semiquantitative scoring

Analyze Immunostained Slides with Semiquantitative Scoring

A  routine task in the lab is to investigate the presence of your favorite protein in a range of histological samples. No doubt, staining your tissue sections using good old immunohistochemistry (IHC) would be your first choice. You just got to love a technique that has celebrated its 70th birthday, and is still used in laboratories around the world. It  tells you a great many things about the localization and the abundance of your protein antigen, both within a tissue or subcellularly. Still, image analysis is difficult.

Below I discuss using a few semiquantitative scoring methods I find helpful.

Types of Stain Variation in Your Slides

So, after you get some work done, you will eventually wind up facing your old friend, the microscope.  You will see  a bunch of immunopositive cells against a counterstained background. Yet, slide after slide, you also notice that your samples vary considerably both in the percentage of stained cells as well as in staining intensity. Sometimes your antigen is widely spread out across the section. Other times it is confined to a small part. Sometimes you have a very intense staining, and sometimes it is very faint. And, just to make things more fun, an intense staining can be narrowly localized, while a vague one can be all over your section, and vice versa.

So is there any way to compare different samples in relation to the extent of IHC staining? Can we subject these variations to statistical analysis? It would be a great to describe the extent of IHC staining within a single tissue section, using a number. Semiquantitative scoring can help.

Semiquantitative Scoring can be Used with Variable Staining

So how do we quantify something we see in an image? So far, there is no perfect, no simple, unbiased or universally accepted answer to this question. The most often used solution is an unfortunate method called  “semiquantitative scoring”. Whoever coined this term probably meant to say, that in spite of the fact that you are trying to quantify your results, it’s just not all about counting.

In common practice, semiquantitative scoring means setting up an arbitrary scale of scores, ranging from zero to whatever number seems suitable to you. Then you assign a score to each slide. This score should represent as accurately as possible the extent of IHC staining. This can be done using two different approaches, with a whole palette in between them.

Combined Stain Intensity and Percent of Cells Positive Approach

You can empirically grade your samples as a combination of staining intensity and the percentage of positive cells. For example: grade 0 for no reaction or focal weak reaction; grade 1 for intense focal or diffuse weak reaction; grade 2 for moderate diffuse reaction; and grade 3 for intense diffuse reaction.

This is a fairly simple method which gives you an ordinal scale of scores from 0 to 3. However, you can have a fair share of your samples that are “somewhere in between” and hard to classify.

Score Staining Attributes Separately

You can independently evaluate the intensity, say, from 0 to 3. Then score the distribution of immunoreactive cells, e.g. grade 0 for less than 10%, grade 1 for 10-50%, grade 2 for 50-80%, and 3 for more than 80% of reactive cells. Lastly, multiply the two numbers to obtain the final score.

This method givese you an ordinal or a continuous variable data, depending on the number of categories you set up for the intensity grade and how precisely you are able to count your stained cells.

Once you are done with your grading, use the data to compare different samples. You can group your cases (e.g. into a high or low expression profile), and analyze the groups independently. Also you can correlate IHC results with other characteristics of the samples or the results of related experiments. You can do all sorts of fun things we like to do with numbers.

A Few More Remarks on Semiquantitative Scoring

To evaluate your IHC staining, you can pick a scoring system from an extensive literature search or  set one up yourself. However, from my experience, there are a few things you should keep in mind:

  • The best scoring system for your experiments is not the one most cited in literature. Nor is it the one your colleague uses in his/her similar set of experiments. It is the one that best represents your results, in your experimental design, and under your experimental conditions.
  • The interpretation of IHC results takes a trained eye. After you finish staining all your samples, spend some time in front of a microscope. Carefully examine each slide before you start scoring, just to get “the feeling of it”. Take your time with the microscope, it won’t be wasted.
  • It can be very useful to have an independent observer go over your samples, to avoid potential observer bias. Reexamine all sections where consensus regarding the score was not reached. If no agreement is made, exclude these cases from the study.
  • Negative and positive control slides give you a sense of what a lowest and a highest graded sample looks like: use them!
  • Different scoring criteria in similar experiments published by other authors, can make it hard to compare your results with the ones in literature. Keep this in mind when citing these articles when you start writing your own.

 

Leave a Comment