Based on the premise that optimal drug delivery might improve the efficacy of locoregional treatment for solid tumors, the authors set up an experimental model for isolation perfusion in surgical specimens from patients resected for carcinoma of the colon. Ten surgical specimens were cannulated, washed internally and externally with saline solution, promptly cooled to 4 degreesC, connected to a circuit, and perfused with Krebs-Henselait modified solution, concentrated red blood cells, albumin, desamethasone, glucose, and heparin for 60 minutes at a target temperature of 37 degreesC. Organ temperature, flow rate, perfusion pressure, and metabolic and functional parameters were checked at 5, 20, and 60 minutes of perfusion. A paraphysiologic perfusion procedure was achieved. Mean values (and ranges) were as follows: temperature 37 degreesC (35. 1-39.6 degreesC); flow rate 10.2 (5.6-17.9) ml/min/100 g; arterial pressure 96 (42-154) mmHg; arterial pH 7.3 (7.1-7.5); arterial PO2 183 (78-304) mmHg; arterial PCO2 36 (31-46) mmHg. No important signs of tissue damage were found at histology. Autonomous or stimulated peristalsis (or both) was present throughout the experiment. Mean O2 extraction was 7.9 ml/min/100 g (range 3.1-11.0). Mean glucose consumption was 229 mg/100 g (range 174-252). The model worked well and appears promising, particularly for future use in various pharmacokinetic and pharmacodynamic studies of antiblastic agents.