Statistics - Level of Agreement

Hi. I'm currently doing a project looking at the predictive value of pre-treatment imaging on the pattern of tumour recurrence. In my data collection, I have divided the imaged anatomy into 6 compartments (1-6). I noted whether that compartment was involved pre-treatment and whether it was involved post-treatment. Thus, I am left with 12 columns per subject: 6 columns representing the 6 compartments pre-treatment, and 6 columns representing the 6 compartments on tumour recurrence. Is there a statistic to determine whether the disease involvement pre-treatment is predictive of the disease distribution/pattern of failure?