Primary myelodysplastic syndromes: diagnostic and prognostic significance of immunohistochemical assessment of bone marrow biopsies.

Material from 63 cases with primary myelodysplastic syndromes (P-MDS) (French-American-British [FAB] types: refractory anemia [RA] = 21; RA with ring sideroblasts [RARS] = 8; RA with excess of blasts (RAEB) = 10; RAEB in transformation (RAEBt) = 6; chronic myelomonocytic leukemia [CMML] = 10 and unclassifiable = 8, ie, bone marrow aspiration was inadequate and stringent FAB criteria were not applicable) was analyzed for bone marrow histologic and immunohistochemical patterns. Standard Giemsa, hematoxylin and eosin (H&E) and reticulin stains were used for morphologic assessment. To identify the cell lineage precisely, chloroacetate esterase staining and an indirect immunoperoxidase technique using mouse monoclonal antibodies CD15, CD68, HLA-DR, and rabbit polyclonal CD3 and UEA-1 (lectin) was developed on formalin-fixed paraffin embedded bone marrow biopsies (BMB). The immunohistochemical assessment permitted accurate identification of dysplastic features such as mononuclear and binuclear megakaryocytes, Pelger-Huet neutrophils, and binuclear erythroblasts. Additional bone marrow histologic and immunohistochemical features observed were heterogeneity of immunohistochemical staining in various cell lineages, megakaryocytic emperipolesis, alteration of bone marrow microarchitecture, intravascular clusters of hematopoietic cells, and the types of benign lymphoid aggregates. The nature of abnormally localized immature precursors (ALIP) was discerned. Three types of clusters of immature cells were found that were difficult to distinguish on Giemsa and H&E morphology, these were erythroid aggregates (n = 18); megakaryocytic aggregates (n = 4), and immature granulocytic and monocytic aggregates (n = 32). The bone marrow histologic and immunohistologic patterns permitted the identification of four groups of clinical relevance: Group 1, cases with predominant erythroid hyperplasia and without ALIP (n = 15); group 2, cases with prominent myeloid hyperplasia and presence of ALIP (n = 32); group 3, cases with hypoplastic MDS (n = 10); and group 4, cases with hyperfibrotic MDS (n = 6). Statistical analysis showed a significant difference in survival and leukemic transformation between groups 1, 2, 3, and 4, with cases in group 2 showing the worst prognosis with early death due to increased propensity to leukemic transformation and cytopenia-related complications (P less than .0001). We conclude that immunohistochemistry is feasible on routinely processed BMB and the information obtained is of diagnostic and prognostic importance in P-MDS. The phenotype of ALIP varies with the morphologic and histologic subtypes of MDS and the term should be reserved for cases in whom the clusters in the intertrabecular region are of myeloid (granulocytic and monocytic) lineage on immunohistochemistry.