AgNORs predictive value of prognosis in non-Hodgkin's lymphoma: comparison with flow cytometric cell cycle analysis.

Paraffin-embedded histopathologic specimens, taken before the commencement of therapy from 14 low-grade and 21 high-grade malignant lymphoma patients, and 9 normal lymph nodes were utilized to analyze six cell DNA-related parameters. The flow cytometry technique was used to determine cell-cycle G0/G1, S and G2/M phases, and silver staining to enumerate nuclear organized regions (AgNORs); nucleus surface area was determined by an image-analyzing system. The six parameters and natural logarithm of cell proportion in the S-phase (LS) were determined according to the histologic tumor type and achievement of the first complete remission (CR). All parameters except cell proportion in G1/M cycle phase differed significantly with respect to histologic cell type, but were not related to the achievement of first CR. Inasmuch as the parameters significantly correlated with each other, multivariate discriminant analysis and proportional hazard regression were applied to estimate their discriminant/predictive values with respect to tumor malignancy. AgNORs proved to be far superior in all three clinical parameters, S-phase was significantly predictive for the achievement of first CR, and LS for tumor histology type. The statistical model applied narrowed down the analysis of seven parameters to two with respect to tumor histology type (AgNORs and LS) and achievement of first CR (AgNORs and S), but only to one for overall patient survival (AgNORs). Only the model for tumor histology type discrimination was statistically significant (R2 = 0.904, p < 0.001). It appears that AgNORs may be of utmost predictive importance for the clinical outcome in NHL.