Several mouse models of human prostate cancer were studied to identify and characterize potential precursor lesions containing foci of atypical epithelial cells. These lesions exhibit a sequence of changes suggesting progressive evolution toward malignancy. Based on these observations, a grading system is proposed to classify prostatic intraepithelial neoplasia (PIN) in genetically engineered mice (GEM). Four grades of GEM PIN are proposed based on their architecture, differentiation pattern, and degree of cytological atypia. PIN I lesions have one or two layers of atypical cells. PIN II has two or more layers of atypical cells. PIN III has large, pleomorphic nuclei with prominent nucleoli and the cells tend to involve the entire lumen with expansion of the duct outlines. PIN IV lesions contain atypical cells that fill the lumen and bulge focally into, and frequently compromise, the fibromuscular sheath. Within the same cohorts, the lower grade PINs first appear earlier than the higher grades. Morphometric and immunohistochemical analyses confirm progressive change. Although the malignant potential of PIN IV in mice has not been proven, GEM PIN is similar to human PIN. This PIN classification system is a first step toward a systematic evaluation of the biological potential of these lesions in GEM.

No abstract

p185erbB-2 overexpression in both CC and SC of FGTCS suggests a common carcinogenic mechanism for both components and supports the conversion-histogenesis hypothesis implicating a dominant role for the CC with the SC arising as a metaplastic change from the CC. EGFR may be expressed in either component and indicates aggressive biologic behavior; however, its prognostic utility is limited by its low predictive value for recurrence (40.3%), inability to foretell recurrence in good prognosis groups, and dependence on stage. High frequency of overexpression and dismal prognosis make FGTCS patients good candidates for trials of therapeutic strategies involving the p185erbB-2 receptor manipulations.