The tumor necrosis factor (TNF) plays a key role in inflammatory processes, as this cytokine is one of the earliest to be produced in such a condition, and triggers the following cytokine cascade. In addition, the TNF and their receptor system are believed to play a key role in the growth, differentiation, and/or apoptosis of the malignant cells. As for TNF receptors, the two types, the 55 kDa (p55, TNFR; TNF-R1) and the 75 kDa (p75, TNFR; TNF-R2) are simultaneously expressed on many cells at different levels. The extracellular domains of these two receptors are released from the cell membrane by cleavage of TNF-Rs as soluble TNF-Rs (sTNF-R1, sTNF-R2). Reportedly the serum TNF-Rs level rise in patients with some malignancies. The aim of the present study was to assess the prognostic significance of serum sTNF-R in aggressive non-Hodgkin’s lymphoma (NHL). Consecutive 110 previously untreated patients with aggressive NHL (diffuse large B-cell lymphoma; 94, peripheral T-cell lymphoma; 16) prospectively participated in this study between 1997 and 2002. The patients were treated with 6–8 cycles of CHOP or THP-COP regimens. To evaluate serum levels of sTNF-Rs (p55; TNF-R1, p75; TNF-R2), venous blood samples were drawn from patients immediately before the initiation of treatment. Serum sTNF-R1 and sTNF-R2 were determined using a sandwich enzyme-linked immunosorbent assay (ELISA). In healthy control subjects, the median of serum sTNF-R1 and sTNF-R2 levels were 1.2 ng/ml (range 0.3–2.9) and 4.17 ng/ml (range 1.91–8.51), respectively. High serum sTNF-R level was associated with some poor prognostic factors and low complete remission (CR) rate. Patients with high sTNF-R1(4 ng/ml and over) and sTNF-R2 (15 ng/ml and over) at onset had significantly lower survival rates (5-year: 19%, 19%) than those with low sTNF-R1 (under 4 ng/ml) and sTNF-R2 (under 15 ng/ml) (62%, 69%), respectively (p<0.0005, p<0.0001). Multivariate analysis employing sTNF-R2 and some conventional prognostic factors demonstrated that sTNF-R2 and performance status for overall survival (OS) and sTNF-R2, sIL-2R, and LDH for event free survival (EFS) were significantly poor prognostic factors. As for TNFa, a serum TNFa level is not related with sTNF-R1 or sTNF-R2 level in aggressive NHL. In addition, serum TNFa level is not associated with OS and EFS. In conclusion, serum sTNF-R2 might be a significant prognostic factor for aggressive NHL and a useful tool for selecting the appropriate therapeutic strategy in the treatment of aggressive NHL. The most reliable prognostic factor and the best combination of some prognostic factors for aggressive NHL should be clarified in order to assist in selecting appropriate treatment.