We describe a biodistribution study using a radiolabelled anti-tumor murine monoclonal antibody F(ab')2 in XF-8 in 6 patients with colorectal carcinoma. One patient had an isolated rectal carcinoma. Five patients had hepatic metastasis of colorectal carcinoma. Among them three had also the primary colorectal carcinoma. Double-label method using simultaneous injection of 131-Iodine monoclonal antibody XF-8 F(ab')2 and 125-Iodine control monoclonal antibody was performed to evaluate the specificity of monoclonal antibody localization in the tumor. Each patient received 4-8 mCi of 131-I XF-8 F(ab')2 infused at doses 0.5-5 mg. Simultaneously 200 microCi of 125-I non relevant antibody were infused. Before each administration, immunoreactivity of antibodies preparations was checked before labelling by immunohistological methods and, after labelling, by cell-binding tests. In biodistribution studies, patients were scanned with a scintillation camera each day since the infusing day to the surgery day (3-6 days). To study specificity of XF-8 F(ab')2 antibody localization we measured by means of scintillation counting in tumors and normal tissues recovered after surgery. In vitro immunoreactivity of monoclonal antibody batches was safe. All our data in vivo are negative but correlated. We concluded that criteria for screening the monoclonal antibodies considered for detection of human tumors are not really known.