Spontaneous Apoptosis in Chronic Lymphocytic Leukemia Is Not an Independent Prognostic Factor for Stability of Disease When Compared with Combined AgNOR and TTM Scores

It has been shown that AgNOR staining can provide useful information for diagnosis in cytology of the oral cavity [22], body fluids [3,18], hematologic smears [10,11,14–17] or lymph node aspiration cytology [12]. We would like draw attention of the readers towards an easy AgNOR evaluation which has shown to be very useful as prognostic marker in chronic lymphoid leukemia (CLL) patients. This disease shows considerable variability of its clinical presentation and evolution. Patients in stable phase may, after months or years, change to a progressive phase. Whereas initially no treatment is warranted, chemotherapy will be necessary in the latter. In a previous exploratory study [17] we suggested that the percentage of CLL cells in peripheral blood with AgNOR clusters at first diagnosis might be an independent prognostic marker for the duration of the stable phase, besides the total tumor mass (TTM). An index created by summing up both values showed to be a more powerful prognostic factor than other common clinical and laboratory parameters [17]. In order to validate this model, we decided to repeat this study with other CLL patients. Since the apoptosis rate has shown to be of prognostic importance for various neoplasias and since resistance to apoptosis could be involved in the disease progression of B-CLL, [19], we also investigated the spontaneous apoptosis rate of CLL cells in culture. Unselected patients with newly diagnosed B-CLL entered the study. Counting of AgNOR clusters and the determination of TTM were done as previously described [10,16,17]. AgNOR staining of cytological preparations from acute leukaemias allows the differentiation of clusters (aggregations of precipitations within a common matrix in the nucleolus) and dots (small singular precipitations without a matrix). CLL clusters showed a longest chord of 2.07 μm (95% CI 1.85–2.4 μm) and compact nucleoli, 1.07 μm (95% CI 0.95–1.2 μm) (Fig. 1) [16]. Apoptosis rate was measured at diagnosis by flow cytometry as the percentage of annexin V positive cells after a 48 hours culture as described previously [19]. Pearson’s correlations were calculated between the variables. Univariate Cox regression analyses were performed to examine the relationship between the treatment-free period and percentage of annexin V positive cells, TTM, percentage of cells with AgNOR clusters in peripheral lymphocytes and the Index (= percentage of cells with AgNOR clusters + TTM). For all further analyses the Index values were logarithmized, in order to get a good approximation to the normal distribution. Then we tried to find out, whether comparing all these parameters in a multivariate Cox-regression, the Index would again be the strongest variable, as postulated previously [17]. Finally we tested the strongest variable from this regression together with the parameter “annexin V positive cells”. In order to estimate the stability of the models we applied the Cox regression to 200 new data sets created by bootstrap resampling [17]. For all calculations SPSS 8.0 software was used. During the study period 32 patients were analyzed. The mean observation time was 18 months and during this period 22 patients fulfilled the criteria for start of chemotherapy. Median time of the stable phase was 13 months in the Kaplan–Meier curve. Pearson’s correlations demonstrated statistically significant inverse correlations between the apoptosis rate and TTM (r = −0.47), the AgNOR score (r = −0.40) and the (logarithmized) Index (r = −0.52). In univariate Cox analyses TTM (Exp(B) = 1.087; CI 95% 1.027 to 1.15; p = 0.004), percentage of cells with AgNOR clusters (Exp(B) = 1.137; CI 95% 1.014 to 1.274; p = 0.027) and the Index (logIndex: Exp(B) = 5.84; CI 95% 1.83 to 18.6; p = 0.003) were unfavourable predictive variables. In a multivariate Cox regression only the logIndex remained in the final model. In the stability test by bootstrap resampling the variable log Index was included in 69.5% of all models, whereas TTM in only 33% and the AgNOR score in only 29.5% of the cases. Spontaneous apoptosis rate was a favourable prognostic factor for treatment-free period in the uni-