Treatment outcome and prognostic factors for relapse after high-dose chemotherapy and peripheral blood stem cell rescue for patients with poor risk high grade non-Hodgkin's lymphoma.

The aim of the study was to determine treatment outcome and identify a particularly high risk group in a consecutive series of 66 patients with poor prognosis high grade lymphoma (NHL) treated with conventional induction chemotherapy followed by high-dose chemotherapy (HDCT) and peripheral blood stem cells (PBSC) rescue. Fifty-one patients with intermediate grade NHL (Kiel) and two or three adverse prognostic features as defined by the age-adjusted International Prognostic Index (IPI) received induction treatment with 7 weeks of doxorubicin, cyclophosphamide, vincristine, bleomycin, etoposide, prednisolone and methotrexate (VAPEC-B) followed by three cycles of ifosfamide/cytarabine. Fifteen patients with high grade Burkitt's and lymphoblastic NHL received 11 weeks of VAPEC-B followed by three cycles of high-dose methotrexate. HDCT for all 66 patients consisted of busulphan/cyclophosphamide followed by autologous PBSC rescue. Thirty-one patients (47%) received HDCT in first complete remission (CR/CRu) and 34 patients (52%) in first partial remission (PR) after conventional chemotherapy. Following HDCT, 42 patients (64%) were in CR/CRu, 19 patients (29%) in PR and one patient had progressive disease. There were four toxic deaths. After a median follow-up period of 27 months (range 7-73) in 46 surviving patients, the actuarial 3-year estimates of overall survival, event-free survival (EFS) and freedom from progression (FFP) were 67%, 65% and 70%, respectively. In univariate analysis, prognostic factors associated with reduced EFS were mediastinal bulk (P = 0.02), > or = 3 extra-nodal sites (P = 0.02), remission status prior to HDCT (P = 0.05), low albumin (P = 0.08) and raised ESR (P = 0.09). No significant difference was observed between patients with intermediate or high grade NHL or between patients with two or three adverse IPI features. Multivariate analysis identified mediastinal bulk (P = 0.01), > or = 3 extra-nodal sites (P = 0.01) and low albumin (P = 0.03) as joint predictors of poor EFS. Remission status prior to HDCT was not found to be significantly associated with reduced EFS, FFP or survival, suggesting early introduction of HDCT may benefit patients with a PR. Based on these three adverse features, three groups (0, 1 or > or = 2 features) could be identified with differing EFS, survival and freedom from progression (FFP) rates at 3 years; 85%, 63% and 20%, respectively for EFS, 84%, 64% and 25% for survival and 85%, 66% and 33%, respectively for FFP. This prognostic model may identify patients with a particularly poor prognosis despite HDCT, who may benefit from other therapeutic approaches.