Abstract 3957

Poster Board III-893

BACKGROUND

Despite recent therapeutic improvements, the clinical course of DLBCL still differs considerably among patients. Here, we present the final results of a study to identify distinct DLBCL subgroups in patients treated with R-CHOP.

METHODS

DNA from 166 frozen DLBCL samples (PMBL,HIV-related DLBCL, post-transplant DLBCL and Richter's Syndromes were excluded) was analyzed with Affymetrix Human Mapping 250K arrays. Affymetrix U133 PLUS 2.0 gene expression profiles (GEP) were available for 54 cases. An optimized non-negative matrix factorization (NMF) was used for unsupervised clustering. The impact of the recurrent lesions on OS, PFS and DFS was evaluated with the log-rank test followed by multiple test correction (MTC).

RESULTS

20 lesions showed a statistical significant impact on OS. Only 8p23.1 loss maintained its significance after MTC and was associated with additional aberrations, such as 17p- and 15q-. Unsupervised clustering identified 5 clusters with distinct genetic profiles, clinical characteristics and outcome: 1 (61/166, 37%), 2 (22/166, 13%), 3 (5/166, 4%), 4 (29/166, 17%), 5 (7/166, 6%). Cluster 1 was characterized by a heterogeneous genomic profile, lacking the most recurrent lesions. When compared to cluster 2 and 4, there were more HCV infections (9/38, 24% vs 1/12, 8% and 0/20; p=0.014) and bone marrow involvement (16/57, 28% vs 1/19, 5% and 2/24, 8%; p=0.008). GEP and morphology data suggested that cluster 1 could represent DLBCL with a high content of infiltrating T-cells, partially explaining the low rate of aberrations. This cluster has characteristics similar to the “host response” cluster previously identified by Monti et al., 2005. The outcome was poor with a high relapse rate (13/49, 26%) and a 5-yr OS of 76%. Cluster 2 (1q+/3q+/6q-/17p-) had also a relatively poor OS 80%. Cluster 4 (7+/12+) had a good outcome with a 5-yr OS of 92%. Clusters 2 and 4 showed differences in the number of pts with a high IPI-score (9/16, 56% vs. 7/26, 27%; p=0.057) and in single IPI factors, CR-rate (17/22, 77% vs. 25/29, 86%) and relapse rate (5/17, 29% vs. 2/25, 8%; p= 0.068). Cluster 3 (6q-/9-/11q+) had a trend for a good outcome. Instead, cluster 5 (3q+/11q+/18q+/6q-/8p-/15q-/17p-) showed a poor outcome. Clusters 2 and 4 were similar to ABC and GCB, respectively.

CONCLUSIONS

ArrayCGH provides the capability to identify genetic features and clusters, associated with a different outcome in patients treated with R-CHOP. The 8p23.1 locus should be investigated further for a gene that may be important in the pathogenesis of DLBCL.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution