, such as age, hemoglobin concentration, leukocyte count, circulating blasts, and constitutional symptoms, constitute the foundation of validated prognostic models, including the International Prognostic Scoring System (IPSS) [4]. In contrast, these same covariates show limited discriminatory capacity in pre-PMF, where most patients cluster within low-risk categories, underscoring the ongoing need to identify determinants of shortened survival in this variant [6,7,8].
In this study, we sought to refine prognostic assessment in early-stage PMF by characterizing the phenotype of long-term survivors. This strategy, complementary to traditional risk modeling approaches, aims to illuminate the biological and clinical features associated with exceptionally favorable long-term outcomes. To this end, we analyzed the clinical and biological characteristics of individuals surviving more than 25 years after diagnosis within a historical institutional cohort of PMF subjects.
Subjects and methods
Study design and population
This retrospective observational cohort study was conducted at the Center for the Study of Myelofibrosis (CSM), IRCCS Policlinico S. Matteo, Pavia, Italy. Clinical, hematological and molecular data were retrieved through a retrospective review of subjects consecutively referred to, and prospectively recorded in, the institutional Pavia‑CSM database. The establishment and use of this database were approved by the Institutional Ethics Committee of the IRCCS Policlinico San Matteo Foundation, and all participants provided written informed consent. Additional consent was obtained at the time of blood sampling for molecular analyses, informing subjects that their genomic material would be stored in an institutional biobank and used for disease‑related molecular studies (Reference 26,20110004143 September 2011). Owing to the retrospective design of the present study, the requirement for renewed individual consent was waived, provided that all subject-level data were fully anonymized prior to the analysis.
The study population comprised consecutive individuals diagnosed with PMF before January 2001, thereby ensuring a minimum potential follow‑up of 25 years from diagnosis. Diagnostic classification into overt-PMF and pre-PMF was based on the WHO criteria applicable at the time of each subject’s initial evaluation and subsequently re‑adjudicated according to the 2022 revised criteria [3]. To avoid misclassification of transplant-related long survivors and to ensure a mature cohort suitable for robust estimation of long-term outcomes, individuals who had undergone allogeneic stem‑cell transplantation were excluded. As a comparison cohort, we selected subjects diagnosed in the same period who had a follow-up duration equal to or shorter than 25 years and were documented as deceased. This strategy minimized loss-to-follow-up bias and ensured comparability of survival ascertainment between groups.
Statistics
Continuous variables were summarized as medians with corresponding ranges, whereas categorical variables were reported as absolute frequencies and percentages. Between-group comparisons were performed using the Chi-square test. Associations between baseline covariates and long-term survival were assessed using logistic regression. Receiver operating characteristic (ROC) curve analysis was applied to identify optimal cut-off values for continuous predictors. Variables achieving statistical significance in univariate analyses were entered into a multivariable logistic regression model to quantify their independent contribution. Model discrimination was assessed using the area under the ROC curve (AUC). Because logistic regression using 25-year survival (alive vs. deceased) captures both early and late mortality without accounting for event timing, and may therefore be disproportionally influenced by early deaths, we examined the robustness of these findings using a Cox proportional hazards model for overall survival (OS) with a 10-year landmark analysis. By this analysis, subjects who died within the first 10 years after diagnosis were excluded, and follow-up began at the landmark time point. Variables included in the multivariable Cox model were selected according to the same strategy used for the logistic regression analysis, namely, the inclusion of variables significant in univariable analyses. This approach mitigates immortal-time bias by excluding individuals who died before the landmark time point.
Study endpoints included progression to active disease, blast transformation (BT), death, and thrombotic events. Progression to active disease was defined by the first occurrence of any of the following: (1) hemoglobin concentration <10 g/dL; (2) splenomegaly >10 cm below the left costal margin; (3) platelet count <150 × 10⁹/L; or (4) leukocyte count <4 × 10⁹/L or >12 × 10⁹/L. To avoid confounding by treatment effects, the onset of these criteria was censored at initiation of disease-modifying therapy. BT was defined as a persistent increase in peripheral blood or bone marrow blasts ≥20%. Thrombotic events included all arterial and venous thromboembolic events, excluding superficial vein thrombosis. Events occurring within two years before diagnosis were classified as PMF-related. These events were expressed as incidence rates, calculated as the number of events per 100 subject-years with corresponding 95% confidence intervals (CI) and compared using Poisson regression models; the incidence rate ratio (IRR) was reported with 95% CI. Competing-risk analysis was considered not informative for intermediate outcomes that were censored at the start of disease-modifying therapy. Statistical significance was defined as p < 0.05. Analyses were performed using STATISTICA© (Dell Technologies Inc., Round Rock, TX, USA) and Stata v18.5 (StataCorp, College Station, TX, USA).
Results
A total of 301 subjects diagnosed with PMF before 2001 were evaluated. After excluding 34 individuals who underwent allogeneic stem cell transplantation, 41 who remained alive for more than 25 years from diagnosis were classified as long survivors. The comparison cohort consisted of 176 subjects diagnosed in the same period who died within 25 years (Supplementary Fig. S1).
Baseline clinical and hematological characteristics
Table 1 summarizes the baseline characteristics of long-term survivors and subjects with an overall survival ≤25 years. Long-term survivors were markedly younger at diagnosis, with a median age of 36 years compared with 58 years in the comparison cohort (p < 0.001). ROC curve analysis identified 50 years as the optimal age threshold for predicting long-term survival. Notably, none of the long-term survivors was older than 65 years at diagnosis (vs. 25% in the comparison cohort), and 90% were younger than 50 years (vs. 26%). Gender distribution also differed substantially, with females representing 53% of long survivors compared with 27% of controls (p = 0.001).
Baseline disease features indicated a distinctly more favorable prognostic profile among long-term survivors. Severe anemia was less frequent, whereas non-palpable splenomegaly, normal blood CD34-positive cell counts, and normal LDH levels were significantly more common. All long-term survivors were classified as IPSS low-risk, compared with 31% of the comparison cohort. White blood cell counts were comparable between groups, whereas platelet counts were significantly higher in long-term survivors (median, 634 × 10⁹/L vs. 368 × 10⁹/L, p < 0.001). Grade-0 bone marrow fibrosis was also markedly more frequent (53% vs. 10%, p < 0.001). ROC analysis identified a platelet threshold >450 × 10⁹/L as the optimal discriminator for long-term survival.
Molecular data further supported biological distinctions between cohorts. Long-term survivors exhibited a lower incidence of the JAK2V617F variant (55% vs. 69%, p = 0.10), corresponding to a proportionally higher, though not statistically significant, frequency of CALR variants (26% vs. 14%, p = 0.08). Among JAK2V617F-mutated individuals, the variant allele frequency (VAF) was significantly lower in long survivors (median, 39% vs. 58%, p = 0.004), with a greater proportion of cases in the class with VAF < 25% (33% vs. 8.9%, p = 0.015). The prevalence of ASXL1 mutations and the distribution of cytogenetic abnormalities were similar between groups (20% vs. 31.2%, p = 0.42; 25% vs. 26%, p = 0.97), although these analyses were limited by the small number of evaluable cases.
Most subjects received hydroxycarbamide, whereas only a minority were treated with ruxolitinib due to its limited availability during the study period. The frequency of hydroxycarbamide use and the rate of splenectomy did not differ significantly between cohorts.
The distribution of PMF clinical variants differed substantially. Pre-PMF phenotypes were markedly more common among long survivors (89% vs. 38%, p < 0.001). Nonetheless, a subset of patients with overt-PMF at diagnosis (11%) also achieved long-term survival.
Predictors of long survival
In univariate logistic regression, several baseline variables were significantly associated with a higher probability of long-term survival: age <50 years, female sex, platelet count >450 × 10⁹/L, grade-0 bone marrow fibrosis, and normal blood CD34-positive cell counts (Table 2). In multivariable analysis, three factors retained independent prognostic significance: age <50 years (OR, 21.26; 95% CI, 5.97–75.7), platelet count >450 × 10⁹/L (OR, 2.88; 95% CI, 1.06–7.99), and grade-0 bone marrow fibrosis (OR, 3.36; 95% CI, 1.25–9.0). When modeled as a continuous variable, age remained an independently predictive: for every 10-years decrease, the probability of achieving long-term survival increased by approximately 14%. The discriminative performance of the multivariable model was high, with an AUC of 0.91.
As a complementary analysis, a 10-year landmark Cox model was performed. A total of 72 patients died before the 10-year landmark and were therefore excluded from the landmark analysis. This analysis showed broadly consistent associations, with younger age, absence of bone marrow fibrosis, higher platelet count, and female sex remaining associated with improved subsequent survival (Supplemental Table 1).
Intermediate outcomes, blast transformation and death
Median follow-up differed markedly between cohorts, reaching 342 months [Interquartile Range (IQR), 327–398 months] in long-term survivors versus 99 months (IQR, 59–171) in controls.
After censoring at the initiation of disease-modifying therapies (hydroxycarbamide, ruxolitinib) or splenectomy, intermediate clinical outcomes indicative of active disease occurred markedly less frequently among long-term survivors (Table 3). In the control cohort, disease progression was chiefly driven by the development of large splenomegaly (incidence rate, 16.03 events per 100 subject-years), followed by leukocytosis (14.16 per 100 subject-years), severe anemia (12.57 per 100 subject-years), and circulating CD34+ cell counts exceeding 100 × 10⁶/L (9.22 per 100 subject-years). Among long-term survivors, the progression pattern mirrored that of non-long-term survivors but with substantially attenuated incidence rates. Large splenomegaly remained the most frequent intermediate event (3.07 per 100 subject-years), followed by severe anemia (1.94 per 100 subject-years), leukocytosis (1.62 per 100 subject-years), and elevated CD34+ cell counts (1.46 per 100 subject-years). Only minimal proportions developed thrombocytopenia (1 per 100 subject-years) or leukopenia (0.75 per 100 subject-years).
Overall, long-term survivors ( ≥ 25 years) showed significantly lower incidence rates of all major intercurrent events compared with subjects who died earlier. Among all intercurrent events, BT displayed the most pronounced difference, with a markedly higher incidence in non-long-term survivors (5.31 vs. 0.48 per 100 person-years), corresponding to an approximately 90% reduction in long-term survivors (IRR, 0.09: 95% CI, 0.04–0.21; p < 0.001).
The overall survival of the entire cohort had a median of 11.25 years (interquartile range, 5.8–21 years) (Supplemental Fig. S2). Among subjects with an available cause of death, BT was the leading cause (n = 72), followed by primary second neoplasia (n = 11), hemorrhage (n = 8), infection (n = 9), and renal insufficiency (n = 5). The distribution of causes of death did not differ significantly between cohorts: BT accounted for 67 deaths (38%) among non–long-term survivors and 5 of 9 deaths (55.5%) among long-term survivors (P = 0.29).
Incidence of thrombosis
Across the interval extending from 2 years before diagnosis through the last available follow-up, at least one thrombotic event occurred in 50 subjects (23%). Eight individuals experienced a second event, and one experienced a third. Over this observation window, the incidence rate was 1.85 events per 100 person-years (95% CI, 1.40–2.43). The crude cumulative incidence of thrombosis was significantly higher among long-term survivors compared with non-long-term survivors (41% vs. 22%, p < 0.001) (Table 4). In contrast, the incidence rate expressed per 100 person-years was higher in the non–long-term survivor cohort (2.20 per 100 person-years; 95% CI, 1.56–3.03) than in long-term survivors (1.36 per 100 person-years; 95% CI, 0.79–2.17; p < 0.001). This discrepancy reflects the markedly longer follow-up time accrued by long-term survivors, which dilutes the event rate despite a higher proportion of individuals ever experiencing thrombosis.
In the overall population, arterial thromboses and deep vein thromboses in typical sites each accounted for 12 events (21% of all thrombotic episodes). Most events (57%) consisted of venous thromboses in atypical sites, including portal vein thrombosis (n = 28), Budd–Chiari syndrome (n = 2), and cerebral sinus vein thrombosis (n = 2). Among long-term survivors, atypical-site deep venous thrombosis represented the most frequent thrombotic manifestation (22% vs. 13% in non–long-term survivors). However, given the limited number of events, this difference did not reach statistical significance (p = 0.15).
Discussion
This study delineates the biological and clinical profile of long-term survivors with PMF, with the aim of enriching current prognostic frameworks, particularly in relation to pre-PMF phenotypes that are frequently classified as low-risk by conventional scoring systems. A comparable analysis by Penna et al. examined determinants of short-term (5-year) and long-term ( ≥ 20-year) survival in PMF [9]. Despite methodological differences, most notably the definition of long survival ( > 25 years in our cohort) and the composition of the comparator group (subjects surviving ≤25 years in our cohort), the two studies yielded remarkably concordant findings. In both analyses, long-term survivors were significantly younger, more often female, enriched for CALR mutations, and less likely to present with anemia or thrombocytopenia. Furthermore, the prevalence of high-molecular-risk non-driver mutations and cytogenetic abnormalities did not differ between long survivors and controls, suggesting that these lesions do not meaningfully discriminate extreme long-term outcomes.
In the Penna cohort, multivariable logistic regression identified seven independent predictors of long survival: age ≤70 years, female sex, hemoglobin ≥10 g/dL (women) or ≥11 g/dL (men), leukocyte count ≤25 × 10⁹/L, platelet count ≥100 × 10⁹/L, circulating blasts ≤1%, and absence of constitutional symptoms. In contrast, our multivariable model isolated three covariates with independent prognostic value: age <50 years, bone marrow fibrosis grade-0, and platelet count >450 × 10⁹/L at diagnosis. Together, these variables delineate a distinct and readily identifiable phenotype that may be seamlessly integrated into clinical decision-making. Because the identification of long-term survivors may be affected by immortal-time bias, we validated these findings using a Cox regression model with landmark analysis, confirming the robustness of the associations.
Among the predictors of long survival, age <50 years emerged as the most powerful correlate with long-term survival. This observation is consistent with the universal inclusion of older age as an adverse factor in established PMF risk scores [10] and aligns with multiple reports documenting superior outcomes in younger PMF populations [11,12,13]. Several biological and clinical mechanisms plausibly account for this association. Younger individuals face fewer competing risks, making survival more directly reflective of disease biology. Inflammation represents a more clearly age-modulated biological axis in PMF. Elevated high-sensitivity C-reactive protein (hs-CRP), a marker of systemic inflammation, correlates strongly with older age and predicts disease progression [14]. Dysregulated iron homeostasis constitutes another age-dependent determinant of PMF prognosis. Increased circulating hepcidin, integrating signals from iron loading, inflammation, and ineffective erythropoiesis, has been associated with age >65 years and inferior survival, further reinforcing the notion that age shapes disease biology through multiple converging pathways [15].
Although age <50 years was the strongest determinant of long survival, a minority of long-term survivors were older than 50 years. Multivariable analysis revealed two additional features distinguishing these individuals: elevated platelet count and grade 0 bone marrow fibrosis. A favorable prognostic effect of thrombocytosis in PMF has been only rarely described; however, in cases of extreme thrombocytosis ( > 1000–1500 × 10⁹/L), improved survival has been reported [16]. Conversely, thrombocytopenia is a well-established adverse factor in the DIPSS-plus score [17]. Mechanistic insights from Varricchio et al. showed that mononuclear cells from thrombocytopenic MPN individuals generate megakaryocyte-biased stem cells that promote a pro-inflammatory megakaryocytic phenotype [18]. These findings support the interpretation that thrombocytosis in PMF reflects preserved megakaryopoiesis and relative absence of the inflammatory mechanisms driving thrombocytopenia, thereby contributing to a more favorable disease trajectory.
Grade-0 bone marrow fibrosis emerged as an additional independent determinant of prolonged survival. Bone marrow fibrosis in PMF represents the morphological imprint of a profoundly perturbed microenvironment, largely orchestrated by inflammatory and fibrogenic cytokines released by dysmorphic megakaryocytes. Although the prognostic significance of fibrosis remains debated, accumulating evidence indicates that incorporating fibrosis grade into established clinical scores improves survival prediction [19, 20]. What distinguishes our findings is the identification of grade-0 fibrosis as a marker of long-term survival, a feature not previously emphasized in prognostic models and not reflected in current diagnostic classifications, where grades 0 and 1 fibrosis are merged within the pre-PMF category. Consistent with this observation, the mere presence of a pre-fibrotic phenotype at diagnosis did not independently predict long-term survival.
Our analyses enabled us to define evolutionary patterns characteristic of long-term survivors. In non-long-term survivors, disease progression was dominated by the emergence of a myelodepletive phenotype, chiefly anemia and thrombocytopenia, accompanied by substantial splenomegaly. In contrast, long-term survivors developed pronounced splenomegaly while maintaining a disease profile defined by persistent myeloproliferation, with BT representing the predominant cause of death. This overarching myeloproliferative signature mirrors the phenotype described in adolescents and young adults with PMF, who typically exhibit higher LDH levels, elevated leukocyte counts, and more frequent splenomegaly, all indicative of sustained clonal proliferative drive [21].
Heterogeneity in thrombotic risk has also been documented in PMF, representing a clinically relevant source of morbidity. Prior studies indicate that overt PMF is associated with only a modest increase in post-diagnosis thrombosis (10.4%) [22], whereas pre-PMF paradoxically displays a substantially higher thrombotic burden both at diagnosis or in the preceding year (30.3%) and during follow-up (25.4%) [6]. In our cohort, the crude incidence of thrombotic events occurring from two years before diagnosis to the last follow-up was higher among long survivors. However, the time-adjusted incidence favored the non-long-term survivor population, consistent with the interpretation that thrombosis clusters with features of disease severity. This aligns with evidence that a primed, thrombo-inflammatory endothelium cooperates with clonal hematopoiesis to amplify thrombotic risk [23] and that the neutrophil-to-lymphocyte ratio, a surrogate marker of systemic inflammation, contributes to the heightened thrombotic propensity characteristic of PMF [24].
This study has limitations. Identification of predictors of long-term survival relied on retrospective data, inherently constrained by missing information, variability in documentation quality, and lack of standardized data collection. Although the predictive performance of the selected covariates was adequate, the preliminary exclusion of subjects who underwent transplantation or were lost to follow-up introduces the possibility of selection bias. Furthermore, potentially informative prognostic variables, such as circulating immature myeloid cells and peripheral blasts, were infrequently recorded in this historical cohort and should be incorporated into future analyses.
The principal contribution of this study lies in its ability to enrich existing prognostic frameworks for PMF by supplying additional, clinically actionable information that supports individualized decision-making. Yet, before these findings can be translated into routine practice, it is essential to demonstrate that the characteristics used to delineate an indolent PMF phenotype are reproducible and clinically meaningful in real-world settings. In particular, it must be shown that the predictors of long-term survival identify a biologically and clinically coherent subgroup whose disease trajectory is sufficiently uniform to inform risk stratification, guide therapeutic choices, and shape future investigative priorities.
Data availability
All data generated or analysed during this study are included in this article. Further inquiries can be directed to the corresponding author.
References
Arber DA, Orazi A, Hasserjian RP, Borowitz MJ, Calvo KR, Kvasnicka HM, et al. International consensus classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140:1200–28.
Khoury JD, Solary E, Abla O, Akkari Y, Alaggio R, Apperley JF, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia. 2022;36:1703–19.
Arber DA, Orazi A, Hasserjian R, Thiele J, Borowitz MJ, Le Beau MM, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127:2391–405.
Cervantes F, Dupriez B, Pereira A, Passamonti F, Reilly JT, Morra E, et al. New prognostic scoring system for primary myelofibrosis based on a study of the International Working Group for Myelofibrosis Research and Treatment. Blood. 2009;113:2895–901.
Guglielmelli P, Pacilli A, Rotunno G, Rumi E, Rosti V, Delaini F, et al. Presentation and outcome of patients with 2016 WHO diagnosis of prefibrotic and overt primary myelofibrosis. Blood. 2017;129:3227–36.
Barbui T, Thiele J, Passamonti F, Rumi E, Boveri E, Ruggeri M, et al. Survival and disease progression in essential thrombocythemia are significantly influenced by accurate morphologic diagnosis: an international study. J Clin Oncol. 2011;29:3179–84.
Coltro G, Mannelli F, Loscocco GG, Mannarelli C, Rotunno G, Maccari C, et al. Differential prognostic impact of cytopenic phenotype in prefibrotic vs overt primary myelofibrosis. Blood Cancer J. 2022;12:116.
Carobbio A, Guglielmelli P, Rumi E, Cavalloni C, De Stefano V, Betti S, et al. A multistate model of survival prediction and event monitoring in prefibrotic myelofibrosis. Blood Cancer J. 2020;10:100.
Penna D, Lasho TL, Finke CM, Vallapureddy RR, Hanson CA, Ketterling RP, et al. 20+ Years and alive with primary myelofibrosis: Phenotypic signature of very long-lived patients. Am J Hematol. 2019;94:286–90.
Mora B, Bucelli C, Cattaneo D, Bellani V, Versino F, Barbullushi K, et al. Prognostic and predictive models in myelofibrosis. Curr Hematol Malig Rep. 2024;19:223–35.
Boddu P, Masarova L, Verstovsek S, Strati P, Kantarjian H, Cortes J, et al. Patient characteristics and outcomes in adolescents and young adults with classical Philadelphia chromosome-negative myeloproliferative neoplasms. Ann Hematol. 2018;97:109–21.
Szuber N, Vallapureddy RR, Penna D, Lasho TL, Finke C, Hanson CA, et al. Myeloproliferative neoplasms in the young: Mayo Clinic experience with 361 patients age 40 years or younger. Am J Hematol. 2018;93:1474–84.
Goulart H, Masarova L, Mesa R, Harrison C, Kiladjian JJ, Pemmaraju N. Myeloproliferative neoplasms in the adolescent and young adult population: a comprehensive review of the literature. Br J Haematol. 2024;205:48–60.
Barosi G, Massa M, Campanelli R, Fois G, Catarsi P, Viarengo G, et al. Primary myelofibrosis: older age and high JAK2V617F allele burden are associated with elevated plasma high-sensitivity C-reactive protein levels and a phenotype of progressive disease. Leuk Res. 2017;60:18–23.
Pardanani A, Finke C, Abdelrahman RA, Lasho TL, Tefferi A. Associations and prognostic interactions between circulating levels of hepcidin, ferritin and inflammatory cytokines in primary myelofibrosis. Am J Hematol. 2013;88:312–6.
Lucijanic M, Krecak I, Soric E, Sabljic A, Galusic D, Holik H, et al. Extreme thrombocytosis in patients with overt myelofibrosis and its clinical associations. Cancers. 2025;17:1390.
Gangat N, Caramazza D, Vaidya R, George G, Begna K, Schwager S, et al. DIPSS plus: a refined Dynamic International Prognostic Scoring System for primary myelofibrosis that incorporates prognostic information from karyotype, platelet count, and transfusion status. J Clin Oncol. 2011;29:392–7.
Varricchio L, Mosoyan G, Elghaity-Beckley S, Mia MB, Handa S, Salib C, et al. Thrombocytopenia in myelofibrosis is characterized by inflammatory megakaryocytes with reduced G6B expression. Blood. 2025;146:1612–24.
Gianelli U, Fiori S, Cattaneo D, Bossi A, Cortinovis I, Bonometti A, et al. Prognostic significance of a comprehensive histological evaluation of reticulin fibrosis, collagen deposition and osteosclerosis in primary myelofibrosis patients. Histopathology. 2017;71:897–908.
Guglielmelli P, Rotunno G, Pacilli A, Rumi E, Rosti V, Delaini F, et al. Prognostic impact of bone marrow fibrosis in primary myelofibrosis. A study of the AGIMM group on 490 patients. Am J Hematol. 2016;91:918–22.
Elkourashy S, Soliman D, Chandra P, Taha R, Elomri H. Primary myelofibrosis in young patients ≤ 50 years: clinicopathological characteristics, genetic landscape and disease outcome. A single-center experience. Hemasphere. 2022;6:3550–1.
Barbui T, Ghirardi A, Carobbio A, De Stefano V, Rambaldi A, Tefferi A, et al. Thrombosis in myeloproliferative neoplasms: a viewpoint on its impact on myelofibrosis, mortality, and solid tumors. Blood Cancer J. 2024;14:188.
Abosabie SAS, Boye-Doe A, Ali M, Podoltsev N, Stegner D, Mendez LM, et al. Thrombo-inflammatory endothelial signatures in JAK2-mutated myeloproliferative neoplasms. bioRxiv. 2026;16:706250.
Cavalca F, Latagliata R, Pugliese N, Palumbo GA, Polverelli N, Musto P, et al. Neutrophil to lymphocyte ratio at diagnosis predicts venous thrombosis in prefibrotic primary myelofibrosis: results from a multicenter cooperative study. Ann Hematol. 2026;105:19.
Funding
VR discloses support for the research of this work from the AIRC 5 × 1000 call “Metastatic disease: the key unmet need in oncology” to the MYNERVA project, #21267 (MYeloid Research Venture AIRC), and the AIRC Individual Grant 2024, project #31013. GB, ADS, and RPG declare no relevant funding.
Author information
Authors and Affiliations
Center for the Study of Myelofibrosis, Fondazione Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Policlinico San Matteo, Pavia, Italy
Giovanni Barosi & Vittorio Rosti
Biostatistics and Clinical Epidemiology, Fondazione IRCCS Policlinico San Matteo, Pavia, Italy
Annalisa De Silvestri
Centre for Haematology, Department of Immunology and Inflammation, Imperial College of Science, Technology and Medicine, London, UK
Robert Peter Gale
Authors
- Giovanni Barosi
- Vittorio Rosti
- Annalisa De Silvestri
- Robert Peter Gale
Contributions
GB designed the study, analyzed the data, and wrote the first version of the manuscript. VR and RPG contributed to writing the manuscript. ADS helped with the statistical analysis. All authors have read and agreed to the published version of the manuscript.
Corresponding author
Correspondence to Giovanni Barosi.
Ethics declarations
Competing interests
Robert Peter Gale is a consultant to BeiGene Ltd., Fusion Pharma LLC, LaJolla NanoMedical Inc., Mingsight Parmaceuticals Inc., Kite Pharma and CStone Pharmaceuticals; Advisor to Antegene Biotech LLC, Medical Director and FFF Enterprises Inc.; Partner in AZACA Inc.; Board of Directors, RakFond Foundation for Cancer Research Support; Scientific Advisory Board, StemRad Ltd.
Ethics
The research was conducted in accordance with the World Medical Association Declaration of Helsinki. Owing to the retrospective design of the study, the requirement for individual consent was waived by the Ethics Committee, provided that all subject-level data were fully anonymized prior to the analysis.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Barosi, G., Rosti, V., De Silvestri, A. et al. Predicting long-term survival in primary myelofibrosis. Leukemia (2026). https://doi.org/10.1038/s41375-026-03177-w
Received: 02 July 2026
Revised: 25 September 2026
Accepted: 07 October 2026
Published: 19 October 2026
Version of record: 19 October 2026
DOI: https://doi.org/10.1038/s41375-026-03177-w
来源:Nature 健康护理主题 · nature.com