Predictability of monthly and yearly rhythms of breast cancer features

In order to evaluate if breast cancer biological characteristics undergo significant menstrual and seasonal variations, we analysed in a consecutive series of 905 breast cancer patients, steroid receptor level (ER and PgR by DCC assay), proliferative activity (3H-Thymidine Labeling Index, 3H-TLI) and size of primary tumour in relation to calendar date and day of menstrual cycle at the time of the surgical procedure. For data analysis, the method of time series construction and classical spectral analyses with Bartlett Kolmogorov-Smirnov test for white noise (BKS test) was utilised. For what concerns menstrual variations, 3H-TLI showed a significant periodicity (t = 0.3146, p < 0.01 by BKS test) with peaks at day 12nd and day 18th; ER showed a significant periodicity (t = 0.3605, p < 0.01 by BKS test) with more evident peak at day 27th; PgR, a significant periodicity (t = 0.160, p = 0.05 by BKS test) with peaks at day 15th and day 24th, similar to that observed for tumour size (t = 0.19, p < 0.05 by BKS test). With respect to yearly fluctuations, 3H-TLI showed only a trend for a significant rhythm (t = 0.16, p = 0.06 by BKS test) with peaks in May and November; ER a significant periodicity (t = 0.2099, p < 0.05 by BKS test) with two evident peaks in January and April; also for PgR a significant periodicity (t = 0.3161, p < 0.05 by BKS test) was demonstrated with a peak in July; finally, tumour size showed a significant rhythm (t = 0.335, p < 0.01) paralleling 3H-TLI behaviour. Finally, the analysis of variance with interaction of menstrual and seasonal timings showed that only the seasonal timing was able to independently influence the 3H-TLI variations (3H-TLI higher in spring). We confirmed that breast biology has significant menstrual and seasonal variations and that the seasonality is probably the timing factor more relevant in periodicity determination.