Prognostic implications of different cell cycle analysis models of flow cytometric DNA histograms of 1,301 breast cancer patients: results from the Multicenter Morphometric Mammary Carcinoma Project (MMMCP).

ABSTRACT

Conflicting prognostic results with regard to DNA flow cytometric cell cycle variables have been reported for breast cancer patients. An important reason for this may be related to differences in the interpretation of DNA histograms. Several computer programs based on different cell cycle fitting models are available resulting in significant variations in percent S-phase and other cell cycle variables. Our present study evaluated the prognostic value of percent S-phase cells obtained using 5 different cell cycle analysis models. Flow cytometric DNA histograms obtained from 1,301 fresh frozen breast cancer samples were interpreted with 5 different cell cycle analysis models using a commercially available computer program. Model 1 used the zero order S-phase calculation and "sliced nuclei" debris correction, model 2 added fixed G2/M- to G0/G1-phase ratio, and model 3 added correction for aggregates. Model 4 applied the first-order S-phase calculation and sliced debris correction. Model 5 fixed the coefficients of variation CVs of the G0/G1- and G2/M-phases in addition to applying the sliced nuclei debris correction and zero order S-phase calculation. The different models yielded clearly different prognostic results. The average percent S-phase cells of the aggregate correction model (model 3) provided the best prognostic value in all cases for overall survival (OS) as well as disease-free survival (DFS) (OS p < 0.0001; DFS p < 0.0001), in lymph node-positive cases (OS p < 0.0001; DFS p = 0.004) and in DNA-diploid subgroups (OS p = 0.004; DFS p = 0.001). For the lymph node negative and DNA-non-diploid subgroups, the percent S-phase of the second cell cycle reached slightly better prognostic significance than the average percent S-phase cells. In multivariate analysis, the average percent S-phase of the aggregate correction model had the best additional prognostic value to tumor size and lymph node status. In conclusion, different cell cycle analysis models yield clearly different prognostic results for invasive breast cancer patients. The most important prognostic percent S-phase variable was the average percent S-phase cells when aggregate correction was included in cell cycle analysis.