Major prognostic factors of adult patients with advanced B-cell lymphoma treated with vincristine, cyclophosphamide, prednisone and doxorubicin (VEPA) or VEPA plus methotrexate (VEPA-M).

Eighty-two adult patients with advanced B-lymphoma, treated between 1981 and 1983 with VEPA (vincristine, cyclophosphamide, prednisolone and doxorubicin) or VEPA-M (VEPA plus methotrexate) in a prospective randomized fashion, were evaluated for pretreatment characteristics. The overall complete response (CR) and the 4-year survival rates were 74% and 45%, respectively. The relapse rate was 51%. Stage of disease only was negatively associated with the CR rate in a multivariate analysis. The primary extranodal tumor site other than upper gastrointestinal (GI) tract and high grade pathology were found to affect disease-free survival adversely in a Cox proportional hazards model. Poor performance status, advanced stage, primary extranodal tumor site other than upper GI tract, advanced age, high grade pathology and prior therapy by either surgery or radiation, were significantly associated with shortened survival in a Cox proportional hazards model. These results indicate advanced B-lymphoma in Japan to be generally similar to advanced non-Hodgkin's lymphoma in the West in terms of prognostic factor characteristics, but the importance of the primary site in predicting survival has not been reported in the West. Also, the lack of a survival plateau in patients with diffuse large cell lymphoma indicates more intensive chemotherapy regimens than VEPA or VEPA-M to be needed. It was also found that the significant prognostic factors in patients with advanced B-lymphoma were very different from those with T-lymphoma. The five factors: pathology, stage, primary site, age, prior therapy by surgery or radiation, for which the risk ratio was more than 2.3, were used to construct a model containing 23 categories of patients running an increasing risk of shortened survival; this divided patients into three groups. The CR and 4-year survival rates of low-, moderate- and high-risk groups were 90% and 74%, 74% and 58%, and 50% and 5%, respectively. The risk-grouping provides indications for determining optimal therapy for individual patients and the need for new therapeutic trials in patients at high risk.