Blood CD34 Count in Myelofibrosis with Myeloid Metaplasia: A Prospective Evaluation of Prognostic Value in 94 Patients.

Background: Previously established adverse prognostic factors in myelofibrosis with myeloid metaplasia (MMM) include anemia (hemoglobin < 10 g/dL), leukocytosis or leukopenia, presence of circulating blasts, and hypercatabolic symptoms (Dupriez et al. 1996;88:1013,

Methods: A concurrent evaluation of PB CD34 count, bone marrow histology, and other clinical and laboratory parameters was conducted in a consecutive cohort of both newly and previously diagnosed patients with MMM. The study patients were prospectively followed for a minimum of 3 years in order to determine the prognostic value of PB CD34 count on survival, leukemic transformation, and disease progression from the time of PB CD34 analysis.

Results:

i. Patient cohort: A total of 94 patients (median age 61 years, range 26–81; 61 males) were accrued between January 2000 and August 2001 and subsequently followed through July 2004. MMM subtypes included agnogenic myeloid metaplasia (AMM; 58 patients), post-polycythemic myeloid metaplasia (PPMM; 12 patients), post-thrombocythemic myeloid metaplasia (PTMM; 12 patients), cellular phase MMM (8 patients) and hypocellular variant MMM (4 patients). At the time of study entry, 19 patients (20%) were newly diagnosed and 75 were established cases (median of 36 months from initial diagnosis, range 4–221). The Dupriez prognostic scores were 0 in 37 (39%), 1 in 37 (39%), and 2 in 20 patients (21%).

ii) Clinical correlates of PB CD34 count: The median PB CD34 count for the entire study cohort was 54.7 x 10⁶/L (range 0–5345) with 85% patients having increased values (> 5 x 10⁶/L). Cellular phase MMM and PPMM had significantly lower PB CD34 counts compared to the other MMM subtypes. On univariate analysis, PB CD34 count significantly correlated with Dupriez prognostic score, leukocyte count, circulating blast or immature myeloid cell percentage, and presence of hypercatabolic symptoms but not with age, gender, anemia (hemoglobin < 10 g/dL), abnormal cytogenetics, disease duration, treatment history, splenectomy status, nucleated RBC count (NRBC), platelet count, or bone marrow histology including fibrosis, osteosclerosis, and angiogenesis. On multivariate analysis, only leukocyte count and PB blast or immature myeloid cell percentage retained their significance.

iii) Prognostic evaluation of PB CD34 count: As of July 16, 2004, 43 patients (46%) have died and median duration of follow up in the surviving patients was 41 months (range 36–54). The projected overall survival at 5 years was 50%. On Cox proportional regression analysis, individual variables that correlated with inferior survival included anemia (hemoglobin < 10 g/dL), advanced age, presence of hypercatabolic symptoms, abnormal cytogenetics, and increased leukocyte count, circulating blast percentage, Dupriez prognostic score, and NRBC. PB CD34 count displayed borderline significance (p=0.06). However, on multivariate analysis, only anemia, leukocytosis, and age retained prognostic significance. Similarly, PB CD34 count did not predict either subsequent leukemic transformation or disease progression.

Conclusion: PB CD34 count is often but not always increased in MMM and correlates significantly with leukocytosis and degree of PB myeloid immaturity. The current prospective study does not support an independent prognostic value for PB CD34 count in MMM.