Bone Marrow Reticulin Fibrosis in 579 Patients with Polycythemia Vera and Essential Thrombocythemia: Effect on Clinical Outcome

Background:

Prognostic value of bone marrow (BM) fibrosis grading in myeloproliferative neoplasm (MPN) is still debated. Polycythemia Vera (PV) and Essential Thrombocythemia (ET) are long term outcome MPN; however, they could evolve to adverse secondary myelofibrosis-MF or acute leukemia-AL.

Aims and Methods:

We retrospectively analyzeda cohort of 579 World Health Organization-defined PV (n=180) and ET (n=399) patients, and examined the prevalence and prognostic relevance of BM reticulin fibrosis.

All patients were diagnosed between 1990 and 2013 and were recruited in Turin (n= 436) and Bologna (n=143), Italy. BM biopsy sample were reviewed by local pathologist and fiber scoring was performed according to a 3-graded system. Eligibility criteria included the availability of BM samples at diagnosis. Patients with grade 2 or 3-fibrosis were excluded.

Overall survival (OS) was evaluated from diagnosis to death using Kaplan Meyer method and Hazard Ratio were estimated with the Cox Model. Cumulative incidence of MF and AL evolution were estimated considering death from any cause as a competing event and compared between groups using the Gray's test.

Results:

Overall, we observed 115 (63%) grade 0 and 65 (36%) grade 1 fibrosis and 291 (72%) grade 0 and 108 (27%) grade 1 among PV and ET patients, respectively (p= 0.028) We analyzed effect on clinical outcome separately.

PV

With a median follow up of 110 months (IQR:70-170), 5 and 10-years OS were 96% and 87%, respectively.

Stratifying patients based on fibrosis degree, we observed 15 (13%) and 16 (25%) deaths for grade 0 and 1 fibrosis respectively, with 5 and 10-years OS of 98% vs 90% and 92% vs 82% for grade 0 and grade 1, respectively (p 0.076). Neither clinical findings nor thrombosis were significantly different between fibrosis degree. JAK2 V617F or exon 12 mutation status and allele burden was similar into subgroup. Cumulative incidence of MF evolution at 5 and 10-years was 2,8% and 7,2% vs 3,8% and 18,7% for grade 0 and grade 1 respectively (p 0.123). Cumulative incidence of AL evolution at 10- year was 4,2 % for both grade whereas at 15-years was 4,2% vs 19% for grade 0 and 1, respectively.

ET

At a median follow up of 75 months (IQR:39-120), 5 and 10-years OS were 98% and 90%, respectively.

We observed 22 (8%) and 16 (55%) deaths for grade 0 and 1 respectively, with 5 and 10-years OS of 98% and 90% for grade 0 vs 97% and 89% for grade 1, respectively (p 0.358).

The mutation status was analyzed in 379 patients and showed: 62% JAK2V617F, 19% CALR (type-1/1-like 14% and type2/2-like 5%), 3% MPLW515, 63 patients were triple negative for the above mutations.

During follow-up, patients with grade 0 fibrosis showed more thrombotic events, 41 cases (14%; the 71% JAK2V617F-positive) vs 17 (16%). At multivariate analysis MPL mutation showed a higher risk of MF evolution compared to triple negative with an HR of 5,8 (p 0.0014) and to JAK2V617 mutation with an HR of 9,5 (p 0.002).

Cumulative incidence of MF evolution was at 5-years 0,5% and 9% and at 10-years 6,2% and 18% for grade 0 and 1, respectively (p 0.0001). Cumulative incidence of AL evolution at 5 and 10-years was 0% for grade 0 and 13% and 7,3% for grade 1, respectively (p 0.096). However grade 0 showed a higher cumulative risk of AL evolution at 15-years (6,9% vs 10% for grade 0 and 1, respectively) (p=0.096).

Conclusion:

According to data recently published, in ET patients grade 1 BM fibrosis seems to correlate with a higher cumulative risk of MF and AL evolution, whereas in PV patients seems to correlate to a trend of higher mortality, even if not statistically significantly.

Data reaffirm the importance of BM examination as part of diagnostic criteria in all MPN.