As mentioned in the title.

For example, the RNA-Seq data of patients, I use one gene set to generate a cluster result A which group the patients into group 1 to 4. And then I use another gene set to generate cluster result B which also have the 4 groups of patients. I want to see if the these patients are more grouped in the associated groups. Say, if the group 1 of cluster result A include 10 patients, and these 10 patients are mostly in the group 2 of the cluster result B, I will say this is a good fit. If all the 4 groups from A and B results can be associated with each other well. I will be very happy. I have the results looks good, but looking for a statistical test.

My summarised results as below:

        A1  A2  A3  A4
    B1  4   0   37  0
    B2  4   0   38  0
    B3  18  10  2   1
    B4  12  1   1   1

Thanks a lot.