Advanced diffuse non‐Hodgkin's lymphoma. Analysis of prognostic factors by the international index and by lactic dehydrogenase in an intergroup study

Background. Recent data have suggested that there are no differences among various anthracycline‐based chemotherapy regimens including cyclophosphamide, vincristine, methotrexate, and prednisone (CHOP), methotrexate, calcium leucovorin, bleomycin, doxorubicin, cyclophosphamide, and dexamethasone (m‐BACOD), methotrexate, doxorubicin, cyclophosphamide, vincristine, prednisone, and bleomycin (MACOP‐B), and cyclophosphamide, doxorubicin, etoposide, prednisone, cytosine arabinoside, bleomycin, vincristine, methotrexate, and calcium leucovorin (PROMACE‐cyta‐BOM) in patients with diffuse aggressive lymphomas. Because outcome appears to depend on certain prognostic factors, risk groups can be identified. Therefore, these prognostic factors were examined for their correlations with survival, time‐to‐treatment failure (TTF), and disease free survival (DFS) in a group of patients with diffuse aggressive non‐Hodgkin's lymphoma who were treated on a single randomized trial with either CHOP or m‐BACOD. Methods. From July 1984 to January 1988, 392 patients with diffuse large cell or diffuse mixed non‐Hodgkin's lymphoma were enrolled in an Intergroup study and were randomly assigned to treatment with CHOP or m‐BACOD chemotherapy. Of these, 325 were eligible for response, toxicity, and survival analysis, and the results were reported. The survival and TTF results now have been updated. The 286 patients who had lactic dehydrogenase (LDH) data available at study entry were analyzed for prognostic features according to the International Index criteria and using Martingale Residuals for proportional hazards regression. Results. There were no differences in survival, TTF, and disease free survival between groups of patients treated with either CHOP or m‐BACOD. In addition, analysis using the International Index criteria confirmed that patients in the lower risk groups had better outcome than patients in the higher risk groups (5‐year survival was 56 and 58% for low and low/intermediate risk groups, respectively, and 37% and 31% for high/intermediate and high risk groups, respectively). There were, however, no differences in survival, disease free survival, or TTF within any risk group when treatment with CHOP or m‐BACOD were compared. In addition, analysis using Martingale residuals for proportional hazards regression identified LDH level (< 3 × normal) as an important prognostic factor that was not captured by the International Index. Thus, 5‐year survival was 57% if LDH was normal or below, 42% if LDH was 1‐3 × normal, and 21% if LDH was < 3 × normal. Conclusion. In patients with advanced diffuse large cell or diffuse mixed non‐Hodgkin's lymphoma, there are no differences in outcome that can be attributed to treatment with CHOP vs. m‐BACOD; this holds for any prognostic group identified by the International Index. However, the level of LDH at time of study entry is an important prognostic factor that is predictive of survival and may help to identify candidates for future clinical trials. Cancer 1995; 75:865‐73.