We have previously demonstrated the high tumor targeting potential of the thymidine analogue 125IUdR in experimental animal models following direct intratumoral or locoregional (intracavitary) administration. The aim of the present work was to evaluate the metabolism and selectivity (based on differential cell proliferation kinetics) of 125IUdR incorporation in patients with breast cancer following a similar approach. 125IUdR (4-8 MBq) was injected intratumorally by ultrasound-guided percutaneous injection in 7 patients with breast cancer 24 hours before ablative surgery. Blood and urine samples were collected up to 72 hours after injection and analyzed by HPLC using a C18 reversed-phase column and methanol:water (20:80) as the mobile phase. Following resection, the radioactivity of the tumor and the surrounding tissues was measured in a gamma counter, and microautoradiography was performed on semithin tissue sections to determine the site of tracer incorporation at the cellular level. Activity in plasma peaked at 0.5 to 1 hour after 125IUdR injection (4.96 +/- 1.08% of injected dose/liter), declining thereafter with a mean T1/2 of 11.24 +/- 2.78 hours. By HPLC analysis, undegraded 125IUdR was about 15-30% of total plasma activity, with a biphasic pattern peaking at both 1-3 hours and approximately 12 hours. In addition to free 125I-, about 10% of early plasma activity was constituted by a labeled metabolite (tentatively identified as radio-iodouracil), rising to about 50-60% at later time points. About 70-90% of urinary radioactivity was 125I-, and 5-20% was undegraded 125IUdR in the first 24-hour samples, while the remainder was iodouracil. High tumor/nontumor ratios were obtained (mean 147.4 +/- 125.2, range 27-397) with average tumor/blood ratios at the time of surgery equal to 32.7 +/- 18.6 (range 5-56). An average 0.0244 +/- 0.0189% of the injected dose was present per gram of tumor (range 0.001-0.061% ID/g). Microautoradiography confirmed the high values of tumor/nontumor incorporation ratios and demonstrated the specificity of 125IUdR incorporation mostly in the tumor cell nuclei, with only occasional incorporation by normal-appearing tubular cells. These results suggest the potential of radiolabeled IUdR for tumor targeting in humans, to be used whenever a satisfactory route of locoregional administration allowing for homogeneous tracer distribution within the tumor mass is accessible and in the presence of favorable tumor cell proliferations kinetics.