Pulmonary "inflammatory myofibroblastic" tumors: a critical examination of the diagnostic category based on quantitative immunohistochemical analysis.

The World Health Organization classification applies the term "pulmonary inflammatory myofibroblastic tumor" to a histologically variegate set of pulmonary inflammatory pseudotumors. However, often these lesions bear little resemblance to tumors of myofibroblastic origin. To elucidate histogenesis, we examined 18 cases from our institution files. The cases were stained with antibodies to smooth muscle actin (SMA), Factor XIIIa, CD3, CD20, CD68, S-100, anaplastic lymphoma kinase (ALK-1), and human herpevirus-8 (HHV-8). The percentage of positive-staining cells within a defined tumor area (400,000 microm(2)) was determined by light microscopy and morphometric analysis. Ten cases (56%) showed myofibroblastic differentiation, as judged by positive SMA staining of spindle cells. All cases showed substantial numbers of CD68+, Factor XIIIa+, and S-100+ monocytoid cells. Fifty percent were ALK-1+, and one was HHV-8+. We conclude that the term "inflammatory myofibroblastic tumor" is a misnomer, as nearly half of cases show no myofibroblastic differentiation. Instead, the results suggest that these lesions are composed predominantly of cells of macrophage-dendritic cell lineage. Although the multiplicity of terms previously applied to these lesions is cumbersome, retaining a descriptive phenomenological terminology may ultimately promote accurate elucidation of pathogenesis.