power calc with coeff of variation

Dear all,
I am a neuroimager and I measure the levels of different chemicals in the brain. We did a test-retest study on our measures (10 subjects scanned 3 times each), and developed a coefficient of variation for our measures. One measure of interest has a shoddy CV of 0.4.
I am doing a study now comparing 2 groups (patients and controls) where I find a 50% difference between group means in this measure.
My question is, is there a way to do a power calculation using the CV of my measure and the means and SDs of my new study? Can you refer me to a website or other resource which addresses power calculations, not just with SDs but with CVs and SDs?
Thanks Jin. I just want to make sure I understand this:

- I derived one CV from repeated measures of the same subjects, so it reflects the reproducibility of the technology.

- Now I have a mean and SD derived from single measures in a whole group. The SD/mean is a little larger than the CV in the experiment above.

Is there a way to incorporate the original CV into my power calculation now?