R-CHOP Therapy Cannot Overcome CD5 Positive Non-GCB Subtype of DLBCL

Introduction: There are various poor prognosis factors in diffuse large B cell lymphoma (DLBCL). CD5 positive (CD5⁺) is estimated as one of the poor prognosis markers in DLBCL. CD5⁺ DLBCL is completely distinguished from CLL or Mantle cell lymphoma, and de Novo CD5⁺ DLBCL is related to a high incidence of cytogenetic abnormalities of 8p21 and 11q13. In many cases, CD5⁺ DLBCL is associated with an aggressive clinical status and advanced stages. Several chromosomal studies have demonstrated the gene expression was similar to non-GCB type of DLBCL. In the clinical phase, it is easy to express chemo-resistance and CNS invasion. However, the clear characterization and mechanisms of chemo-resistance have not been demonstrated yet. In this study, we examined our previous CD5⁺ DLBCL patients in our institute, then evaluated the prognosis and clinical characteristics regarding GCB or non GCB type.

Methods: We studied 372 newly diagnosed DLBCL patients including 42 cases of CD5⁺ from 2005 to 2015 in our institution retrospectively. The pathological diagnoses were performed with immunohistochemical analysis by two or three hematological pathologists. CD5 expression was evaluated by immunohistostaining and flowcytometry, then cyclin D1 positive cases were excluded. GCB or non-GCB subtype was evaluated with CD10, bcl-6, and MUM-1 of immunohistostaining. The clinical stage of patients and evaluation of the effect of the therapy were performed using PET-CT scan. The statistical analyses were performed by Dr. SPSS II.

Results: In all our treated patients, 350 DLBCL patients (female;161) and 41 CD5⁺ patients (female 22) could be evaluated. 192 patients were GCB type and 158 patients were non-GCB type. The median age was 64.5 yrs (25-86 yrs) and the median follow up time was 45 months (1-133 months). All patients were treated by rituximab-CHOP (R-CHOP) therapy. CR rate of CD5^- DLBCL was 94.2% and CD5⁺ was 73.1% respectively (p =0.0093). 3.5-year event free survival (EFS) was 79.16% and 52.20% (p <0.0001) and overall survival (OS) was 85.82% and 54.35 %(p <0.0001) (CD5^- and CD5⁺ respectively). In CD5⁺ DLBCL, GCB type was 11 (36.6%) and non-GCB type was 30 (63.3%). In the non-GCB type of CD5⁺ cases, IPI low (L) and low intermediate (LI) was 86.6% (18 and 8 each) and high intermediate (HI) was 13.3% (4/30). CR of GCB type in CD5⁺ was 81.8% and relapse rate was 22.2% compared with CR of non-GCB in CD5⁺ was 70.0% and relapse rate was 52.4% respectively. All HI cases of non-GCB type could not get CR and died within 6 months. 3.5-year EFS of CD5⁺ DLBCL according to GCB and non-GCB were 81.82% and 40.6% (p =0.1214) and OS was 90.9% and 40.05% (p =0.0154), respectively. 3.5-year EFS of CD5^- DLBCL according to GCB and non-GCB were 82.12% and 69.23% (p =0.0173), and OS of CD5^- was 88.51% (GCB) and 82.12%(non-GCB) (p =0.0061) respectively.

Discussion: CD5⁺ showed poor prognosis in our treated DLBCL, however, GCB type of CD5⁺ had similar OS and EFS as CD5^- DLBCL. Non-GCB type of CD5⁺ demonstrated significantly poor prognosis compared to GCB subtype or CD5^- DLBCL. The clinical status of CD5⁺ cases have been generally reported as having aggressive status, however, in our CD5⁺ of non-GCBcases, IPI L and LI were 86%. In spite of many populations of L and LI of IPI, the prognosis was poor. From these results, it suggested that it is not necessary to estimate CD5⁺ of GCB subtype as poor prognosis in R-CHOP therapy, however, CD5⁺ non-GCB type indicated independent poor prognosis factors in DLBCL. Additionally, R-CHOP therapy could induce CR at once in IPI L and LI cases, however it could not contribute to good prognoses. We should investigate other treatment strategies for improving outcomes of CD5⁺ non-GCB subtype of DLBCL.