Abstract
DLBCL can be divided into two prognostically different subgroups, germinal center B-cell-like (GCB) and activated B-cell-like (ABC), according to twodifferent gene patterns identified with cDNA microarray technology. Though valuable, this technology is expensive and not generally available. However, the identification of individual antigens related to different stages of B-cell differentiation using immunohistochemistry, can be used to assess the two profiles yielding comparable results with respect to cDNA microarray technique. We have retrospectively investigated 105 patients (pts) diagnosed with de novo DLBCL and treated at our centre between November 2001 and June 2004; the only inclusion criteria was the availability of a tissue biopsy at diagnosis. Median age was 62 (19–85); stage at diagnosis was I–II in 49 pts (47%), III in 14 pts (13%) and IV in 42 pts (40%); according to IPI, 74 (70%) pts were defined as low (0–2) and 31 (30%) as high risk (3–5). Interestingly, the majority (53%) of our pts, had a primary extranodal lymphoma at diagnosis. TMA analysis was performed with antibodies to CD10, bcl-6 and MUM1 allowing the following classification: 50 pts (48%) were considered having CGB lymphoma and 55 pts (52%) having ABC disease. According to IPI risk score, 38 pts with CGB and 36 with ABC lymphoma were at low risk (0–2) whereas 10 with CGB and 16 with ABC lymphoma were at high risk (3–5). All pts received a median of 6 cycles of a CHOP-like, antracycline-based polychemotherapy. Observed ORR was 89% (94/105); 62 (59%) pts achieved a CR, 32 (30.5%) a PR while 11 (10.5%) failed to respond to treatment. Median follow-up of the surviving pts was 45 months ( 5–110). The 3-year OS for the entire group was 71% and the 3-year EFS was 54%. In terms of CR, PD and resistance to therapy, no difference was observed between the two TMA types of DLBCL. Pts obtaining a CR after 1st line treatment were equally distributed in both groups (28.6% in CGB vs 30.5% in ABC) as were those not responding to therapy (3.8% in CGB versus 6.7% in ABC). A separate analysis in pts with stage IV disease at diagnosis was also performed and showed similar results (40.5% of CR, 23.8% of NR; no difference was observed between the two TMA defined subgroups). Pts who experienced a PD at any time, were equally distributed between the 2 subgroups, either if they relapsed after first line therapy (6 pts in CGB group , 10 in the ABC one) or after any other subsequent treatment (12 in the CGB group, 18 in the ABC one). Furthermore, 39 of 56 (69%) patients with extranodal presentation at diagnosis showed a CR independently of their subgroup distribution (37.5% in the CGL group vs 32.1% in the ABC one), and pts experiencing a PR were equally distributed between the two groups. Our data do not support the use of TMA to predict outcome in DLBCL. However, previously published data supporting the prognostic impact of TMA, have not enrolled pts with lymphomas of primary extranodal origin which, instead, were the majority in our study. Indeed, the study by Colomo et al, which enrolled 39% of pts affected by extranodal lymphomas, also failed to demonstrate a role of TMA in predicting outcome in DLBCL pts. Furthermore, the use of a limited number of antigens to define different subtypes of DLBCL may not be sufficient to identify the same patterns as defined by cDNA microarray technology.
Whose sample is it anyway? Widespread misannotation of samples in transcriptomics studies