Fine needle aspiration biopsy of solitary fibrous tumors. Report of two cases with histologic, immunohistochemical and ultrastructural correlation.

BACKGROUND: Although the fine needle aspiration biopsy (FNAB) technique is widely applied to thoracic and soft tissue tumors, the cytologic morphology of solitary fibrous tumors is not fully established. CASES: We report the cytologic features observed on computed tomography (CT)-guided FNAB obtained from two benign solitary fibrous tumors of the pleura in two males. The cytologic findings were compared to the light and electron microscopic features. The smears were composed of spindle cells dispersed amid a delicate network of small blood vessels. Scattered among these spindle cells were also epithelioid cells arranged singly or in small clusters. These cells were polygonal, with moderately dense, homogeneous or finely vacuolated, basophilic cytoplasm and a moderately high nuclear/cytoplasmic ratio. The nuclei were oval to round and slightly hyper-chromatic and rarely had small nucleoli. Abundant spindle-shaped, round, naked nuclei were seen in the background. Stains for CD34 were positive in the epithelioid and spindle cells. Histologically the lesions were composed of spindle cells embedded in a collagenous matrix. Hypercellular nodules and focal hemangiopericytic and storiform patterns were also observed. There were no features suggesting malignancy. In electron microscopic evaluation there were fibroblastic cells in a collagenous matrix as well as polygonal cells singly and in clusters with scanty to moderate amounts of cytoplasm containing sparse organelles. The cells were attached to each other by tight junctions. Occasionally, cells were surrounded by short segments of basal lamina. CONCLUSION: The cytomorphologic pattern described above correlates well with the histologic and ultrastructural findings in benign solitary fibrous tumors (SFT) and corresponds perfectly to the histogenetic concept of this lesion. As with other tumors, fine needle aspiration biopsy is helpful in the evaluation of SFT.