Q. I’ve Googled and YouTubed this thing to death, and I still can’t grasp the meaning of “differentiation.” It seems the opposite of its definition. To “differentiate” means to recognize what makes something different. But according to your post on tumor differentiation, well-differentiated tumors resemble (don’t look different from) their tissue of origin. I would think if something is well-differentiated, it would look very different from the thing it’s being compared to. Why is the use here opposite of its meaning?

A. I totally get where you’re coming from. It’s REALLY frustrating in pathology when things are described in terms that don’t seem to make sense. You are not alone in questioning the use of this term!

The problem is that the word in question – differentiation – has a specific meaning in the real world. You’re exactly right in your definition: to differentiate between two things means to recognize what’s different or unique.

So you’d think that “differentiation” in the pathology world would mean the same thing: the recognition of things that are unique, different, or not the same. By logical reasoning, then, a “differentiated” tumor would be one that looked different from its cell of origin. And you’d think a “well-differentiated” tumor would be one that looked very different from its cell of origin.

Unfortunately, “differentiation” doesn’t have the same definition in the pathology world. So we have to put aside our logic and knowledge of vocabulary for a moment, irritating as that may be, and learn a new definition for this word.

The definition of “differentiation” in pathology-speak.

When we’re talking about tumors, the definition of “differentiation” is simply this: the degree to which tumor cells resemble their cell of origin. A well-differentiated tumor is one in which the tumor cells look very much like their cell of origin. A poorly-differentiated tumor (like the poorly-differentiated squamous cell carcinoma shown above) is one in which the tumor cells barely resemble their cell of origin.

That’s it. Yes, it’s an annoying word choice, because it is used here in a way that seems counterintuitive. But maybe it’s not as far off as it seems.

Maybe this will help.

I think about it (okay, rationalize it) this way. When cells are really immature, they don’t have a lot of features that make them look different from other cells. Myeloblasts don’t look very different than lymphoblasts, for example. So we could say that these immature cells are undifferentiated; it’s hard to tell what kind of cell they really are, and hard to tell them apart from other cells.

The same thing is true of the cells in poorly-differentiated tumors! The cells show practically no features that give away their identity; it’s hard to even tell what kind of cells they are. They are, in effect, undifferentiated.

If you think about “differentiation” this way (undifferentiated cells lack identifying features; it’s hard to tell what kind of cell they are), then the concept of tumor differentiation is a little easier to swallow. A little.