This paper addresses the problem of detecting weak incorporation of BrdUrd and the related efficiency of the denaturation protocols used to unmask this thymidine analog. Evidence is presented that measuring the distribution of BrdUrd-tagged fluorescence intensities by image cytometry generates a standard deviation threshold that discriminates between positive and negative MRC5 cells in vitro. A comparison of the thresholding by standard deviation (SDT) with the usual thresholding by the nuclear total fluorescence intensity (FIT) demonstrated that SDT has a significantly higher sensitivity (99.4-100%, depending on the denaturation protocols) than FIT (94.7 and 74.3%, respectively), although both tests have a high specificity (93% and 100%, respectively) for detecting S cells. Since detecting the S cells is not only dependent on the test used, but also on the denaturation protocols, a quality index (QI) was derived from the standard deviation and the mean value of the non-specific fluorescence of negative cell population versus BrdUrd fluorescence of positive cell population. The following DNA denaturation protocols have been assessed according to QI: acidic denaturation, thermal denaturation in formamide, and thermal denaturation in distilled water. Each denaturation procedure was preceded or not by incubation in either proteinase K or Triton X-100. The results showed that thermal denaturation in formamide, especially when preceded by proteinase K incubation, revealed the largest difference between negative and positive cells. This work also demonstrated that image cytometry of BrdUrd-labelled cells can be suitable for clinical application because of the high sensitivity provided and the small samples needed.