Ipi

Risk group Number of adverse factorsa 5 year RFS (%) 5 year OS (%)

Low-intermediate 2 50 51

High-intermediate 3 49 43

High 4 or 5 40 26

Age adjusted IPI

Risk group Number of adverse factorsb 5 year RFS (%) 5 year OS (%)

Low 0 86 83

Low-intermediate 1 66 69

High-intermediate 2 53 46

High 3 58 32

RFS, relapse-free survival; OS, overall survival.

aAdverse risk factor for IPI are stage III or IV disease, age >60 years, elevated LDH, ECOG performance status >2, >2 extranodal sites. bAdverse risk factor for age-adjusted IPI are stage III or IV disease, elevated LDH, ECOG performance status >2.

Table 66.4 Examples of individual immunophenotypic eatures with reported prognostic significance in DLBCL

Immunophenotype

Impact on prognosis

bcl-2 expression

Adverse

Mutated p53

Adverse

High proliferative rate defined by Ki-67

Adverse

Cyclin D2 positive

Adverse

CD 5 expression

Adverse

MUM-1 positive

Adverse

bcl-6 expression

Favorable

HLA class II expression

Favorable

Tumor infiltrating lymphocytes

Favorable

Alizadeh et al. initially reported results using the lymphochip cDNA microarray to analyze biopsies from 44 patients with DLBCL.23 In this study, patients with high levels of expression of genes characteristic of normal germinal center B-cells (GCB) were shown to have higher overall survival rates than those with expression profiles characteristic of activated B-cells (ABC). A study of material from 77 uniformly treated DLBCL patients reported by Shipp et al. using olignonucleotide microarrays identified a predictive model based on expression of 13 genes.24 Rosenwald et al. subsequently described a 17-gene predictive model based on further studies using a cDNA microarray which identified four gene expression signatures characteristic of GCB, proliferating cells, reactive stromal and immune cells in the lymph node, and expression of major histocompatibility (MHC) class II antigens.25 There was no concordance between individual genes identified in the 13-and 17-gene models previously described, possibly because the microarrays used in the studies differed and different techniques were used to develop the predictive models. The potential clinical utility of prognostic scores based on microarrays is limited by these technical differences, as well as the requirement for fresh, or optimally cryopreserved samples, and by the costs of these techniques. In an attempt to overcome some of these problems Lossos et al. have described a simplified six-gene model using quantitative RT-PCR in 66 patients with DLBCL, all treated with CHOP or related regimens.26 Genes correlated with shorter survival were BCL2, CCND2, and SCYA3. Those associated with longer survival were BCL6, LMO2, and FN1.

Results from all of these studies have been shown to provide prognostic information which helps to refine the predictive value of the IPI. However, in view of the limitations of GEP studies outlined earlier, there has been recent interest in the use of TMAs which allow immunohistochemical analysis of protein expression from multiple tissue sections on single slides, and which can be performed on routinely fixed and paraffin-embedded clinical specimens.

A recently published study of TMAs in DLBCLs has examined expression of CD10, bcl-6, MUM1, FOXP1, cyclin D2, and bcl-2 in samples from 152 patients with DLBCL, of which 142 had previously been analyzed using GEPs.27 Using bcl-6, CD10, and MUM1 expression, it was possible to identify samples as GCB-like or ABC-like. As shown in Table 66.5, there was a marked difference in event-free and overall survival according to phenotype, the survivals being very similar to those described for the same series classified according to GEPs. High IPI score and non-GCB phenotype were independent adverse prognostic factors in multivariate analysis. Subsequent studies have confirmed the adverse prognostic significance of non-GCB pheno-type identified by TMAs.28'29 These early results suggest that TMAs are likely to have more clinical utility than GEPs since the availability of frozen material will remain a limitation for genetic studies. A small number of immunohistochemical stains, rather than multiple stains used in TMAs may prove to have adequate predictive value in future studies. Prospective evaluation of GEPs, TMAs, and expression of individual proteins by immunohistochemistry should continue to be an integral part of new clinical trials in DLBCL in an attempt to produce more specific prognostic factors. This is particularly true since the addition of rituximab to combination chemotherapy for DLBCL. Several studies have demonstrated the superiority of ritux-imab/chemotherapy combinations compared with chemotherapy alone in the treatment of DLBCL. However, to date, all of the published data regarding TMAs and GEPs as prognostic tools is based on samples from patients treated with chemotherapy only. Recent reports have shown that the addition of rituximab to chemotherapy can modify the prognostic significance of certain factors. For example, the adverse prognostic effect of absent bcl-6 expression in DLBCL patients treated with CHOP does not apply to patients receiving CHOP-rituximab.30 Two recent studies have demonstrated that the adverse prognostic significance of bcl-2 expression in DLBCL is lost in patients treated with chemotherapy/rituximab combinations.31,32

At present, the IPI and aa-IPI should be regarded as the standard system for identifying risk groups in DLBCL. Ongong studies of GEPs and TMAs based in samples from patients treated with rituximab-based

Table 66.5 Tissue microarray criteria for GCB versus non-GCB derivation of DLBCL
0 0

Post a comment