Specific cutaneous infiltrates in patients with myelogenous leukemia: A clinicopathologic study of 26 patients with assessment of diagnostic criteria

Background: Few recent studies have analyzed the clinicopathologic features of specific cutaneous manifestations of myelogenous leukemia in a large number of patients. Objective: We characterize the clinical and histopathologic spectrum of specific cutaneous manifestations in acute (AML) and chronic (CML) myelogenous leukemia, ascertain further diagnostic criteria, and examine current prognosis. Methods: Thirty-six lesions of specific cutaneous infiltrates from 26 patients with myelogenous leukemia (AML: 17 patients; M:F = 1:2.4; mean age: 52.6 years; AML-French-American-British [FAB] classification subtypes:Ml =1, M2 = 3, M4 = 8, M5 = 5. CML = 9 patients; M:F = 4.5:1; mean age: 60.6 years) were retrospectively collected for the study. Results: Cutaneous manifestations presented as solitary or multiple reddish to violaceous papules, plaques, and nodules (17 lesions), or as a generalized erythematous maculopapular eruption (9 lesions). Concurrent extramedullary involvement. in other peripheral sites (eg, gums, pharynx, orbits) was observed in 10 patients. Histopathologically, lesions revealed nodular/diffuse infiltrates, often with perivascular and periadnexal accentuation, sparing of the upper papillary dermis, and prominent single arraying of neoplastic cells between collagen bundles. Extension to the subcutis was noted in all deep biopsy specimens (26 lesions). Cytomorphologically, medium to large-sized mononuclear cells (myeloblasts and atypical myelocytes) predominated in AML-M1 and M2, whereas M4 and M5 mainly showed small, medium-sized, or large mononuclear cells with slightly eosinophilic cytoplasm and indented, bi-lobular, or kidney-shaped nuclei (atypical monocytoid cells). In CML, either a variable mixture of mature and immature cells of the granulocytic series (myelocytes, metamyelocytes, eosinophilic metamyelocytes, and neutrophils) or a rather monomorphous infiltrate of mononuclear cells were found. Staining for naphthol AS-D chloroacetate-esterase (NASD) was positive in 24 of 36 lesions (66.6%; AML: 16; CML: 8). Immunohistochemical analysis on paraffin sections using a large panel of antibodies (16 lesions: AML: 13; CML: 3) showed strong reactivity for LCA (CD45), lysozyme, myeloperoxidase (MPD), LN2 (CD74), HLA-DR, and MT1 (CD43) in the majority of cases, and variable staining for monocyte/macrophage markers (KP1/CD68, PGM1/CD68, Mac387, Ki-M1p). The neuronal cell adhesion molecule (NCAM) marker CD56 was reactive in 2 cases of CML, but negative in all cases of ANLL. MIB1(Ki67) stained 20% to 80% of neoplastic cells. CD34, CD15, CD20, and CD3 were negative in all cases. No correlation between histochemical/immunohistochemical features with type of leukemia or FAB-subtype of AML was observed. All patients with CML and AML with adequate follow-up died within 24 months after onset of skin lesions (mean survival, AML: 7.6 months; CML: 9.4 months). Conclusion: Specific cutaneous lesions in AML and CML show distinctive clinicopathologic features that allow diagnosis in most cases. Immunohistochemistry on routinely fixed, paraffin-embedded tissue sections provides useful adjunctive information. Simultaneous expression of lysozyme, MPD, CD45, CD43, and CD74 militates in favor of a diagnosis of specific cutaneous infiltrate of myelogenous leukemia. Pitfalls in immunohistologic diagnosis mainly include lack of expression of some myeloid markers (lysozyme, MPD), and aberrant expression of T-cell markers (eg, CD45RO). Regardless of type of myelogenous leukemia, onset of specific skin manifestations correlates with an aggressive course and short survival.