Leukemic transformation of polycythemia vera

A single center study of 23 patients

Authors


Abstract

BACKGROUND

Acute leukemia (AL) may occur as rare and late event of polycythemia vera (PV).

METHODS

The current study included 23 patients who developed acute leukemia in a cohort of 414 consecutive PV patients with long-term observation (3208 person years of follow-up). Kaplan–Meier Product-Limit method was used to estimate the cumulative probability of survival; Gehan–Wilcoxon test was applied to compare survival in different groups of patients.

RESULTS

Median age was 68 years, and 18 patients (78%) were > 60 years of age. At diagnosis of AL, most patients had a white blood count > 10 × 109/L (n = 17; 74%), Hgb < 10 g/dL (n = 13; 57%), and platelet count > 50 × 109/L (n = 17; 74%). Of 14 patients in whom cytogenetic analysis was available at leukemic transformation, 12 showed high-risk abnormalities including complex karyotype (n = 10), del (7)(q22) sole (n = 1) and del (X)(q26) sole (n = 1), whereas 2 had a normal karyotype. In patients whose karyotype was available at diagnosis of PV, cytogenetic evolution was documented at progression to AL. Treatment consisted of supportive care and/or low-dose chemotherapy (n = 15), or induction chemotherapy (n = 8). This included idarubicin plus cytarabine (n = 3), high-dose cytarabine (n = 4), and fludarabine-based regimen (n = 1). Allogenic stem cell transplantation was offered to a single patient, who is alive at Day + 70. The outcome of patients was poor, with a median survival of 2.9 months (range, 0.6–20.1 mos), with no significant differences between palliation and intensive treatments.

CONCLUSIONS

AL following PV has distinct clinical and biologic features. Outcome of patients is poor irrespective of the treatment employed. Cancer 2005. © 2005 American Cancer Society.

Polycythemia vera (PV), with essential thrombocythemia (ET) and myelofibrosis with myeloid metaplasia (MMM), belongs to the group of Philadelphia-negative chronic myeloproliferative disorders.1, 2 Acute leukemia (AL) may occur as a late event in all these disorders.3 In our prior study on PV, the incidence of AL was 5.3 × 1000 person years.4 Studies with long-term follow-up indicate that transformation to AL is part of the natural evolution of PV.5, 6 Myelosuppressive agents such as hydroxyurea or pipobroman, when used as the sole treatment, seem not to increase the natural risk of AL,4, 6 but the risk increases with sequential use of different cytotoxic agents.4 The incidence of AL is lower in ET (1.2 × 1000 person years) than in PV,4 whereas MMM has the highest risk of transformation among these disorders.7 The poor prognosis of patients with AL following MMM has been recently described.7

Studies providing clinical information on clinical features and treatment outcome of AL following PV are limited,8, 9 with only few studies reporting cytogenetic findings.10, 11 In the current study, we describe presenting characteristics and outcomes of 23 patients who developed acute leukemia in a cohort of 414 consecutive PV patients under long-term observation.

MATERIALS AND METHODS

Patients

The current study is focused on the 23 patients who developed AL within a cohort of 414 consecutive PV patients. Leukemia was diagnosed between 1981 and November 2004 at the Division of Hematology, San Matteo Polyclinico of the University of Pavia, Italy. The study was conducted in accordance with the institutional guidelines established for retrospective studies. Diagnosis of PV was made according to criteria in use at the time of first observation.12–14 Diagnosis of AL was according to World Health Organization (WHO) criteria, with 20% blast-cell threshold for diagnosis.1 The morphologic classification of blast cells was according to French–American–British (FAB) criteria.15 We categorized treatments for leukemic transformation as follows: palliation (supportive care only, or low-intensity chemotherapy) or induction chemotherapy (de novo leukemia-like therapy aimed at remission). Complete remission (CR) was defined by National Cancer Institute criteria and required bone marrow cellularity > 20% with < 5% blast cells and normal myeloid maturation.16

Statistical Analysis

Continuous variables are summarized by median and range; categoric variables are reported as count and relative frequency. Survival analysis was calculated from the onset of AL, considering death for all causes as endpoint. Kaplan–Meier Product-Limit method was used to estimate the cumulative probability of survival; Gehan–Wilcoxon test was applied to compare survival in different groups of patients. The incidence rate (IR) of AL, with the corresponding 95% confidence interval (CI), was estimated using a Gaussian approximation to the Poisson log likelihood. All computations were carried out using STATISTICA for Windows 6.1 (StatSoft, Tulsa, OK) and Microsoft Excel 2000) Microsoft, Redmond, WA).

RESULTS

From 1970 to December 2004, 414 consecutive patients with PV (3208 person years of follow-up) were diagnosed and followed at the San Matteo Polyclinico of the University of Pavia, Italy. Among these 414 patients, 23 progressed to AL after a median time of 12.8 years from diagnosis of PV (range, 1.5–22.8 yrs). The incidence of AL post-PV was 7.1 × 1000 person years.

Clinical and Hematologic Findings at Leukemic Transformation

Treatment of PV consisted of pipobroman alone (n = 14; 61%), sequential use of 2 or more myelosuppressive agents (n = 6; 26%: 4 with pipobroman and busulphan, 2 with pipobroman and hydroxyurea), phlebotomy and pipobroman (n = 2; 9%), phlebotomy alone (n = 1; 4%). The clinical characteristics of patients at diagnosis of AL are shown in Table 1. The median age was 68 years, and 18 (78%) patients were > 60 years of age. Hemorrhagic symptoms were present in 2 (9%) patients, and fever in 8 (34%). At diagnosis of AL, most patients had a white blood count (WBC) > 10 × 109/L (n = 17; 74%), hemoglobin (Hgb) < 10 g/dL (n = 13; 57%), and platelet count > 50 × 109/L (n = 17; 74%). Twelve (52%) patients had splenomegaly; at diagnosis of PV, 5 already had splenomegaly. All cases had an abrupt onset. Leukemia was of myeloid origin in 22 out of 23 (96%) patients. The most frequent subtype was M0-M1 (Table 1). One patient had acute lymphoid leukemia 47 months after PV diagnosis (L2 according to FAB classification, with B phenotype). Two (9%) patients who evolved in AL had overt myelofibrosis 12 and 25 months before leukemic transformation, respectively. Another patient with moderate splenomegaly showed only focal reticulin fibrosis at bone marrow histology (1 year before AL), without allowing a definitive diagnosis of myelofibrosis. Before leukemic transformation, three patients with splenomegaly received bone marrow biopsy that showed hypoplasia without evidence of reticulin fibrosis or megakaryocyte abnormalities.

Table 1. Characteristics of 23 Patients with Acute Leukemia Evolved from Polycythemia Vera
 No. (%)
No. patients23
Male/female12/11
Median age in yrs (range, 46–84)68
Hemoglobin in g/dL; median, 9.2; range, 7–14.2 
 <8 2 (9)
 8–1011 (48)
 >1010 (43)
WBC count in × 109/L; median, 15; range, 1.4–83 
 <7 6 (26)
 7–3013 (56)
 >30 4 (18)
Platelet count in × 109/L; median, 63; range, 10–300 
 <50 6 (26)
 50–10012 (52)
 >100 5 (22)
French–American–British classification 
 M0–M1 8 (36)
 M2 6 (27)
 M4 5 (23)
 M5 2 (9)
 M6 1 (5)

Cytogenetic Findings at Leukemic Transformation

At the time of AL, cytogenetic analysis was available in 14 (61%) patients, whereas 3 patients (13%) showed insufficient number of metaphases. In 6 (26%), the test was not performed. Karyotype aberrations (balanced rearrangement, loss or gain of chromosome or its part) involved the following chromosomes: 1, 2, 4, 5, 6, 7, 8, 9, 10, 12, 13, 14, 16, 17, 19, 20, and X (Table 2). According to de novo AL risk stratification,17 12 of 14 (86%) patients showed high-risk abnormalities including complex karyotype (n = 10), del (7)(q22) sole (n = 1), and del (X)(q26) sole (n = 1). Two (14%) patients showed a normal karyotype. For three AL patients, a cytogenetic study was performed at the time of diagnosis of PV. At the onset of AL, 2 of them progressed from normal cytogenetics to complex karyotype (n = 1), or del 7 (n = 1). The latter patient was treated with phlebotomy alone. The third patient had t(1;7)(p11;p11) at diagnosis of PV and maintained the same translocation inside a complex karyotype at transformation.

Table 2. Karyotype Abnormalities in 14 Patients with Acute Leukemia Evolved from Polycythemia Vera
Cytogenetic findingNo. (%)
Normal karyotype 2 (14)
Single structural abnormality 2 (14)
Complex karyotype10 (72)
 Involvement of chromosome 55
 Involvement of chromosome 77

Treatment Outcome of AL

Treatment of AL was categorized on the basis of the aim of treatment as palliation or induction of remission. The choice of treatment was based upon the individual patient according to performance status and comorbidities (Table 3). Of 15 patients assigned to palliation, 14 (93%) died, and 1 (7%) is alive 4 months after diagnosis of AL. The median survival was 2.5 months (range, 0.6–20.1 mos). The causes of death were as follows: infection in 9 (64%), multiorgan failure in 3 (22%), hemorrhage in 2 (14%). Of 8 patients treated with induction chemotherapy, only 1 (13%) obtained a complete hematologic response, whereas 7 (87%) died without obtaining a response. Among the latter, one patient obtained a transient reduction of splenomegaly and two patients a transient clearance of marrow blast cells. The median survival was 5.6 months (range, 0.7–8.3 mos). The cause of death was infection in 5 (71%) and hemorrhage in 2 (29%). The single patient (51 yrs of age) obtaining CR is still alive at Day + 70 after unrelated donor bone marrow transplantation.

Table 3. Treatment of 23 Patients with Acute Leukemia Evolved from Polycythemia Vera According to the End-Point of Therapy
TreatmentNo. (%)
  1. p.o.: orally; i.v.: intravenous; RIV: rapid intravenous infusion; CIV: continuos intravenous infusion

End point: palliation15 (65)
 Supportive care alone7
 6-thyoguanine (100 mg/m2 p.o., Days 1–7)4
 Cytarabine (100 mg/m2 RIV, Days 1–5)4
End point: induction of remission 8 (35)
 Idarubicin (10 mg/m2 i.v. Days 1–3), plus cytarabine (100 mg/m2 24-hr CIV, Days 1–7)3
 Idarubicin (5 mg/m2 i.v. Day 1) plus cytarabine (800 mg/m2/12 hrs, 2-hr CIV, Days 1–5)3
 Idarubicin (6 mg/m2 Day 1), plus cytarabine (600 mg/m2/12 hrs, 2-hr CIV, Days 1–3), plus etoposide (150 mg/m2, 1-hr CIV, Days 1–3)1
 Idarubicin (10 mg/m2/day, Days 1–2), cytarabine (1000 mg/m2/12 hrs, 2-hr CIV, Days 1–5), fludarabine (15 mg/m2/12 hrs, Days 1–5)1

As of December 2004, 21 out of 23 (91%) patients with AL had died. Median survival (Fig. 1) was 2.9 months (range, 0.6–20.1 mos). Survival was not significantly different for patients receiving palliation (median, 2.5 mos; range, 0.6–20.1 mos) or induction chemotherapy (median, 5.6 mos; range, 0.7–8.3 mos) (Fig. 2).

Figure 1.

Survival estimate of 23 patients with acute leukemia following polycythemia vera.

Figure 2.

Survival estimate of 23 patients with acute leukemia following polycythemia vera according to treatmentas follows: induction chemotherapy (de novo leukemia-like therapy aimed at remission) and palliation (supportive care only or low-intensity chemotherapy).

DISCUSSION

Transformation to acute leukemia is a rare event in polycythemia vera, with an estimated incidence of about 5 × 1000 person years.4, 6 Prognostic factors at diagnosis of PV that could be used to predict leukemic transformation are lacking. Leukemia is part of the natural evolution of the disease but may be favored by the use of cytotoxic myelosuppressive agents. The issue of the leukemogenic potential of different treatments in PV has been addressed in two large studies with a long follow-up.4, 6 Both studies, although not randomized, reassure the leukemogenic potential of hydroxyurea. In addition, a study from our group4 on patients exclusively treated with pipobroman shows that the risk of AL, when this agent is used alone, is not different from that of hydroxyurea.

To define the hematologic features and outcome of AL following PV, we studied 23 patients with acute transformation from a cohort of 414 patients with PV (3208 person years of follow-up). In our series, 22 out of 23 patients showed a myeloid phenotype. Only one patient had acute lymphoid leukemia, and his clinical picture at diagnosis and outcome parallel those of other PV patients with lymphoid transformation.18 During PV, lymphoid transformation is unusual,18 and its correlation with treatments is unlikely;19 probably it may be considered a de novo event.

The current study shows a consistently poor outcome with conventional induction chemotherapy in patients with AL post-PV. It may result from distinct clinical and biologic features of this type of AL, namely, relatively old age (median, 68 yrs), and unfavorable cytogenetics. In fact, elderly patients often have comorbidities that are exacerbated by chemotherapy,20 resulting in a reduced tolerance of antileukemia treatments. Patients developed post-PV AL after a median time from diagnosis of 12.8 years, which implied a long lasting exposure to myelosuppressive agents. This could induce a greater susceptibility to infections at the time of AL transformation.

Concerning cytogenetics, patients with post-PV AL typically showed unfavorable karyotype. This may reflect the high median age of patients. In fact, elderly patients with de novo acute myeloid leukemia have more frequently unfavorable cytogenetics such as abnormalities of chromosomes 5 or 7, or complex karyotype.21 These represent the most frequent abnormalities observed in patients with AL post-PV,11, 22, 23 as well as in our cohort of patients. Abnormalities involving chromosome 5, 7 and 17 are considered therapy-related in patients with AL evolved from chronic myeloproliferative disorders. In fact, monosomy or deletion of chromosome 5 or 7 are commonly observed in patients developing AL following treatment with alkylating agents,24 whereas 17p- has been observed following hydroxyurea.25 However, in a rare familial platelet disorder with propensity to develop leukemia,26 abnormalities of chromosomes 5 or 7 are the most frequent karyotype associated with leukemic transformation, suggesting the involvement of second mutation independent from therapy. In the current study, the role of second mutation in leukemic transformation of PV is supported by the sequential analysis of three patients, whose karyotype was available at diagnosis of PV. A karyotype evolution was documented in all three patients, comprising a patient treated with phlebotomy alone.

Clinical and biologic features of patients with AL post-PV may explain the poor outcome after induction chemotherapy, with only 1 out of 8 patients (13%) obtaining a CR. On the contrary, about 50% of elderly patients with de novo AL achieve CR with standard induction chemotherapy.20 As a rule in this setting of patients, chemotherapy is generally based upon the individual patient.7 In this study, remission induction regimens included standard-dose cytarabine plus idarubicin, high-dose cytarabine, or a fludarabine-based regimen. This study does not permit definitive conclusions concerning the best treatment for AL post-PV. However, patients receiving intensive regimens had a higher rate of death (29%) within 1 month than patients receiving palliation (14%). In addition, intensive regimens did not confer a significant increase in survival. Allogenic stem cell transplantation was offered to the single patient achieving CR after induction. This patient is alive and well at Day + 70 after transplant, indicating that transplantation might be the treatment of choice for rare patients who achieve CR after induction chemotherapy. Few data are available on stem cell transplantation in patients with AL post-PV. In the Seattle experience on 10 patients with AL post-PV or ET, all patients died of transplant-related complications.27 Allogenic stem cell transplantation with reduced-intensity conditioning regimens may be an option for these patients, as demonstrated in MMM.28

In conclusion, the current study indicates that AL evolved from PV has distinct clinical and biologic features. The outcome of patients is poor whatever the treatment used. Stem cell transplantation remains an option for rare patients who achieve CR.

Ancillary