It is the a measure of the quality binary (2 - class)classfications.It is similar to the Pearson coefficiient when interpretated.Yes it is important when reporting a bio-maker or a test as it can explain cross validation when using the mathews co efficient .

e.g.Correaltion of:

c = 1 indictes agreement

c = 0 indicates predication better than random

c = -1 indicates diasagreement

I hope this has helped

val

Thanks a lot  Valerie for your answer.

hello see this research 

Thanks a lot Noor for your answer.