This study investigates cryodestruction of the Dunning AT-1 rat prostate tumor at the single cell, tissue slice, and in vivo levels. The thermal history around a 3-mm-diameter cylindrical cryosurgical probe was predicted by solving the bioheat equation in a one-dimensional cylindrical geometry. At various radial positions in the iceball this thermal history was approximated by a constant cooling rate and a final, steady-state temperature (or end-temperature). The predicted cooling rates and end temperatures ranged from > or = 1000 degrees C/min to 5 degrees C/min and -196 degrees C to -20 degrees C, respectively. These cooling rates and end-temperatures were then imposed on single AT-1 cells, AT-1 tissue slices in vitro and AT-1 tumors in vivo. The single cells and tissue slices were frozen by LN2 immersion, copper block slam-freezing, or controlled cooling on a cryomicroscope or a directional solidification stage. LN2 immersion is lethal to AT-1 cells (presumably due to intracellular ice formation), while cooling at 5-100 degrees C/min leaves some viable cells (at end-temperatures ranging between -20 and -40 degrees C). AT-1 tumor slices show extensive intracellular ice formation due to slam cooling, extensive dehydration at 100 degrees C/min, and total dehydration at rates < or = 10 degrees C/min to end temperatures below -10 degrees C. Postfreeze culture and histology of the AT-1 tissue show that extensive intracellular ice formation is lethal, while cellular dehydration and vascular engorgement leave viable cells (at end-temperatures between -20 and -40 degrees C). Based solely on the single cell and in vitro tissue damage achieved by cooling rates and end-temperatures, a sizable portion of a cryosurgically frozen tumor would be expected to survive. However, in vivo cryosurgery performed on AT-1 tumors demonstrated that the tissue was damaged throughout the cryolesion, even at the periphery where the thermal history would be expected to allow single cells and tissue slices to survive in vitro. Taken together, these results suggest that damage mechanisms other than those due to cooling rate and end-temperature may be responsible for the increased cellular destruction at the periphery of the iceball in vivo and that cooling rate is less important than end-temperature in determining cryosurgical damage in AT-1 tumors. Experiments are ongoing to determine if the time held at an end temperature, thawing rate, vascular response, or other mechanisms are primarily responsible for the enhanced destructive capability in vivo.