Here's a plot that I made from data taken from the PATCH estrogen patch therapy study. It plots the % reduction in PSA from baseline (100%) at time = 0 as a function of testosterone levels over a period of 4 weeks.
As you know, estrogen effectively castrates men, causing a drop in testosterone over time to castrate levels (< 50 ng/dL). My question was: is this a linear relationship? The answer is "Yes", and one can extrapolate the trend a short distance down to zero PSA at zero testosterone level.
It's exciting to discover linear relationships from complex data sets!
Bob