Comments on "A generalization of Fisher's exact test in PxQ contingency tables using more concordant relations"