The optimal goal for pathologists is to provide important information to clinicians in order to predict tumor biology. Specific morphologic features may serve as predictive markers of tumor behavior. Macroscopic invasion of the perisellar tissues, defined as radiographic or gross operative finding, is considered a more consistent prognostic indicator. Regarding morphology, cytologic atypia is not a reliable feature. In contrast, the number of mitoses is very important for prognosis. Given that only scarce mitoses can be identified in some aggressive cases, Ki-67 represents an alternative key feature to assess tumor proliferation. In the recent World Health Organization classification, the Ki-67 labeling index (LI) represents a major prognostic indicator for pituitary adenomas. Expression of the p53 gene product is very important for tumor biology. Adenomas with more than 3% Ki-67 LI and extensive p53 immunoreactivity are classified as 'atypical adenomas'. Apoptosis and mitoses represent two adverse and asynchronous events, maintaining the optimal cell numbers. Using DNA labeling techniques, we can identify apoptotic cells. A higher apoptotic LI was found in functioning compared with nonfunctioning adenomas, in microadenomas, particularly in corticotrope adenomas, and in untreated adenomas, particularly prolactinomas. Cytogenetic analysis of chromosomes may provide important information regarding tumor development and progression. Increased chromosome 11 copies are more frequent in functioning, aneuploid pituitary adenomas. Monosomy or partial loss of chromosome 11 in adenomas with a normal or increased DNA LI indicates complex genomic abnormalities of chromosomes, other than chromosome 11. Immunohistochemical detection of somatostatin receptors is important, as their density in the cytoplasmic membrane is directly related to the effectiveness of somatostatin analogues. Therefore, morphologic assessment of the somatostatin receptor profile can predict the responsiveness and validate the effectiveness of treatment with somatostatin analogues. We can conclude that among the currently available predictive factors, tumor invasiveness is important, whereas the presence of mitoses, the Ki-67 LI, p53 expression and apoptosis are very important; DNA ploidy and fluorescent in situ hybridization analysis, although important, are difficult to apply. Finally, in the near future, immunohistochemistry for somatostatin receptors will be a very important application.