WHO classification of myeloid neoplasms and leukemia

To the editor:

Vardiman et al have focused their paper¹  on major changes in the 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and leukemia compared with the 2001 edition and have provided the rationale for those changes. Many of these changes and new definitions follow biologic features and include important information for prognosis. They pave the way not only to a better understanding of acute myeloid leukemia (AML) but also will advance outcome for patients. However, we cannot agree with the rationale for maintaining the category of “acute myeloid leukemia with multilineage dysplasia” (MLD), first established in the third edition in 2001, that is now subgrouped in the group of “AML with myelodysplasia-related changes.”

We have shown in 2 large AML studies^2,3  of 2 different study groups (Study Alliance Leukemia and German AML Cooperative Group) in 2380 patients that MLD has no independent prognostic relevance if compared for patients when cytogenetics are also available (a must in WHO classification). Even more, MLD per se has absolutely no prognostic significance in patients 60 years of age or younger with de novo AML and, additionally, in the important subgroup of patients with normal karyotype.

We could show that it is of prognostic relevance to include now “MDS [myelodysplastic syndrome]–related cytogenetic changes”^1p945 in the definition of this new WHO subgroup. However, to define only by morphology AML that “exhibit dysplasia in 50% or more of the cells in 2 or more myeloid lineages”^1p946 cannot be justified based on published data. Thus, MLD as a marker of an AML subgroup should be omitted because it is prognostically and clinically misleading.

Vardiman et al further stated that there is no data concerning the correlation of “morphologic dysplasia” and the molecular mutations NPM1 and FLT3-ITD. As published in our paper in Blood,²  we could show in more than 1200 patients with AML that NPM1 was mutated in 30% of patients with AML and MLD, which was exactly the same percentage as in patients without MLD. FLT3-ITD mutations were, interestingly, significantly more prevalent in MLD-negative versus -positive patients (34% vs 24%, P < .001), that is, appear to be associated with de novo disease. In a multivariate analysis including MLD, age, cytogenetics, history of AML, and different NPM1/FLT3-ITD combinations only the combination NPM⁺/FLT3-ITD⁻ has shown a significant prognostic relevance besides age and cytogenetics as the most powerful prognostic factors.² 

We conclude that a more biologic understanding of AML, as requested by the WHO classification, should omit a group of patients classified only by the morphologic criteria of MLD in the future but further extend the cytogenetic, molecular, and other biologic criteria to define clinically significant disease entities.