Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Outcome in patients with nonleukemic granulocytic sarcoma treated with chemotherapy with or without radiotherapy


Granulocytic sarcoma (GS) is an extramedullary tumor composed of immature myeloid cells. The objectives of this study were to describe the frequency, presenting characteristics, and survival in patients with nonleukemic GS by conducting a review of all untreated patients presenting to the MD Anderson Cancer Center between January 1990 and June 2002. In all, 21 patients with nonleukemic GS, 1520 patients with acute myeloid leukemia (AML), and 402 patients with high-risk myelodysplastic syndrome (MDS) were identified. GS occurred in 1.4% of patients with AML, and 1.1% of patients with AML or high-risk MDSs. The median patient age was 57 years (range, 7–81). Among 20 patients with available cytogenetics in tissue and/or bone marrow, six had chromosome 8 abnormalities. The median follow-up of surviving patients is 12 months (range, 7–75). In all, 20 patients were treated. Patients were treated with AML-type chemotherapy (n=16), chemotherapy and radiotherapy (n=3), or radiotherapy alone (n=1). A total of 13 patients (65%) achieved complete remission and one patient (5%) achieved partial remission. The median overall survival was 20 months (range, 1–75), median overall failure–free survival was 12 months (range, 1–75). The median survival of patients with chromosome 8 abnormalities was 12 months compared with 40 months of those without (P=0.17). Novel therapies for patients with GS are required.


Granulocytic sarcoma (GS) or chloroma is an extramedullary tumor composed of immature myeloid cells.1 The latter term derives from the Greek word χλωρóς (green), describing the typical green appearance of myeloblasts on myeloperoxidase staining.2 GS was first used to describe an extramedullary solid destructive tumor mass composed of immature cells of the granulocytic series,3 but was subsequently used to describe all forms of extramedullary myeloid leukemic infiltrates, the so-called extramedullary myeloid tumor.4 GS most commonly presents concurrently with, or after, the diagnosis of chronic myeloproliferative disorders (MPD) or acute myelogenous leukemia (AML). It may also present without overt hematologic disease, that is, normal marrow and no history of AML or MPD.5 Many of these cases are misdiagnosed as lymphoma.6,7 Patients with nonleukemic GS may eventually develop acute leukemia.7,8 Owing to the rarity of the disease, prospective studies to determine the best induction and maintenance therapy would be difficult in terms of patient accrual, and therefore therapeutic decisions are based on retrospective studies and case reports. We report on the incidence, presentation and the clinical outcome of patients with nonleukemic GS, who were treated mainly with chemotherapy and/or radiation therapy.

Patients and methods

The database of the Department of Pathology of the University of Texas, MD Anderson Cancer Center (MDACC) was searched for patients presenting with extramedullary myeloid diseases from January 1990 to June 2002. This database includes patients with chronic MPD, AML, and MDS. Also the database of the Department of Leukemia was searched for patients with GS, and all patients with AML or high-risk MDS, that is, refractory anemia with excess blasts (RAEB) or RAEB in transformation (RAEB-T), which presented at MDACC during the same time period. Patients were included in the study if they had biopsy-proven extramedullary disease and no evidence of AML, RAEB, RAEB-T, chronic myelomonocytic leukemia, chronic myelogenous leukemia or other MPD in the bone marrow or the peripheral blood. All slides were reviewed at the MDACC to confirm the diagnosis.

All records were reviewed to investigate clinical and pathologic presentation, treatment and outcome. Patient evaluation included history, physical examination, complete blood count, platelet counts, and differential, biochemical profile, bone marrow aspirates and biopsies, cytogenetic analysis of bone marrow and available extramedullary tissue and available reports of chest X-ray and computerized tomography of head and neck, chest, abdomen and pelvis. Chromosomal analysis by conventional criteria was performed on bone marrow aspirates and/or available tissue at the time of diagnosis.9

Treatment according to standard or investigational regimens was administered either at the MDACC. A signed informed consent was obtained before all therapy, in keeping with the policies of the MDACC.


In all, 16 patients were treated with chemotherapy, three with chemotherapy and radiotherapy and one (with skin disease) radiotherapy only. One patient refused therapy. In all, 11 patients received idarubicin and ara-C (IA), two fludarabine. ara-C, idarubicin and G-CSF (FLAG); two cyclophosphamide. cytarabine, topotecan and G-CSF (CAT-G); one daunorubicin and ara-C. Three patients with central nervous system (n=2) or skin (n=1) disease received both radiation and chemotherapy.

End points and statistical methods

Complete remission (CR) was defined as complete disappearance of all detectable clinical and radiographic evidence of disease and disappearance of all disease-related symptoms. Partial remission (PR) was defined as a reduction by ≥50% of the sum of the products of the greatest diameters of bidimensionally measurable disease. Any other response was considered a failure. Survival was measured from the time of diagnosis until death from any cause, or last follow-up. Failure-free-survival (FFS) was defined as the time from time of diagnosis until relapse, progression to leukemia, death or last contact. The χ2 test was used to investigate the independence between two categorical variables. Survival curves were estimated using the Kaplan–Meier method.10 Adverse events were evaluated using the National Cancer Institute common toxicity criteria (NCI-CTC).11



A total of 21 patients with nonleukemic GS were identified. The numbers of patients with AML and high-risk MDS are shown in Table 1. GS occurred in 1.4% of patients with AML, and 1.1% of patients with AML or high-risk MDSs.

Table 1 Granulocytic sarcoma, AML and high-risk MDS: 1990–2002

Pretreatment characteristics

The presenting clinical and laboratory characteristics of the patients with nonleukemic GS are listed in Table 2. Median age was 57 years (range, 7–81). The median number of sites of involvement was 2 (range, 1–4). The most common extramedullary sites of involvement were skin, lymph nodes, testes, and central nervous system.

Table 2 Clinical and laboratory characteristics of patients with GS

In total, 20 patients had cytogenetic analysis in the bone marrow (n=18) and/or extramedullary mass (n=4). Six patients (28%) had chromosome 8 abnormalities: trisomy 8 in the bone marrow (n=4); 8q- in the bone marrow and the lymph node (n=1); complex cytogenetic abnormalities including trisomy 8 and trisomy 7 in the bone marrow (n=1). One patient had inv(16) in the bone marrow; one patient had -7 and inv(2) in the bone marrow; one patient had deletion 12 in the lymph node; one patient with complex cytogenetics, including 11q- and 1 p- in the lymph node; and one patient had dup(1) and deletion 3 in the paraspinal mass and normal cytogenetics in the bone marrow. Nine patients (45%) had normal cytogenetics.


Among the 20 patients, who received treatment and were evaluable for response, 13 patients (65%) achieved CR, one patient (5%) achieved PR and six patients failed to respond (early death, two patients; and refractory/progressive disease, four patients). Response according to treatment is shown in Table 3. The median time to respond to chemotherapy was 28 days (range, 20–84). Three out of six patients (50%) with chromosome 8 abnormalities achieved CR compared with nine who achieved CR (64%) and one with PR out of 14 patients with other cytogenetics (P=0.55).

Table 3 Response, FFS and survival in 20 patients with non leukemic GS according to treatmenta


The median follow-up of surviving patients is 12 months (range, 7–75). Among the 20 patients who received treatment, nine patients (45%) are still alive. Eight patients died from progressive disease, one patient from fever of unknown origin and heart failure, and two patients died from multiorgan failure, including one patient with sepsis. The patient who was not treated died 4 months after diagnosis from progressive disease. The median survival of patients treated with chemotherapy was 15 months and the projected 3-year survival was 30% with an apparent plateau after 2 years (Figure 1).

Figure 1

Survival in patients with nonleukemic GS treated with chemotherapy.

In Figure 2, the survival of patients with chromosome 8 abnormalities is compared with that of patients with other cytogenetics. Although survival was not significantly different between these two groups (P=0.17), five out of six patients (83%) with chromosome 8 abnormalities have died at the time of this writing (median survival of 12 months) compared with seven out of 14 patients with other cytogenetics (50%; median survival of 40 months).

Figure 2

Survival according to the presence or absence of chromosome 8 abnormality.

Failure-free survival

A total of 12 patients have failed, including nine patients who died (Table 3). Five patients developed AML with a median time to progression to AML of 5 months (range, 4–26); these patients were treated with chemotherapy group. Three patients died from progression of their GS. The projected 3-year FFS in patients who received chemotherapy was 21%.


Among the 20 patients who were treated with chemotherapy with or without radiotherapy, nine patients developed infections and/or sepsis (bacterial, six; fungal, one; viral, one; sepsis and pneumonia with negative cultures, one). There were six episodes of fever of unknown origin and two episodes of neutropenic fever. Six patients developed Grade 3-4 hemorrhagic episodes (gastrointestinal, two; subarachnoidal, one; hemorrhagic conjunctivitis, one; pulmonary hemorrhage, one; and vaginal, one). Other Grade 3–4 toxicities were as follows: arrhythmia, three patients; pulmonary edema, one patient; adult respiratory distress syndrome, one patient; hyperbilirubinemia, three patients; elevated serum creatinine, one patient, mucositis, one patient; diarrhea, one patient; and rash, one patient. Grade 1–2 toxicities were mainly electrolytic abnormalities, nausea, vomiting, and diarrhea.


This is a retrospective analysis of nonleukemic GS patients, based on our referral population. In this population, the overall ratio of nonleukemic GS to AML and high-risk MDS is 0.011. The median age of patients with nonleukemic GS was similar to that of patients with AML and high-risk MDS in our institution.12 Nonleukemic GS presented most frequently with skin lesions and lymphadenopathy. Chromosome 8 abnormalities were the dominant abnormalities in the bone marrow and/or extramedullary sites. In all, 13 patients achieved CR with chemotherapy, radiotherapy, or combined therapy, with a CR rate of 65%. The median FFS in all groups was 12 months and although the number of the patients is small, the overall survival was prolonged in patients treated with combined chemotherapy and radiotherapy (median, 40 months).

Our patient population was older (median age, 57) than that of a previous report on two new cases and 72 cases reviewed from the literature (median age, 33).13 Skin, lymph nodes, breast, testis, and spleen were more frequently involved by GS in our study compared with previous reports, some of which are based on case reports.13 In patients with nonleukemic GS, although CR rates are high, the FFS survival is short (Table 3). In the current study, among the five patients who progressed to AML, the median time to progression was 5 months (range, 4–26). This interval is comparable with a previous report from our institution on 16 patients who presented between 1962 and 1985; seven out of 14 patients with adequate follow-up progressed to AML, 1 week to 13 months after the diagnosis of nonleukemic GS.7 Other investigators have reported a 10–12-month interval between the time of diagnosis and the time to progression to AML.13,14 However, it is unclear whether this interval refers to the overall population with nonleukemic GS or patients who progressed to AML alone.13 Interestingly, none of the three patients treated with chemotherapy combined with radiotherapy progressed to AML in the current study.

Previous studies support the use of intensive chemotherapy in patients with nonleukemic GS. Imrie et al15 have reported that antileukemic chemotherapy at the time of diagnosis is associated with a significantly lower probability of developing AML and with prolonged survival. Stem cell transplantation has also been used as intensified postremission therapy in second CR.15,16

Patients with chromosome abnormalities appeared to have lower CR rates, shorter survival and FFS in our study. Although these differences did not reach statistical significance (Figure 3) – probably because of the small numbers of patients – an abnormal marrow karyotype may indicate a poor prognosis with GS and more intensive therapy may be required for these patients. This also leads one to consider if patients with a GS and an abnormal karyotype can be truly considered to have ‘nonleukemic’ disease – it is more likely that bone marrow cytogenetic analysis is a more sensitive test of marrow involvement. Extramedullary leukemia has been associated with poor prognosis in patients with AML and t(8;21).17 Byrd et al17 reported that among AML patients with t(8;21)(q22;q22) karyotype, those presenting with extramedullary leukemia have lower rates of CR and survival compared with those without extramedullary leukemia; and suggested that this finding is probably related to inadequate local therapy for patients with involvement of the central nervous system.

Figure 3

Survival according to the presence or absence of an abnormal bone marrow karyotype.

The optimal therapy for patient with nonleukemic GS has not been determined. Our data indicate that AML-type intensive chemotherapy with or without radiotherapy was moderately effective. The addition of radiotherapy was associated with a prolonged FFS and should be considered, especially in patients with leukemia involvement of the central nervous system. Intensive therapy should be considered for patients with nonleukemic GS with chromosome 8 abnormalities. Novel approaches are required for patients with GS.


  1. 1

    Muss HB, Moloney WC . Chloroma and other myeloblastic tumors. Blood 1973; 42: 721–728.

    CAS  PubMed  Google Scholar 

  2. 2

    King A . A case of chloroma. Mon J Med 1853; 17: 97.

    Google Scholar 

  3. 3

    Rappaport H . Tumours of the hematopoietic system. In: Flo A (ed). Pathology: Atlas of Tumor Pathology. Armed Forces Institute of Pathology, Washington, DC, 1966, pp 241–243.

    Google Scholar 

  4. 4

    Davey FR, Olson S, Kurec AS, Eastman-Abaya R, Gottlieb AJ, Mason DY . The immunophenotyping of extramedullary myeloid cell tumors in paraffin-embedded tissue sections. Am J Surg Pathol 1988; 12: 699–707.

    CAS  Article  Google Scholar 

  5. 5

    Menasce LP, Banerjee SS, Beckett E, Harris M . Extra-medullary myeloid tumour (granulocytic sarcoma) is often misdiagnosed: a study of 26 cases. Histopathology 1999; 34: 391–398.

    CAS  Article  Google Scholar 

  6. 6

    Eshghabadi M, Shojania AM, Carr I . Isolated granulocytic sarcoma: report of a case and review of the literature. J Clin Oncol 1986; 4: 912–917.

    CAS  Article  Google Scholar 

  7. 7

    Meis JM, Butler JJ, Osborne BM, Manning JT . Granulocytic sarcoma in nonleukemic patients. Cancer 1986; 58: 2697–2709.

    CAS  Article  Google Scholar 

  8. 8

    Byrd JC, Edenfield WJ, Dow NS, Aylesworth C, Dawson N . Extramedullary myeloid cell tumors in myelodysplastic-syndromes: not a true indication of impending acute myeloid leukemia. Leukemia Lymphoma 1996; 21: 153–159.

    CAS  Article  Google Scholar 

  9. 9

    Mitelman F. ICSN. In: Karger S (ed). An International System for Human Cytogenetic Nomenclature. Basel, Switzerland, 1995.

  10. 10

    Kaplan EL, Meier P . Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481.

    Article  Google Scholar 

  11. 11

    National Cancer Institute. ‘Guidelines for Reporting of Adverse Drug Reactions’. Bethesda, MD: Division of Cancer Treatment, National Cancer Institute, 1988.

  12. 12

    Estey EH, Thall PF, Cortes JE, Giles FJ, O'Brien S, Pierce SA et al. Comparison of idarubicin+ara-C-, fludarabine+ara-C-, and topotecan+ara-C-based regimens in treatment of newly diagnosed acute myeloid leukemia, refractory anemia with excess blasts in transformation, or refractory anemia with excess blasts. Blood 2001; 98: 3575–3583.

    CAS  Article  Google Scholar 

  13. 13

    Yamauchi K, Yasuda M . Comparison in treatments of nonleukemic granulocytic sarcoma: report of two cases and a review of 72 cases in the literature. Cancer 2002; 94: 1739–1746.

    Article  Google Scholar 

  14. 14

    Neiman RS, Barcos M, Berard C, Bonner H, Mann R, Rydell RE et al. Granulocytic sarcoma: a clinicopathologic study of 61 biopsied cases. Cancer 1981; 48: 1426–1437.

    CAS  Article  Google Scholar 

  15. 15

    Imrie KR, Kovacs MJ, Selby D, Lipton J, Patterson BJ, Pantalony D et al. Isolated chloroma: the effect of early antileukemic therapy. Ann Intern Med 1995; 123: 351–353.

    CAS  Article  Google Scholar 

  16. 16

    Huter O, Brezinka C, Nachbaur D, Schwaighofer H, Lang A, Niederwieser D . Successful treatment of primary extramedullary leukemia (EML) of the uterus with radical surgery, chemotherapy, autologous bone marrow transplantation (BMT) and prophylactic local irradiation. Bone Marrow Transplant 1996; 18: 663–664.

    CAS  PubMed  Google Scholar 

  17. 17

    Byrd JC, Weiss RB, Arthur DC, Lawrence D, Baer MR, Davey F et al. Extramedullary leukemia adversely affects hematologic complete remission rate and overall survival in patients with t(8;21)(q22;q22): results from Cancer and Leukemia Group B 8461. J Clin Oncol 1997; 15: 466–475.

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to F J Giles.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Tsimberidou, AM., Kantarjian, H., Estey, E. et al. Outcome in patients with nonleukemic granulocytic sarcoma treated with chemotherapy with or without radiotherapy. Leukemia 17, 1100–1103 (2003).

Download citation


  • granulocytic sarcoma
  • therapy
  • survival
  • cytogenetics

Further reading


Quick links