[Non-Hodgkin's lymphomas. II. Analysis of the prognostic factors in a series of 307 patients].

PURPOSE

To analyse different clinico-biologic data in order to assess their prognostic value in non-Hodgkin's lymphoma (NHL) patients.

MATERIAL AND METHODS

The series comprises 307 patients with NHL diagnosed and treated between 1975 and 1988. The histopathologic diagnosis was revised in accordance with the working formulation system, three prognostic groups being thus considered: low-grade (LGL), intermediate-grade (IGL) and high-grade (HGL) lymphomas. Age, sex, clinical course prior to diagnosis, presence of B symptoms, histologic type, number of lymph-node areas involved, bulky disease, anaemia, thrombocytopenia, LDH, stage and response to therapy were all evaluated for the study. Survival curves were drawn with the Kaplan-Meier method, and the log-rank test was used for comparison of median survival. Whenever the univariate analysis achieved statistical significance, a multivariate analysis was performed by means of a multiple correlation and regression study in accordance with the Cox's model, in which the variables were expressed in a binary model.

RESULTS

The following 8 values were found significant in the univariate study of low-grade lymphomas: age, number of involved areas, bulky disease anaemia, thrombocytopenia, LDH, stage, and initial response to treatment. In intermediate-grade lymphomas, the significant findings were age, number of affected areas, bulky disease, thrombocytopenia, LDH, stage, and initial response. For high-grade lymphomas, number of affected areas, thrombocytopenia, LDH, stage and initial response were found statistically significant. Although no significant differences were found for survival within each of the three grades, such differences were significant between them. In the multivariate analysis, age was significant only in the LGL (p < 0.0001) in IGL, age (p < 0.07) and initial response to therapy (p < 0.0001) achieved significant value, and in HGL, stage (p < 0.02) and initial response to treatment (p < 0.0001) attained significance.

CONCLUSIONS

The univariate analysis provides various prognostic factors of statistically significance, as reported in the literature, but these after the multivariate analysis was applied, were reduced to age, stage and initial response to treatment.