Granular cell tumor: immunohistochemical assessment of inhibin-alpha, protein gene product 9.5, S100 protein, CD68, and Ki-67 proliferative index with clinical correlation.

CONTEXT: Granular cell tumor (GCT) is a rare tumor of nerve sheath origin with a predilection for upper aerodigestive tract, skin, and soft tissue. The neoplastic cells typically express S100 and CD68 (KP-1), the latter due to cytoplasmic lysosome content. However, the histogenesis of this tumor is unknown. Additionally, distinction between benign and malignant GCT is difficult because of histologic similarity and lack of reliable criteria that can predict clinical behavior.

OBJECTIVE: To perform a comparative, side-by-side immunohistochemical assessment of the traditional immunohistochemical markers for GCTs (S100, CD68), along with the newer markers (inhibin-alpha, protein gene product 9.5) for these tumors.

DESIGN: To address diagnostic and prognostic issues, we studied 30 specimens of GCT (27 primary and 3 recurrent tumors, 2 of which occurred consecutively in the same patient) for (1) nuclear pleomorphism, prominent nucleoli, necrosis, spindling, high nuclear-cytoplasmic ratio, and mitoses; (2) immunohistochemical expression of inhibin-alpha, protein gene product 9.5, S100, CD68 (KP-1), and Ki-67 using the avidin-biotin complex method on formalin-fixed, paraffin-embedded sections; and (3) correlation between tumor grade, proliferative fraction, and clinical data.

RESULTS: Twenty-seven of 27 primary GCTs and 1 of 3 recurrent GCTs had typical histologic features, while the 2 consecutive recurrent GCT specimens from the same patient were atypical (moderate nuclear atypia and prominent nucleoli alone). The mean age for primary GCT was 37.3 years (range, 5-67 years), and mean size was 1.89 cm. None of the cases metastasized. All 30 specimens showed diffuse (2+ to 3+) staining for S100, CD68, and inhibin-alpha, and 3+ staining for protein gene product 9.5; pseudoepitheliomatous hyperplasia was nonreactive. The Ki-67 proliferative index was less than 1% to 20% in typical nonrecurrent cases, 1% in the typical recurrent case, and 1% and 10% in 2 sequential recurrences of the atypical case.

CONCLUSION: Our study expands the immunophenotype of GCT (S100, CD68, protein gene product 9.5, and inhibin-alpha) regardless of location and supports a neural origin. Intensity of immunohistochemical staining had no prognostic significance. Although 1 of the 2 recurrent GCTs had atypical features, the Ki-67 proliferative index did not distinguish reliably between typical (nonrecurrent) and atypical or recurrent GCTs. The significance of inhibin expression with regard to cell differentiation and pathogenesis is unclear and warrants further investigation.