ABSTRACT
PURPOSE: The predictive potential of tumor cell kinetic parameters may be improved when they are studied in relation to other microenvironmental parameters. The purpose of this investigation was to quantitatively categorize human tumor samples according to proliferation patterns. Second, it was examined whether these characteristics are retained after xenotransplantation. METHODS AND MATERIALS: Fifty tumor samples from head-and-neck cancer patients were immunohistochemically stained for Ki-67 and vessels. Also, parts of the samples were transplanted into nude mice. Tumors were categorized according to previously described patterns of proliferation. Vascular and proliferation patterns were analyzed using an image processing system. RESULTS: The 50 tumors were categorized into four patterns of proliferation by visual assessment: marginal (6), intermediate (10), random (21), and mixed (12). One tumor could not be classified. These patterns were quantified by calculating the Ki-67 labeling index in distinct zones at increasing distance from vessels yielding good discrimination and significant differences between patterns. The probability of growth after xenotransplantation was significantly higher for tumors with a labeling index and vascular density above the median value compared to tumors with both parameters below the median (82% vs. 35%). Fifty percent of the tumors retained their proliferation patterns after xenotransplantation. CONCLUSION: The categorization by proliferation pattern previously described by others was reproduced quantitatively and spatially related to the vascular network using a computerized image processing system. The combination of quantitative and architectural information of multiple microenvironmental parameters adds a new dimension to the study of treatment resistance mechanisms. Tumor models representative of the various patterns can be used to further investigate the relevance of these architectural patterns.
