Surface-antigen expression profiling of B cell chronic lymphocytic leukemia: from the signature of specific disease subsets to the identification of markers with prognostic relevance

Abstract

Studies of gene expression profiling have been successfully used for the identification of molecules to be employed as potential prognosticators. In analogy with gene expression profiling, we have recently proposed a novel method to identify the immunophenotypic signature of B-cell chronic lymphocytic leukemia subsets with different prognosis, named surface-antigen expression profiling. According to this approach, surface marker expression data can be analysed by data mining tools identical to those employed in gene expression profiling studies, including unsupervised and supervised algorithms, with the aim of identifying the immunophenotypic signature of B-cell chronic lymphocytic leukemia subsets with different prognosis. Here we provide an overview of the overall strategy employed for the development of such an "outcome class-predictor" based on surface-antigen expression signatures. In addition, we will also discuss how to transfer the obtained information into the routine clinical practice by providing a flow-chart indicating how to select the most relevant antigens and build-up a prognostic scoring system by weighing each antigen according to its predictive power. Although referred to B-cell chronic lymphocytic leukemia, the methodology discussed here can be also useful in the study of diseases other than B-cell chronic lymphocytic leukemia, when the purpose is to identify novel prognostic determinants.

Figure 1

From the hierarchical clustering to the immunophenotypic signature of subsets with different prognosis. See text for details.

Figure 2

From the signature of subsets with different prognosis to the definition of a comprehensive prognostic scoring system and the division of patients into prognostic groups. See text for details.

Figure 3

Surface antigens identified by the nearest shrunken centroid algorithm and their role as putative prognosticators for B-CLLs. For each antigen identified as part of the immunophenotypic signature, are here reported: i) its expression levels in the three immunophenotypic groups as identified by nearest shrunken centroids (↑ or ↓ indicate above- or below-average expression, respectively; = indicates average expression level) [17]; ii) its association (Y) or lack of association (N) with good or bad prognosis and the relative z scores as resulted by Cox proportional hazard regression analysis [22]; iii) the cutoff values, as resulted by Maximally selected log-rank statistics analysis, and the score values assigned when its expression was found to be above or below the established cutoff [22]. Data and values of the antigens selected for the final prognostic score are reported in red.

Figure 4

Translation of the prognostic values for the identified six antigens into a scoring system. Each of the six antigens previously identified as the most powerful prognosticators [17,22] was associated with a score of 0, 1 or 2 according to its predictive power (z score) and the expression above or below the established cut-off values. Each of the reported bar, corresponding to the theoretical expression values (reported as % of positive cells) for each antigen, is depicted in grey (score "0" zone), azure (score "1" zone) or blue (score "2" zone). See text for detailed discussion of score assignment and cut-off values.