Abstract 2621

Current data on prognostic significance of extramedullary leukemia (EML) in general and for specific genetic subgroups in children with acute myeloid leukemia (AML) are limited and controversial. AML-BFM data from previous studies indicated that skin and testicular involvement were independent adverse prognostic factors in high-risk (HR) patients (core-binding factor [CBF] AML [=t(8;21) and inv16] patients were standard risk patients) (Creutzig, Blood (suppl)2001;98(11). This was supported by others for leukemia cutis (Dusenbery, J.Pediatr. Hematol. Oncol.2003). Orbital involvement, which is associated with t(8;21), was reported to be unfavorable in Turkish and Indian children. A Dutch study could not confirm the prognostic relevance of EML (Bisschop, Leukemia,2001). Therefore, we analyzed the data of pediatric CBF-AML and MLL gene-rearrangement AML (MLL-AML) patients <18 years old treated according to protocols AML-BFM 98 and -2004 (7/1998-12/2010) with regard to their initial clinical features and the prognostic relevance of EML including EML-sites. EML was diagnosed in 57/232 patients with de novo CBF-AML and in 80/241 patients with MLL-AML. EML sites included skin, orbit, kidney, testes and other rare localizations like tonsils and salivary glands. In this analysis, central nervous system involvement was not defined as EML and isolated EML was excluded.

Treatment was similar for all patients in both trials (4–5 intensive courses including high-dose cytarabine, prophylactic cranial irradiation, 1-year maintenance and stem cell transplantation in HR patients, if an HLA-identical donor was available. Patients with single site EML received local irradiation. According to the risk classification of the AML–BFM studies, CBF-leukemia patients with <5% bone marrow blasts on day 15 were stratified as standard risk (SR) and MLL-patients as HR patients.

Patient characteristics:

EML occurred significantly less frequently in CBF than in MLL patients [57/232 (24.6%) vs. 80/221 (36.2%); pchi=0.007] (i. e. patients with available data on EML). Sites of EML were a) skin: CBF 7/229 (3.1%) vs. MLL 32/220 (14.5%), pchi=0.00002; b) orbit: CBF 19/228 (8.3%) (i. e. mainly patients with t(8;21): 17/132 (13%)) vs. MLL 13/216 (6.0%), pchi=0.34; c) kidneys: CBF 14/230 (6.6%) vs. MLL 29/215 (13.5%), pchi=0.008; d) tonsils: CBF 9/222 (4.1%) vs. MLL 7/211 (3.3%), pchi=0.68; e) salivary glands: CBF 1/224 (0.4%) vs. MLL 5/213 (3.3%), pchi=0.088; f) testes: CBF: 1/226 (0.4%) vs. MLL 6/220 (2.7%), pchi=0.052; g) other sites (including multiple sites): CBF 27/224 (12.1%) vs. MLL 28/209 (13.4%) patients, pchi=0.67. Patients with MLL (with or without EML) were generally younger than CBF-patients (age <2 years: 49% vs. 7%, pchi<0.0001). Hyperleukocytosis was more frequent in MLL than in CBF patients (30% vs. 11%, pchi<0.001). EML-morphology was FAB M2Auer/M4eo in 86% of AML-patients with CBF and with FAB M4/M5 in 95% of patients with MLL-rearrangement.

Results:

Outcome for EML-patients with CBF-AML was significantly better than for EML-patients with MLL-AML with a 5-year event-free survival (pEFS) of 63+3% vs. 45+6%, respectively (plogrank=0.0006). Extramedullary site of the disease was found to have no prognostic significance for patients with CBF-AML: pEFS with or without EML was 63+3% vs. 75+3% (plogrank=0.38). Also, there was no difference in EFS of patients with t(8;21)- or inv(16)-AML with or without EML (plogrank=0.28 and 0.69, respectively). In particular, orbital involvement was not of prognostic significance in patients with t(8;21) (plogrank=0.26).

Further, pEFS results in MLL-AML patients with or without EML were similar (45% both, plogrank=0.48). However, prognosis for patients with MLL-rearrangement and skin involvement was significantly more unfavorable when compared to those with MLL-AML at other extramedullary sites (EFS 32%+9% vs. 54%+7%, plogrank=0.03). In addition, testicular involvement in MLL-patients might also be unfavorable (4 events in 6 patients). Excluding MLL patients with skin and testicular involvement, extramedullary site was not of prognostic relevance (pEFS MLL with other EML vs. MLL without EML = 55%+8% vs. 45%+4%, plogrank=0.44).

Conclusion:

EML did not influence the favorable prognosis of CBF-AML, whereas skin and testicular involvement in MLL-AML were unfavorable prognostic factors.

Disclosures:

No relevant conflicts of interest to declare.

Topics:
core-binding factor, leukemia, myeloblastic leukemia, pediatric acute, prognostic factors, brachytherapy, cranial irradiation, cytarabine, hematopoietic stem cell transplantation, human leukocyte antigens, leukemia, myelocytic, acute

This icon denotes a clinically relevant abstract

Author notes

*

Supported by the Deutsche Krebshilfe e.V.

© 2012 by The American Society of Hematology
2012
Sign in via your Institution