Bromodeoxyuridine (BUDR) is a thymidine analog which is incorporated into the DNA of proliferating cells. Since the dose of BUDR needed to label cells is not toxic, cell labelling can be accomplished in vivo, by infusing the substance in patients. A monoclonal antibody against BUDR is then used to identify BUDR-labelled cells. The same cell population can also be stained for DNA content with propidium iodide (PI). Using bivariate flow cytometry (FCM) for measurements, both the percentage of BUDR-labelled cells and their total DNA content can be evaluated. This technique allows one to obtain the labelling index (LI) and the DNA synthesis time (TS). The potential doubling time (Tpot) and the fractional turnover rate (FTR) can be mathematically derived, so that a complete picture of tumor growth can be obtained. Our aim was to ascertain whether this method is clinically applicable and whether the kinetic values obtained are reliable. We studied 22 patients with benign and malignant brain tumors, and observed no immediate toxicity from BUDR administration. The BUDRLI obtained ranged from 0.9% to 3.9% (median: 2.0%) in meningiomas and from 3.8% to 7.6% (median: 6.3%) in malignant gliomas (P less than 0.01). The fraction of S-phase cells determined with the BUDR FCM technique was statistically similar to that found by single DNA flow cytometric analysis performed on duplicate samples of both benign and malignant brain tumors. The TS obtained in malignant gliomas ranged from 10.5 to 227 h (median: 12.8). The calculated Tpot ranged from 7.6 to 26.8 days (median: 11.6), and the calculated FTR ranged from 3.7 to 13.1 cells/100 cells/day (median: 8.8). These data suggest that in vivo BUDR infusion coupled with FCM can be performed in clinical settings, and it is reliable and can easily be used for kinetic studies in clinical trials aimed at evaluating the prognostic relevance of proliferative parameters and in planning tumor treatment.