Determination of the lower limit positivity - (Feb/16/2015 )

Hi! I am new in biostatistics, and I have a question. I ’m analyzing immunohistochemical expression of a protein in different samples. For example, the 56% of cells were positive in Case1, 65% of cells were positive in Case2 and so on. My problem is that I don’t know how to determine the lower limit of the positivity (or threshold) of the analyzed protein in my experiment? What is lower limit when I could say this case is positive yet or negative below that value? Anybody has an idea?

Hmmm, that's a tricky one, and depends entirely on the protein being investigated. For instance the HER2 (herceptin) protein used to stage and grade breast cancers shows as a peri-cytoplasmic membrane ring, and is considered to be positive if more than 50% of the cell membrane stains. This is backed up by functional studies which showed that below this 50%, breast cancers would not respond to the drugs used to treat these HER2+ cancers.

In your case, you need to determine this point, which could be difficult, but have a look in the literature.