A Rapid Approach for the Integrated Central Review of Acute Myeloid Leukemia Diagnosis In a Nationwide Clinical Trial for Children

Abstract 4833

Diagnosis in accordance with the WHO classification provides important information for prognosis and risk stratification in the clinical trials of therapy for AML. However, each of the diagnostic parameter contained in the WHO classification, such as morphologic, cytochemical, immunophenotypic, cytogenetic, and genetic features, requires standardization and accuracy to ensure the validity of such clinical trials. In addition, discrepancies among the results of these parameters are encountered in some patients. Based on these considerations, we attempted to develop a comprehensive and rapid system for the central review of AML diagnosis in a nationwide clinical trial conducted by a multi-institutional cooperative study group. JPLSG AML-05 is a study that was conducted on children with de novo AML, excluding APL and myeloid leukemia associated with Down syndrome. Morphological diagnosis and flowcytometric immunophenotyping were performed through central laboratory systems. Cytogenetic analyses were carried out in regional laboratories, and the reports were reviewed centrally. Expression of chimeric genes including AML1-ETO, CBFB-MYH11, PML-RARA, MLL-AF9, MLL-AF6, MLL-ELL, FUS-ERG, and NUP98-HOXA9, and mutations of FLT3-ITD were also examined for all patients at a single laboratory. In case where cellular bone marrow aspirate smears could not be obtained, we recommended bone marrow biopsy, for which the histology was subjected to central pathologic review. When all the data had been submitted to our data center, a central review of the diagnosis based on the 2001 edition of the WHO classification was performed for each individual case by multidisciplinary conferencing based on electronic mail discussion. Between November 2006 and July 2010, the diagnoses of 383 patients enrolled in AML-05 were reviewed. The diagnosis was undetermined for seven patients because of inadequate submission. Nine patients did not meet the inclusion criteria for AML-05. Among the remaining 367 patients, 184 (48.8%) had AML with recurrent abnormalities, 32 (8.7%) had AML with multilineage dysplasia (AML-MLD), and 151 (41.1%) had AML, not otherwise categorized. Among the 184 patients with AML with recurrent abnormalities, 104 had AML with t(8;21)(q22;q22)/AML1-ETO. A higher prevalence of AML with t(8;21)(q22;q22)/AML1-ETO in our study, compared with those in previous reports from European countries, suggests a difference in genetic background. Two patients with t(8;21)(q22;q22)/AML1-ETO and one patient with inv(16)(p13q22) were assigned to each category through FISH analysis, because their cytogenetic analyses indicated a normal karyotype or suboptimal quality. Among 55 patients who were suspected of having AML with 11q23 (MLL) abnormalities by chromosome or chimeric gene screening tests, 30 had concordant results between chromosomes and chimeric genes. The remaining 25 patients were assigned to AML with 11q23 (MLL) abnormalities by either FISH for MLL or other chimeric gene tests involving MLL (MLL-AF10, MLL-ENL,MLL-AVI1 or MLL-TET1). The prevalence of AML-MLD in this study was higher than that (2.6%) in a previous report for Japanese children with AML, where a central review system of morphology had not been established (Adachi S., et al. Int J Hematol 86;358-63,2007). Patients with AML-MLD had a higher frequency of unfavorable cytogenetics (6/32) and/or FLT3-ITD mutations (7/32) in comparison with other types of AML. One patient with acute megakaryoblastic leukemia was diagnosed on the basis of bone marrow histology in our central pathologic review because the percentage of blasts on the smear did not meet the criteria for AML. We have successfully developed a rapid system for the integrated central review of AML diagnosis based on the WHO classification, which will enable us to evaluate trial data on the basis of standardized diagnosis and risk classification.