Power analysis for Cox regression...a twist?


I am planning prospective trial using secondary data.

My sample size is 480. I am using an artificial grouping based on an experimental variable where:

Group 1 n= 72

Group 2 n = 408

I have not analyzed the data yet. I am seeking to determine if the experimental variable impacts PFS or OS. I am planning on using a Cox PH model but am struggling with the appropriate calculations regarding power.

The median follow up time is approximately 6 years.

The "plan" is to first evaluate:

1. The experimental variable in the entire sample population of 480.

2. The experimental variable in Group 1 compared to Group 2.

Any help or direction is truly appreciated! I did search the forum first before posting.

Thank you!