Background High levels of cyclin E (CCNE) are accompanied by shorter survival in cyclophosphamide, hydroxydaunorubicin, oncovin and prednisone (CHOP)-treated diffuse large B-cell lymphomas (DLBCL), independent of the international prognostic index (IPI). Data on the prognostic role of CCNE in the ‘rituximab (R)-era’ are lacking.

Methods To test reproducibility and applicability of observations from the ‘pre-R era’ to the ‘R era’, we examined the prognostic role of CCNE expression by immunohistochemistry in 1579 DLBCL on tissue microarrays (TMA); 339 patients were treated by CHOP and 635 by R-CHOP.

Results 1209 samples (77%) were evaluable; failures were due to missing TMA punches and fixation artefacts. Mean expression of CCNE was 13% (0–85%); applying a cut-off of >16%, 382 DLBCL (31%) were positive. CCNE did not correlate with any of the known variables (IPI, primary site, cell of origin, proliferation, and BCL2- or C-MYC rearrangements). We were able to reproduce data suggesting an IPI- and response to therapy independent, negative prognostic impact of CCNE in CHOP-treated DLBCL patients: CCNE-positive cases had a median survival of 16 months compared with 57 months in negative ones (p=0.012). In R-CHOP-treated patients the prognostic impact of CCNE was abrogated and only IPI, cell of origin and response to therapy had a prognostic significance.

Conclusions Addition of R to CHOP overcomes the negative prognostic impact of CCNE in DLBCL. Thus, R not only prolongs survival in DLBCL but also serves a cautionary note that prognostic factors should not be transferred into the ‘R era’ without proper scientific studies.