Analysis of ZAP-70 Expression in Chronic Lymphocytic Leukemia (CLL) Using Kolmogorov-Smirnov (KS) Statistics.

ZAP-70 is a prognostic marker in CLL and has been shown to correlate closely with V_H mutational status in most studies. Assays that require the setting of markers using either T-cells, or a matched isotype control to determine ZAP-70 expression in CLL are subjective and prone to analytical and inter-operator variability. We examined the use of the KS statistical test to analyze histogram overlays as an objective approach to the assessment of ZAP-70 expression in 18 CLL patients with known V_H mutational status (V_H mutated; n=11, V_H unmutated; n=7). Use of the KS statistical test to compare ZAP-70 expression of T cells with B cells in CLL has been recently described by Van Bockstaele et al.(Cytometry Part B 70B:302-308. 2006). In contrast to this approach we used a B cell gating assay with a pre-conjugated alexa-fluor 488 ZAP-70 antibody clone 1E7.2 and an alexa-fluor 488 matched isotype control manufactured by Caltag Laboratories (Burlingame, CA) as previously described by our group. Modifications were made to allow the determination of KS ‘D’values. Statistical analysis of our resulting KS data using a t-test demonstrated a significant difference in ZAP-70 KS ‘D’ values between the two V_H mutational groups (p=0.0001). All patients with a KS ‘D’ value of 0.5 or below were V_H mutated, whilst all patients with a KS ‘D’ value of 0.7 or above were V_H unmutated, with some cross-over between these values. At a threshold KS ‘D’ value of 0.6, 10 of the 11 V_H mutated patients were correctly classified (1 patient had a ‘D’ value of 0.68), whilst 6 of the 7 V_H unmutated patients were correctly classified (1 patient had a ‘D’ value of 0.59) The sensitivity and specificity of using a threshold KS ‘D’ value of 0.6 was 86% and 91% respectively. Analysis of ZAP-70 expression using the KS statistic is operator independent and thus provides a means for a more reproducible analytical approach in the laboratory. We acknowledge that our initial sample size was small and requires further investigation using a larger sample size to fully validate the diagnostic value of using KS ‘D’ values to determine ZAP-70 expression in CLL.