Extensive Molecular Analysis Strongly Improves the Distinction Between AML and ALL in Adult Acute Leukemias of Ambiguous Lineage

Abstract

BackgroundIn 1995 the European Group for the Immunological Classification of Leukemias (EGIL) presented guidelines to classify bilineage and biphenotypic acute leukemias (BAL). Subsequently, the WHO in 2008 defined mixed phenotype acute leukemias (MPAL), as leukemias coexpressing antigens of both lymphoid and myeloid lineage, and leukemias with distinct blast populations of more than one lineage. However, a classifying diagnosis towards myeloid or lymphoid leukemia based on an extensive molecular-biological characterization could provide much better distinction, which could be useful for selecting AML or ALL like treatment approaches. We investigated whether extensive molecular analysis might improve a diagnosis of BAL/MPAL into predominant AML or ALL.Patients and methodsFor this study 25 adults, diagnosed between 2000 and 2009 with BAL/MPAL were identified. Flowcytometry and cytogenetics were performed at diagnosis. In addition, molecular analysis was performed to identify BCR-ABL, MLL, AML1-ETO, CBF-MYH11, SIL-TAL1, SET-NUP, NUP-ABL fusion transcripts, FLT3-ITD, CEBPA, NPM1, NOTCH1 mutations and IKZF1, CDKN2A, CDKN2B deletions. Furthermore, multiplex PCR assays to detect immunoglobulin (Ig) and T-cell receptor (TCR) gene rearrangements were performed. Gene expression profiles (GEP) were generated, and results were compared to an AML/ALL classifier profile. A classifying diagnosis of either ALL or AML was made if at least two of the 5 assays applied, were highly suggestive for a specific lineage without explicit contradicting results of the remaining assays.ResultsTwenty-four of the 25 BAL/MPAL cases could be diagnosed into predominant AML or ALL by extensive molecular analysis (Table 1). In 11 cases the WHO2008 criteria classified BAL cases as AML or ALL, and extensive molecular analysis also indicated the same lineage. Adding standard cytogenetic and molecular analysis improved the WHO classification in another 4 patients. Remarkably, extensive molecular analysis clearly showed preference for either AML or ALL in all of the MPAL cases. As a result, we were able to improve a diagnosis in all MPAL cases. Only 1 of the 25 patients remained unclassified. Statistical analysis did not show significant correlation between the different assays. Twelve patients received AML-based chemotherapy, including 3 patients ultimately classified as ALL. Moreover, among 12 patients who received ALL type therapy, 1 was retrospectively classified as AML. In general, outcome appeared very poor with a 3 year overall survival of 32% (Kaplan-Meier estimate, 95%CI: 14%-51%).ConclusionsThis study suggests that a differential diagnosis of AML versus ALL is partly improved by applying the WHO2008 criteria instead of the EGIL classification. However, extensive molecular analysis, including GEP and Ig/TCR, more strongly enabled the identification of the lineage of origin, ultimately resulting in a classifying diagnosis in 24 out of 25 patients. So far no assay appeared as golden standard but different complementary techniques showed additive value. Collectively these results strongly argue for advanced, centralized diagnostic procedures in patients with acute leukemia of ambiguous lineage, which may subsequently enable the development of specific therapeutic approaches in these poor-risk patients.Table 1Patient characteristicsAge (year)EGILWHO 2008Molecular aberrationsGEP probabilityIg/TCRClassificationAMLALL51TMMPAL*0,0001,000POSALL52TMMPALFLT3-ITD0,9700,030NEGAML20BMMPALMLL-AF4 CEBPA0,0001,000POSALL18BMMPALIKZF1 CDKN2A-2B0,0001,000POSALL49BMMPALBCR-ABL IKZF1NDNDPOSALL57BMMPALCDKN2A0,0001,000POSALL40BMMPALCDKN2A-2BNDNDPOSALL23TMPALCDKN2ANDNDPOSALL24BMPAL*0,0001,000POSALL65BMPALBCR-ABL IKZF1 CDKN2ANDNDPOSALL58BMAML*1,0000,000NEGAML52BMAMLNPM1NDNDNEGAML59BMAMLFLT3-ITD NPM1NDNDNEGAML41BMAMLFLT3-ITD CEBPA NPM1NDNDPOSAML69BMAMLMLLNDNDNEGAML18BMAMLCBF-MYH11NDNDPOSAML71BM#FLT3-ITD CDKN2A-2B0,7090,291NEGAML67BM#*0,0030,997POSALL15BMALLIKZF1 CDKN2A-2BNDNDPOSALL23BM#*0,9980,002POS?32BTM#*0,9380,062POSAML27BTMALLIKZF10,0010,999POSALL69BTMALLCDKN2A-2B0,0980,902NEGALL16BTMALLCDKN2A-2BNDNDPOSALL54TMALLNOTCH10,1480,852NEGALLAbbreviations: ND, not determined.* No abnormalities# Unclear classification according to WHO2008 DisclosuresNo relevant conflicts of interest to declare.