Author(s): Clarke MR, Weyant RJ, Watson CG, Carty SE
Abstract Share this page
Abstract Clinical and histopathological features do not reliably distinguish between benign and malignant pheochromocytomas. Additional markers that might be useful prognostic indicators in the pathological assessment of these tumors are sought. Immunohistochemical expression of MIB-1, Bcl-2, cathepsin B, cathepsin D, basic fibroblast growth factor (bFGF), c-met, and type IV collagenase were studied on formalin-fixed tissue from 33 nonconsecutive cases of pheochromocytoma, selected on the basis of reliable long-term follow-up, to determine associations with malignancy. The study group included 33 patients, 19 men and 14 women, with a mean age of 45 years, including five cases of neurofibromatosis (NF), three familial, and one MEN IIb. Mean follow-up was 63.2 months. Ten patients were determined to have malignant pheochromocytomas by the presence of metastatic disease. Features found to be associated with malignancy included MIB-1 labeling index (5\% vs 1\%) (P = .0009), male gender (90\% vs 43\%) (P = .008), extra-adrenal location (40\% vs 9\%) (P = .03), tumor weight (481 g vs 124 g) (P = .05), and young age (38 years vs 49 years) (P = .05). None of the five cases with NF were malignant (P = .04). S-100 positivity showed a significant (P = .02) but nonlinear association with benign tumors. Absent S-100 correlated with greater tumor weight. Malignancy was not associated with right versus left side or bilaterality, although bilateral tumors were smaller. C-met, bFGF, cathepsin B, cathepsin D, and collagenase were strongly expressed in most tumors and were not predictive of outcome, nor was bcl-2, which was variably expressed. Using multiple logistic regression with malignancy as the dependent variable, MIB-1 continued to show a significant association with malignancy (P = .005) independent of any association with sex, age, or extra-adrenal location. Using a cutoff value of MIB-1 labeling of greater than 3\% yielded a specificity of 100\% and a sensitivity of 50\% in predicting malignancy.
This article was published in Hum Pathol
and referenced in Journal of Addiction Research & Therapy