Methods

Top Methods Results Discussion Author & Article Info References

To identify all patients with isolated chloroma, we did a retrospective review of records from 1980 to 1994 at three teaching hospitals associated with the University of Toronto. In all cases, the diagnosis was reviewed by one of three reference hematopathologists, and the histologic diagnosis was confirmed using a naphthol-ASD-chloroacetate esterase stain and appropriate enzyme cytochemistry and immunocytochemistry. We excluded all patients with a history of a myeloproliferative syndrome, myelodysplasia, or acute myeloid leukemia diagnosed before or concurrent with diagnosis of chloroma. A normal complete blood count and bone marrow aspirate at diagnosis were required for inclusion.

Patient characteristics evaluated were age, sex, site of chloroma, and treatment given at diagnosis, including surgical resection, chemotherapy, radiotherapy, and bone marrow transplantation. The chemotherapy regimen was judged to be an acute myeloid leukemia induction regimen if cytosine arabinoside was incorporated. Consolidation or maintenance therapy was not evaluated. Outcome measures included survival from diagnosis of chloroma and the interval between diagnosis of chloroma and the development of leukemia.

We did a MEDLINE search of the English-language literature from 1980 to 1994 and identified all articles in which the following terms appeared in the title or abstract: granulocytic sarcoma, chloroma, and myelosarcoma. Inclusion and exclusion criteria were identical to those used for locally identified cases but, in addition, we excluded patients for whom therapy or outcome was not reported.

Survival from diagnosis of chloroma was estimated using the Kaplan-Meier method [7]. We evaluated the effect of the following variables on survival and the development of leukemia: patient age, site of chloroma, and chemotherapy, local radiotherapy, or surgical resection done at diagnosis. We tested for the significance of the effect of the type of initial therapy on the development of leukemia using the Pearson chi-square test. The effect of treatment on disease-free survival and overall survival was tested using the log-rank test. Multivariate analysis was done using Cox regression analysis. All statistical testing was done using SPS for Windows version 6.1.

Results

Top Methods Results Discussion Author & Article Info References

Review of the literature showed 208 English-language papers with one of the requisite keywords in the title or abstract. A total of 149 reports were excluded: Eighty-one included only patients with a previous or concurrent diagnosis of acute myeloid leukemia, the nature of therapy or outcome was not evaluable in 30, the final diagnosis was not chloroma in 21, and 17 referred only to patients with previous myeloproliferative or myelodysplastic disorders. The remaining 59 reports, describing a total of 83 cases, were included. Seven patients from three hospitals in Toronto were identified in the defined time period. Details of the entire group of 90 cases are summarized in Table 1. The median age at diagnosis was 35 years. Sixty-eight percent of patients were male. The site of chloroma was visceral in 38% of patients, soft tissue or skin in 20%, head or neck in 16%, bone in 12%, the central nervous system in 6%, and other sites in 8%. Eleven patients (13%) had multiple lesions. Cytogenetic analysis at diagnosis of chloroma was not available from biopsy material in any patients due to lack of marrow involvement and a low index of suspicion before biopsy.

Forty-nine of the 90 patients (54%) received chemotherapy at diagnosis. Of these, 24 received standard leukemia induction chemotherapy and 20 received other treatment including chemotherapy for lymphoma or palliation. Data were insufficient to assess consolidation or maintenance chemotherapy. In 5 patients, the type of chemotherapy given could not be determined. Thirty-seven patients (41%) had surgical resection and 56 (62%) had local irradiation at diagnosis.

Fifty-nine of the 90 patients (66%) developed acute myeloid leukemia at a median of 9 months (95% CI, 7 to 13 months) after diagnosis of chloroma. The FAB subtype of the leukemia at relapse was described in only 8 patients (9%): It was M₂ in 3 patients, M₄ in 2 patients, M₁ in 1 patient, M₃ in 1 patient, and M₇ in 1 patient. The karyotype of the malignant clone in the bone marrow at diagnosis of acute leukemia was reported in 9 patients: It was t(8;21) in 3 patients, +8 in 2 patients, inv(16) in 1 patient, t(15;17) in 1 patient, t(1;7) in 1 patient, and normal in 1 patient. Forty-eight patients (53%) received induction chemotherapy at the time of diagnosis of overt leukemia regardless of initial therapy. A significantly lower proportion of the patients who received any form of chemotherapy at initial diagnosis subsequently developed leukemia (41% ± 14% compared with 71% ± 14%; P = 0.001). In addition, leukemia occurred significantly later in those patients who received systemic chemotherapy. The median time from diagnosis of chloroma to leukemia was 36 months (CI, 12 to 66 months) in patients treated with chemotherapy compared with 6 months (CI, 4.2 to 7.8 months) in untreated patients (P < 0.001).

Figure 1 is a Kaplan-Meier plot of survival for all 90 patients. Median survival was 22 months (CI, 13.2 to 30.8 months). Sex, site of chloroma, and treatment with radiotherapy or surgical resection did not influence survival or development of acute leukemia. Being more than 50 years of age was associated with a poorer prognosis (P = 0.003). Survival was significantly longer in patients receiving any form of chemotherapy at diagnosis. Median survival was 13 months (CI, 8.6 to 17.4 months) for those who did not receive chemotherapy. In contrast, median survival could not be calculated for patients who received chemotherapy, because more than half of these patients were alive with a median follow-up of 25 months (range, 1 to 192 months; P = 0.001) when their cases were submitted for publication. The group treated with chemotherapy and the untreated group did not differ with respect to age, sex, or site of chloroma.

View larger version (12K): [in this window] [in a new window]   Figure 1. Survival of 90 patients with isolated chloroma.

We next analyzed the difference in survival between patients receiving accepted leukemia induction and patients receiving other chemotherapy at diagnosis of isolated chloroma. Median survival was 39 months (CI, 0 to 78 months) for the latter group, whereas the median survival of the former group had not been reached at a median follow-up of 35 months (range, 1 to 144 months; P = 0.03). Cox regression analysis showed only younger age (< 51 years) and chemotherapy at diagnosis to be factors favorably affecting survival (P = 0.003 and P < 0.001, respectively).

Seven patients had bone marrow transplantation (four allogeneic and three autologous), and four of the seven had transplantation done before any evidence of leukemia was seen. Six of the seven were alive in complete remission with a median follow-up of 22 months.

Discussion

Top Methods Results Discussion Author & Article Info References

We confirm that chloroma in patients without acute myeloid leukemia or other bone marrow disorders is associated with a poor prognosis. The median survival of 22 months found in our study is similar to that found in the two series that included 10 or more patients [2, 4]. The median age in our study is similar to that in other published series [2, 4], but is lower than the median age of onset of acute myeloid leukemia [8]. Older age is a significant adverse prognostic factor and, although this association has not previously been noted, is in keeping with the outcome for acute myeloid leukemia [9]. Our series shows a slight preponderance of men among patients with isolated chloroma (62%), similar to that seen with acute myeloid leukemia [9].

Cytogenetic and histologic data at relapse were available for only a minority of patients but were consistent with the reported association of chloroma with FAB M₂ leukemia and the t(8; 21) translocation [10, 11].

We show a significantly lower rate of progression to leukemia and longer survival among patients who received any form of chemotherapy at diagnosis of chloroma; the best results were seen in those patients receiving acute myeloid leukemia induction. The nature of the chemotherapy varied because many of the published cases had initially been misdiagnosed, often as non-Hodgkin lymphoma.

It is important to emphasize that our finding of longer survival in patients initially treated with chemotherapy is based on a retrospective review of case reports and short case series and is therefore subject to bias. Publication bias may exclude patients who have relapse after aggressive chemotherapy. We have attempted to minimize this effect by evaluating all reports published during a 14-year period and, specifically, by including in our review not only case series published in hematology or oncology journals but also many individual case reports in the medical, surgical, and radiology literature. Another potential confounding factor is that fewer patients with adverse prognostic features, such as poor performance status, are likely to be offered aggressive chemotherapy. The lack of detailed information on consolidation or maintenance therapy is an additional concern. Although they cannot be excluded, such factors are likely to have only a minor effect in patients without systemic disease. Moreover, age, sex, and site of chloroma were similar in the two groups.

A randomized trial is needed to confirm our conclusions, but, in view of the rarity of this presentation and the difficulties with initial diagnosis, such a trial is unlikely to be done. On the basis of our analysis, and in the absence of a prospective trial, we conclude that patients with isolated chloroma should receive standard induction chemotherapy for acute myeloid leukemia at diagnosis.

Dr. Keating: The Toronto Hospital, General Division, MLW 2-036, 200 Elizabeth Street, Toronto, Ontario, M5G 2C4, Canada.

Dr. Kovacs: Victoria Hospital, 375 South Street, PO Box 5375, London, Ontario, N6A 4G5, Canada.

Drs. Lipton and Patterson: Princess Margaret Hospital, 500 Sherbourne Street, Toronto, Ontario, M4X 1K9.

Dr. Pantalony: The Toronto Hospital, General Division, E 3-306, 200 Elizabeth Street, Toronto, Ontario, M5G 2C4, Canada.