Hello researchers,

I was wondering if anyone had experience in the following:

I would like to translate two separate continuous biomarkers (A and B), that are mechanistically linked to each other, to the same scale. However, I want to avoid using an equation to estimate A from B or B from A.

Both biomarkers are rarely measured in the same cohort and so some cohorts measure biomarker A while others measure biomarker B. This is due to different companies making assays for either A or B and the established relationships between these companies and individual laboratories. Transforming both biomarker levels so that measurements of A or B can be used would help with statistical power and with repeatability of studies.

I was thinking that maybe several cut points can be established for biomarkers A and B (to make high, medium, and low count categories or something), separately, based on each biomarker's association with prognosis so that each can be mapped to the same scoring system. In doing this, I would like to have at least 3 categories.

But I am not sure how to do this and if it is even the best solution.

I hope that all makes sense. Any help will be much appreciated.

Spencer