Clinical Implications of Acute Myeloid Leukemia Presenting as Granulocytic Sarcoma

Granulocytic sarcoma (GS), also known as chloroma, is described as an extramedullary tumor composed of immature myeloid cells. It has been reported to develop in 2–8% of patients with acute myeloid leukemia (AML) and can occur prior to, concomitantly with, or following the diagnosis of AML. Rarely patients present with GS as an isolated mass without evidence of AML (Byrd JC, J Clin Oncol. 1997). Due to the rarity of this disorder, large series of patients are seldom reported and the prognosis and optimal treatment of patients presenting with GS are not clear.

In this retrospective study we analyzed the presenting characteristics, treatments and overall survival of all patients presenting with isolated granulocytic sarcoma (GS) or GS with concomitant acute myeloid leukemia (AML) at presentation and compared them to all AML patients, treated at our institution during the same period.

We identified cases who were diagnosed as having GS (with or without bone marrow involvement by AML) and cases of AML (without evidence GS at diagnosis) using ICD-9 codes in the hospitalization data base of the Hadassah Medical Center between 1990 and 2005. We excluded patients with GS at the time of relapse. All GS cases were biopsy-proven.

The study population consisted of 19 GS and 235 AML non-GS patients. The median age of these patients was 41 and 48 years, respectively. There was no statistically significant difference between the groups regarding gender, age, cytogenetic risk groups, rate of complete remission (CR), number of cycles of chemotherapy needed to achieve CR and rate of first relapse. Hematopoietic stem cell transplant (HSCT) (either autologous or allogeneic) was performed in 10 of 19 patients (52.6%) in the GS group and in 66 of 235 (28.1%) in the AML group (p=0.025). The overall survival in the GS group (median 16 months) was not significantly different (p= 0.60) from the AML group. The median time to death in subjects who had radiotherapy (6/19) was identical (median 21 months, p=0.79) to that of subjects who did not receive radiotherapy. Transplantation was associated with prolonged survival in both GS and AML groups (p=0.018 and p<0.0001 respectively). At the end of the follow up, 4 patients in the GS group who underwent HSCT were alive, compared to none in the group who did not undergo HSCT. Karyotype was found to be a prognostic factor in both the GS and AML groups (p=0.0034 and p<0.0001 respectively). In a multivariate analysis there was no statistically significant difference in the risk of death between subjects in the AML and the GS group; hazard ratio (HR) = 0.83, 95% CI (0.456-1.516), p=0.55, after controlling for age, karyotype and transplantation. As expected, in the model, age less than 47.5 years (HR=0.646 (p=0.0022)) and favorable and intermediate karyotype (HR=0.13 (p<0.0001) and HR=0.44 (p<0.0001) respectively, compared to unfavorable) were associated with a lower risk of death. Subjects who did not undergo transplantation had an increased risk of death compared with subjects who underwent the procedure (HR=1.88, p=0.0023) (Fig. 1).

To the best of our knowledge this is the first retrospective series of GS patients with concomitant AML at diagnosis compared to all other AML patients in one medical center. The addition of radiotherapy as a treatment modality for GS patients did not appear to change survival. Patients with GS at diagnosis might benefit from upfront aggressive treatment with HSCT. Due to the rarity of this disorder and the remaining open questions, a prospective multi-center study is necessary to address these issues.

Fig 1:

View largeDownload slide

Time to death in subjects who underwent HSCT (median time: 12 [3.0 – 17.0] months) was significantly longer than in subjects who did not undergo HSCT (median time: 27 [16.0 –NA] months). p-value: 0.02 (log-rank test).

Fig 1:

View largeDownload slide

Time to death in subjects who underwent HSCT (median time: 12 [3.0 – 17.0] months) was significantly longer than in subjects who did not undergo HSCT (median time: 27 [16.0 –NA] months). p-value: 0.02 (log-rank test).

Close modal

No relevant conflicts of interest to declare.