Abstract
Background: Several biomarkers indicating poor prognosis have been reassessed in patients with diffuse large B-cell lymphoma (DLBCL) treated by rituximab combination chemotherapy. However, no studies have investigated the outcome by combining these biomarkers for rituximab treatment. In addition, no large studies have reassessed the outcome of patients with CD5-positive DLBCL treated by rituximab. To determine the most influential biomarkers, we reassessed and investigated the predictive value of three biomarkers, namely Bcl-2 expression, GC phenotype and CD5 expression, in DLBCL patients treated with rituximab in combination with CHOP as an initial therapy.
Methods: A total of 121 patients with CD20-positive DLBCL receiving RCHOP at our institute between April 2002 and October 2005 were enrolled in the study. To classify the samples into immunohistochemically defined GC or non-GC phenotypes, we used a previously described algorithm using expression of CD10, Bcl-6, and MUM1 (Hans et al. Blood, 2004). For examination of the CD5 expression, we defined the cases as CD5-positive if CD5 expression was detected by immunohistochemical and flow cytometric analyses. We also assessed the outcome of patients according to International Prognostic Index (IPI).
Results: Expression of Bcl-2 was detected in 79 of the 121 (65%) patients, while CD5 expression was positive in 11 patients (9%). Overall, 48 of the 121 (40%) patients expressed a non-GC phenotype. CD5-positive patients displayed a significantly poorer progression-free survival (PFS) and overall survival (OS) than CD5-negative patients (2-year PFS, 18% vs. 73%, P<0.001; OS, 45% vs. 91%, P=0.001). However, there were no significant differences in PFS and OS according to Bcl-2 expression (PFS, 76% vs. 91%, P=0.08; OS, 77% vs. 97%, P=0.06) or GC phenotype (PFS, 70% vs. 90%, P=0.08; OS, 77% vs. 91%, P=0.12). In contrast, the differences between high or high-intermediate and low or low-intermediate of IPI for PFS and OS were significant (2-year PFS, 52% vs. 91%, P=0.001; OS, 64% vs. 92%, P=0.01). Multivariate analysis revealed that CD5 expression and the IPI were significant prognostic factors for PFS and OS (RR: PFS, 13.57 and 9.65, respectively; OS, 18.15 and 10.72, respectively) (Table 1).
Conclusion: These results demonstrated that CD5 expression was the most influential prognostic factor among the biomarkers examined and associated with a very poor outcome, even after rituximab treatment. To confirm this conclusion, further large studies of CD5-positive DLBCL patients are required.
Cox’s hazard regression analysis of PFS and OS
| . | Variable . | RR . | 95% CI . | P . |
|---|---|---|---|---|
| RR indicates relative risk; CI, confidence interval; GC phenotype, non-GC type worse; and IPI, higher IPI worse | ||||
| PFS | CD5 | 13.57 | 4.260–42.947 | <0.001 |
| BCL2 | 2.00 | 0.518–7.761 | 0.314 | |
| GC phenotype | 2.59 | 0.851–7.857 | 0.094 | |
| IPI 0 to2 or 3 to 5 | 9.65 | 2.911–31.987 | <0.001 | |
| OS | CD5 | 18.15 | 3.023–109.029 | 0.002 |
| BCL2 | 1.29 | 0.240–6.966 | 0.764 | |
| GC phenotype | 1.43 | 0.367–5.573 | 0.606 | |
| IPI 0 to2 or 3 to 5 | 10.72 | 1.945–59.085 | 0.006 | |
| . | Variable . | RR . | 95% CI . | P . |
|---|---|---|---|---|
| RR indicates relative risk; CI, confidence interval; GC phenotype, non-GC type worse; and IPI, higher IPI worse | ||||
| PFS | CD5 | 13.57 | 4.260–42.947 | <0.001 |
| BCL2 | 2.00 | 0.518–7.761 | 0.314 | |
| GC phenotype | 2.59 | 0.851–7.857 | 0.094 | |
| IPI 0 to2 or 3 to 5 | 9.65 | 2.911–31.987 | <0.001 | |
| OS | CD5 | 18.15 | 3.023–109.029 | 0.002 |
| BCL2 | 1.29 | 0.240–6.966 | 0.764 | |
| GC phenotype | 1.43 | 0.367–5.573 | 0.606 | |
| IPI 0 to2 or 3 to 5 | 10.72 | 1.945–59.085 | 0.006 | |
Author notes
Disclosure: No relevant conflicts of interest to declare.
