Introduction: Involvement of >1 extranodal site is regarded as a poor prognostic factor for patients with diffuse large B-cell lymphoma (DLBCL). It is necessary to clarify the prognostic impact of specific extranodal sites.

Methods: Patients with newly diagnosed DLBCL treated at the University of Heidelberg between 06/2001 and 07/2015 were identified and included in this retrospective analysis. Data on clinical characteristics and treatment modalities were obtained by review of medical charts. 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 and by multivariate analysis using the Cox proportional hazards model.

Results: A total of 1001 patients were identified of whom 646 (65.1%) presented with extranodal involvement. Median age was 64.6 years [range 18.6-92.5], 540 (53.9%) were male. 820 patients had an available international prognostic index (IPI) score, 295 (36%) IPI 0-1, 398 (48.5%) IPI 2-3, and 127 (15.5%) IPI 4-5. The most frequently involved extranodal sites were central nervous system (CNS) (15.3%), soft tissue (14.4%), gastrointestinal tract (11.9%), bone (11.1%), lungs/pleura (9.7%), liver (6.2%), kidney (3.3%), pancreas (2.6%), and testis (2.5%).

878 patients were treated in curative intent and were further analyzed regarding the prognostic impact of extranodal involvement on outcome. 69.7% of them received CHOP-like therapies, 28.9% received chemotherapy more aggressive than CHOP, typically addition of etoposide or treatment with high-dose methotrexate in case of CNS involvement, 86.4% received additional rituximab, and 24% additional radiotherapy.

All involvement sites were evaluated by univariate analysis for OS. Sites associated with decreased OS (P <.05) were CNS, lungs/pleura, liver, kidney, malignant ascites and pharynx. Multivariate Cox Regression analysis of the 20 most common involvement sites revealed that patients with involvement of CNS, lungs and/or pleura, liver, kidney, and pharynx had significantly worse OS (P <.05).

We identified three involvement sites that had worse PFS by univariate analysis: CNS, lungs/pleura, and kidney. Multivariate Cox Regression analysis revealed that patients with involvement of CNS, lung and pleura, bone marrow, and kidney had significantly worse PFS (P <.05). On the contrary, involvement of the salivary glands was significantly associated with higher PFS in multivariate analysis (P=.044).

The involvement of gastrointestinal tract (n=107) did not have any significant prognostic impact related to OS or PFS.

Conclusions: In this retrospective registry analysis of patients with newly diagnosed DLBCL, involvement of CNS, lungs/pleura, and kidney were identified as risk factors for both inferior PFS and OS. In contrast, involvement of salivary glands was related with better PFS. Further analyses are required as toward which treatment modalities might be best suited to overcome the inferior prognosis of some extranodal involvements and to avoid the overtreatment of other involvements.

Disclosures

Kriegsmann: BMS: Research Funding; Celgene: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution