Purpose: The aim of this study was to assess the potential of pre-treatment cell kinetic parameters to predict outcome in head and neck cancer patients treated by conventional radiotherapy.Materials and methods: Data from 11 different centers were pooled. Inclusion criteria were such that the patients received radiotherapy alone, and that the radiotherapy was given in an overall time of at least 6 weeks with a dose of at least 60 Gy. All patients received a tracer dose of either iododeoxyuridine (IdUrd) or bromodeoxyuridine (BrdUrd) intravenously prior to treatment and a tumor biopsy was taken several hours later. The cell kinetic parameters labeling index (LI), DNA synthesis time (Ts) and potential doubling time (Tpot) were subsequently calculated from¯ow cytometry data, obtained on the biopsies using antibodies against I/BrdUrd incorporated into DNA. Each center carried out their own¯ow cytometry analysis.Results: From the 11 centers, a total of 476 patients conforming to the inclusion criteria were analyzed. Median values for overall time and total dose were 49 days and 69 Gy, respectively. Fifty one percent of patients had local recurrences and 53% patients had died, the majority from their disease. Median follow-up was 20 months; being 30 months for surviving patients. Multivariate analysis revealed that T-stage, maximum tumor diameter, differentiation grade, N-stage, tumor localization and overall time correlated with locoregional control, in decreasing order of signi®cance. For the cell kinetic parameters, univariate analysis showed that only LI was signi®cantly associated with local control (P 0:02), with higher values correlating with a worse outcome. Ts showed some evidence that patients with longer values did worse, but this was not signi®cant (P 0:06). Tpot showed no trend (P 0:8). When assessing survival in a univariate analysis, neither LI nor Tpot associated with outcome (P 0:4, 0.4, respectively). Surprisingly, Ts did correlate with survival, with longer values being worse (P 0:02). In the multivariate analysis of local control, LI lost its signi®cance (P 0:16). Conclusions: The only pretreatment kinetic parameter for which some evidence was found for an association with local control (the best end-point for testing the present hypothesis) was LI, not Tpot, and this evidence disappeared in a multivariate analysis. It therefore appears that pretreatment cell kinetic measurements carried out using¯ow cytometry, only provide a relatively weak predictor of outcome after radiotherapy in head and neck cancer. q