Significant circadian cell cycle variations with a maximal number of cells in S-phase during the night have been found in a series of 24 patients (18 men and 6 women) with histologically established non-Hodgkin's lymphomas. Pathological lymph nodes of a total of 26 patients were punctured and aspirated by fine needle technique every 4 h during a single 24-h time span. Twenty-four patients (92.3%) had Stage III or IV disease. Twelve patients (46.1%) had low grade, 10 patients (38.5%) had intermediate grade, and 4 patients (15.4%) had high grade lymphomas according to the Working Formulation. The samples were analyzed by flow cytometry, and DNA synthesis (S-phase) and ploidy were determined according to circadian stage. The individual mean 24-h S-phase varied from 2.2 ± 1.2% (mean ± SD) to 24.0 ± 3.3%. Within the group of patients with low grade lymphomas, a wide range in mean S-phase from 2.4 ± 1.2% to 9.2 ± 2.8% was observed.

The percentage variation within each patient between the lowest and highest S-phase as compared to the lowest value (range of change) during the 24-h time span varied from 21 to 353%, with a mean range of change of 128 ± 19%. When each individual S-phase series was converted to percent of mean and combined for analysis by one-way analysis of variance to test for time-effect across 2 12-h time spans (8 p.m.-8 a.m. versus 8 a.m.-8 p.m.), S-phase variation according to circadian stage was found to be statistically significant (P < 0.004), with higher values found in the 8 p.m.-8 a.m. time span. By single cosinor analysis, S-phase yielded a near significant P value of 0.069 for the least-squares fit of a 24-h cosine to all data as percent of mean, with the acrophase found to be near midnight (0.05 h). For those patients with low and intermediate grade lymphomas and with mean S-phase values <10.0%, we found that mean S-phase was higher during winter (5.8 ± 0.4%) than during spring (3.8 ± 0.3%) or during fall (3.6 ± 0.3%) (P < 0.001, analysis of variance). Twenty-one of the 26 patients (80.8%) had an aneuploid, hypodiploid, or near diploid population in one or several of the repeated samples. For the whole series, the DNA indices for the aneuploid populations varied from 1.09 to 1.96, the median DNA index being 1.20. A combination of already established prognostic factors together with an optimally timed estimation of DNA synthesis activity as well as DNA content, i.e., ploidy, might increase the predictive accuracy relative to response in the individual patient, as well as guiding the selection of dose and a properly timed chronotherapeutic schedule of drug delivery.