Analysis of Prognostic Factors and Outcome of Patients with Newly Diagnosed Diffuse Large B-Cell Lymphoma of the Gastrointestinal Tract

Introduction: The gastrointestinal tract is one of the extranodal sites more frequently involved in patients with diffuse large B-cell lymphoma (DLBCL). However, DLBCL with gastrointestinal involvement represents a very heterogeneous disease encompassing various biological entities. Therefore, optimal therapeutic strategies still remain a matter of debate.

Methods: Patients with newly diagnosed DLBCL and gastrointestinal involvement treated at the University Hospital of Heidelberg, Germany, between January 2000 and December 2011 were included in this analysis. Medical charts were retrospectively reviewed in regard to clinical characteristics, treatment and outcome. Progression-free survival (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method. The impact of variables on PFS and OS was evaluated by univariate log-rank tests as well as by multivariate Cox regression.

Results: 82 patients were identified, 47 were male, 35 female. Median age was 60.5 years [range 19-86 years]. The site of gastrointestinal involvement was stomach in 37 patients (45%), duodenum in 4 (5%), small intestine in 19 (23%), colon in 8 (10%) and multiple gastrointestinal sites in 14 (17%). According to the international prognostic index (IPI) 24 patients (29%) were low risk (IPI 0-1), 27 (33%) intermediate risk (IPI 2-3), and 13 (16%) high risk (IPI 4-5); 18 patients had no evaluable IPI score. 5-year PFS was 65%, 5-year OS was 79%. There was no significant difference in regard to survival between the different sites of gastrointestinal involvement both in uni- and multivariate analysis. While neither age, sex, stage, and IPI score had a significant impact on survival in multivariate analysis, presence of B-symptoms was significantly associated with inferior PFS (p=.02) yet not with OS.

In regard to therapy, 66 patients (80%) were treated with a rituximab containing regimen. The chemotherapy backbone consisted of CHOP or a comparable regimen in 57 patients (70%), 21 (26%) received a more aggressive regimen (in our cohort mainly CHOP with additional etoposide (CHOEP)); four elderly patients underwent palliative treatment. Consolidative radiotherapy to the site of extranodal involvement was applied in 22 patients (27%). The addition of rituximab did not show a significant impact in regard to survival both in uni- and multivariate analysis. However, escalation of chemotherapy from CHOP to a more aggressive regimen had a significant positive impact on PFS (p=.004) and OS (p=.04) in univariate analysis; in multivariate analysis, it was significantly associated with prolonged PFS (p=.03) and showed a trend towards prolonged OS (p=.09). Patients receiving CHOEP had an excellent outcome with 5-year PFS 90% and 5-year OS 100%. While consolidative radiotherapy did not have an impact in univariate analysis, there was a trend towards beneficial impact on PFS (p=.09) and OS (p=.06) in multivariate analysis.

Conclusions: In our cohort of patients with newly diagnosed DLBCL and gastrointestinal involvement presence of B-symptoms was significantly associated with inferior PFS. Regarding treatment modalities, escalation of chemotherapy to a more aggressive regimen than CHOP as well as consolidative radiotherapy seem to be potentially beneficial therapeutic strategies.