Cell of Origin in Richter's Transformation of CLL

Abstract 4574

Richter's transformation (RT) from chronic lymphocytic leukemia (CLL) to diffuse large B-cell lymphoma (DLBCL) occurs in about 5% of pts and confers significant morbidity and mortality. In de novo DLBCL the cell of origin (COO), either activated B-cell like (ABC) or germinal center B-cell like (GCB), is prognostic. The ABC group has been shown to have both a less favorable response to chemoimmunotherapy and a shorter overall survival than the GCB group. However, COO has not been previously described for pts with RT. The aim of our study was to determine the pattern of COO in RT and to determine if prognostic correlates exist.

Clinical information on 30 pts with CLL and pathologically-confirmed RT were obtained from the CLL cohort treated at our institution from 1980 to July, 2012, and, of these, 13 pts had sufficient data for inclusion in the analysis. Paraffin-embedded DLBCL tissue specimens were subjected to antigen retrieval and antibody staining via commercially manufactured products. CD10, BCL-6, and MUM-1 were the antigens identifiable by staining and used in the Hans method for COO determination. Immunoglobulin heavy chain VDJ-PCR was performed on CLL and RT tissue specimens and compared in order to establish clonality on some tissue specimens. Kappa and lambda immunophenotyping was compared between CLL samples and RT tissue samples for clonal assessment as well.

Of the 13 pts evaluated, 10 were ABC-type and 3 were GCB-type. Median age was 59 yrs, 7 male, 6 female, 3 ZAP 70 +, 6 IgVH mutated. EBER staining was positive in 2/13 RT pts. For the ABC-type, median time to transformation (TTT) was 25 mos (0–288 mos) for the GCB- type median TTT was 48 mos (48–100 mos). The one-year OS for pts after RT was 38%; 40% for ABC-type and 33% for GCB-type. For the ABC-type, survival after transformation (SAT) was a median of 10.5 mos (1–96 mos), for the GCB-type median survival was 3 mos (9 days-14+ mos). The pts with the longest TTT (98 and 288 mos), had an IPI of 2. Those with shortest TTT (0, concurrent diagnosis of RT and CLL) had IPIs of 5.

Of the pts with ABC-type: 5 were previously treated with chemotherapy for CLL; 4/5 fludarabine-based regimens, and 2 with rituximab monotherapy. Available treatment information for RT in ABC-type: 1 with HyperCVAD, 4 with RCHOP, 1 with HyperCVAD followed by REPOCH. One pt went on to BMT, and one required IT chemotherapy for CNS involvement. For the 3 pts with GCB-type, 2 received prior treatment for CLL with FR, the other pt was previously untreated. Treatment for RT in the GCB-type included: R-CHOP, DCDT-2980S (an investigational drug–antibody conjugate), and ongoing treatment on a clinical trial with ibrutinib in one pt who remains alive, radiation therapy for another who died 3 mos after RT, and one pt who died 9 days after RT and was not treated for transformed disease.

Five of the 11 evaluable samples were found to be of a different clone from the original CLL and 6/11 were the same clone based on VDJ-PCR and immunophenotype. Longer TTT (48 mos or more) correlated with the finding of a different clone in the RT. The 2 pts who had synchronous diagnoses of RT and CLL were from the same clone.

A higher prevalence of ABC-type DLBCL was noted in this series of RT pts. Among 10 ABC-type RT pts, there was a shorter median TTT compared with GCB–type RT pts. In the ABC-type RT pts there was a correlation between longer TTT and longer survival, which may be reflective of the underlying biology of the lymphoma within the ABC-type. This correlation was not seen in the GCB group. An association between the type of chemotherapy used in the CLL setting and the COO was not identified. In the ABC-group, survival was not impacted by type of chemotherapy used in RT. In contrast to de novo DLBCL, there was a suggestion of a more favorable outcome with the RT ABC-type. In those cases of longer TTT (4 yrs +), the RT clone appeared different than the original CLL based on PCR or immunophenotype. This difference may represent a CLL clone that escaped chemotherapeutic destruction and then developed into RT, or the possibility of a newer clone that emerged independent of the original lymphoma and confirms what has been previously described, that all RT are not clonally evolved but a de novo event in some cases.

Though a small sample size, the results of this series of pts is hypothesis-generating. Improved prognostic accuracy and risk-directed therapeutic strategies may become possible with an improved classification of COO of RT.